[Sursound] Call for papers: IEEE Multimedia Signal Processing 2019

[Politis Archontis]

(Apologies for cross-posting)

Call for papers: Special session on Computational Audio Intelligence for 
Immersive Applications

IEEE 21st International Workshop on Multimedia Signal Processing (MMSP) 2019.

Special session website: https://dr-costas.github.io/caiia-mmsp2019/

=

Important dates:
Paper submission: 18th of May
Notification of acceptance: 18th of July
Camera ready submission: 30th of July

=

Scope and topics:
Audio is a key component of immersive multimedia applications. Beyond simply 
communicating information to the user through speech, music, and other 
context-specific sound, the audio and acoustics can serve to convey a sense of 
spatial awareness and presence. This is particularly important in the context 
of mixed reality, where virtual content is presented spatially to the user to 
augment or enhance the real environment.

Recent advances in machine listening and audio analysis proved successful at 
extracting information about the environment, e.g., by performing acoustic 
scene analysis, estimating acoustic parameters for spatially positioning 
virtual acoustic objects, or localizing sound sources and events. The 
availability of this type of information has the potential to greatly improve 
the user experience in immersive multimedia applications. Examples include 
providing real-time feedback about the surroundings, automatically suppressing 
or enhancing specific sounds, and altering the spatial information of spatially 
rendered content for artistic purposes or to better blend it with the real 
environment.

The scope of this special session is to explore this intersection between 
computational audio intelligence and immersive audio research, and provide a 
venue for hosting recent advances in these two audio research directions. This 
will create the opportunity for scientific interaction that can significantly 
bolster research in both computational audio intelligence and immersive audio 
from a joint perspective.

The special session targets researchers and professionals working in 
computational audio intelligence and immersive audio processing, and in 
particular focusing on applications which synergically exploit techniques 
related to these research fields.

The topics of the special session include (but are not limited to):

* Machine listening for analysis of acoustic scenes
* Localization, tracking, and separation of sources and sound events
* Generative and data-based enhancement and synthesis of immersive audio
* Spatial audio analysis, modification, and synthesis
* Context-aware immersive audio processing and applications
* Deep learning for analysis of real-life audio recordings
* Audio augmented and mixed reality
* Machine learning for acoustic simulation and personalization

On behalf of the organising committee,
Archontis Politis

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Enquiry on upmixing from 1st order ambisonics to 3rd order ambisonics.

[Politis Archontis]

Hi David,

These upmixing methods extract a lot of information from the FOA recording that 
is then re-used to essentially “synthesize" the HOA signals, with a spatial 
resolution that would not be possible with the FOA recordings. They are 
“active” in that sense, and signal-dependent, compared to the “passive" 
classical ambisonic decoding. Their success depends of course on how effective 
is their underlying model and how robustly they are implemented.

In that sense there isn’t necessarily a large benefit in parametric upmixing 
from FOA to 3rd-order, compared to parametric decoding for playback, since 
these methods can also upmix directly from FOA to, say, 40 speakers or 
headphones, with their maximum sharpness. However, the HOA upmixing could be 
useful for people that are working with a HOA processing pipeline, and they 
want to integrate FOA or lower-order material seamlessly.

Regards,
Archontis Politis


> On 22 Feb 2019, at 19:45, David Pickett  wrote:
> 
> At 02:37 22-02-19, Wilson Lim wrote:
> 
>> ot sure if I have missed a discussion about upmixing with ambisonics on
>> Sursound.
>> 
>> Just wondering if anyone is willing to share some information on how to
>> implement upmixing algorithms from 1st Order Ambisonics A&B-format to 3rd
>> Order Ambisonics B-format.
> 
> I am curious to know what advantage there is to playing back 1st order 
> upmixed to 3rd order. Doesnt it still sound like 1st order, since there no 
> information is actually added? Or are the images expectated to be more stable 
> on account of more loudspeakers being involved, on the analogy of stereo 
> played back through three loudpeakers?
> 
> David
>  
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Enquiry on upmixing from 1st order ambisonics to 3rd order ambisonics.

[Politis Archontis]

Hi Steven,

As far as I remember the Rode plugin does not offer this functionality 
(upmixing FOA to HOA).

Regards,
Archontis Politis


On 22 Feb 2019, at 14:04, Steven Boardman 
mailto:boardroomout...@gmail.com>> wrote:


And lets not forget ‘Soundfield’ by Rode.
This also uses some form of parametric direction detection. Either on B-format 
directly, but also on the Rode’s A-format.


Steve


Oh my mistake then, I thought the HARPEX vst was doing direct rendering only, 
and its HOA upmixing functionality was done only by the Blue Ripple Sound 
upmixer (which is using HARPEX).
Three it is then :-).

Best,
Archontis

On 22 Feb 2019, at 11:05, Axel Drioli 
mailto:axeldri...@gmail.com> 
>> wrote:

To add on top of the previous email, there are three then ( I didn t know
about COMPASS), because Harpex can be a vst itself, which sounds
magnificently good.

Regards
Axel
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Enquiry on upmixing from 1st order ambisonics to 3rd order ambisonics.

[Politis Archontis]

Oh my mistake then, I thought the HARPEX vst was doing direct rendering only, 
and its HOA upmixing functionality was done only by the Blue Ripple Sound 
upmixer (which is using HARPEX).
Three it is then :-).

Best,
Archontis

On 22 Feb 2019, at 11:05, Axel Drioli 
mailto:axeldri...@gmail.com>> wrote:

To add on top of the previous email, there are three then ( I didn t know
about COMPASS), because Harpex can be a vst itself, which sounds
magnificently good.

Regards
Axel

On Fri, 22 Feb 2019 at 07:52, Politis Archontis 
mailto:archontis.poli...@aalto.fi>>
wrote:

Hi Wilson,

First of all you have to understand the upmixing algorithms themselves
before starting implementing them. Some methods that can do that are
Directional Audio Coding (DirAC), which you can read all about in the
“Parametric Time-frequency Domain Spatial Audio” book of last year, and
numerous articles from AES journal and conferences. Then there is the
HARPEX method by Svein Berge for FOA signals, you can find the articles
online. Another option is the recent COMPASS method which can also upmix
second-order or 3rd-order if needed. Finally, an alternative is the sparse
recovery approach by Wabnitz, Jin and Epain, published mostly in IEEE
conferences.

There are two plugins I know of offering this functionality at the moment,
one by Blue Ripple Sound that uses HARPEX, and one based on COMPASS by us,
which is to be released this month. In general if you have no previous
experience with the various parts of these methods, time-frequency
transforms, parameter estimation, adaptive filtering, maybe decorrelation,
and others.. expect to spend many months testing things till you manage to
get good audio quality.

Best regards,

Archontis Politis

On 22 Feb 2019, at 03:37, Wilx Wilson 
mailto:wilxson@gmail.com>mailto:wilxson@gmail.com>>> wrote:

Hi Everyone,

Not sure if I have missed a discussion about upmixing with ambisonics on
Sursound.

Just wondering if anyone is willing to share some information on how to
implement upmixing algorithms from 1st Order Ambisonics A&B-format to 3rd
Order Ambisonics B-format.

I am really curious to know on how they are implemented mathematically.
Because I am developing a prototype plugin for my research project.

Have a good day!

Much Obliged,
Wilson Lim
-- next part --
An HTML attachment was scrubbed...
URL: <
https://mail.music.vt.edu/mailman/private/sursound/attachments/20190222/c16a21c1/attachment.html

___
Sursound mailing list
Sursound@music.vt.edu<mailto:Sursound@music.vt.edu>
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
edit account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: <
https://mail.music.vt.edu/mailman/private/sursound/attachments/20190222/782f0c2f/attachment.html

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
edit account or options, view archives and so on.

--
*Axel Drioli*
*axeldrioli.com<http://axeldrioli.com> <http://axeldrioli.com/>*

*For immersive recordings visit 
ImmersiveFieldRecording.com<http://immersivefieldrecording.com>
<http://immersivefieldrecording.com/>*
*For Spatial Audio production visit 
SpatialAudioLabs.com<http://spatialaudiolabs.com>
<http://SpatialAudioLabs.com<http://spatialaudiolabs.com>>*

*Tel-Facetime: +44 7460 223640*
*Skype: axel.drioli*
*E-mail: a...@spatialaudiolabs.com<mailto:a...@spatialaudiolabs.com> 
mailto:a...@spatialaudiolabs.com>>*




*Mostly LDN. UK.*
-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20190222/037cb2c2/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu<mailto:Sursound@music.vt.edu>
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20190222/ad6213fc/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Enquiry on upmixing from 1st order ambisonics to 3rd order ambisonics.

[Politis Archontis]

Hi Wilson,

First of all you have to understand the upmixing algorithms themselves before 
starting implementing them. Some methods that can do that are Directional Audio 
Coding (DirAC), which you can read all about in the “Parametric Time-frequency 
Domain Spatial Audio” book of last year, and numerous articles from AES journal 
and conferences. Then there is the HARPEX method by Svein Berge for FOA 
signals, you can find the articles online. Another option is the recent COMPASS 
method which can also upmix second-order or 3rd-order if needed. Finally, an 
alternative is the sparse recovery approach by Wabnitz, Jin and Epain, 
published mostly in IEEE conferences.

There are two plugins I know of offering this functionality at the moment, one 
by Blue Ripple Sound that uses HARPEX, and one based on COMPASS by us, which is 
to be released this month. In general if you have no previous experience with 
the various parts of these methods, time-frequency transforms, parameter 
estimation, adaptive filtering, maybe decorrelation, and others.. expect to 
spend many months testing things till you manage to get good audio quality.

Best regards,

Archontis Politis

On 22 Feb 2019, at 03:37, Wilx Wilson 
mailto:wilxson@gmail.com>> wrote:

Hi Everyone,

Not sure if I have missed a discussion about upmixing with ambisonics on
Sursound.

Just wondering if anyone is willing to share some information on how to
implement upmixing algorithms from 1st Order Ambisonics A&B-format to 3rd
Order Ambisonics B-format.

I am really curious to know on how they are implemented mathematically.
Because I am developing a prototype plugin for my research project.

Have a good day!

Much Obliged,
Wilson Lim
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Sursound Digest, Vol 126, Issue 14

[Politis Archontis]

Thanks for the feedback Florian, is that on Reaper? The 3rd-order version can 
be quite heavy. Can you check the CPU usage on the plugin window (if on Reaper? 
maybe elsewhere in another DAW)

Cheers,
Archontis


On 22 Jan 2019, at 04:14, Florian Grond 
mailto:floriangr...@gmail.com>> wrote:

Hi Archontis, I just tried the COMPASS plugin, really great stuff like the
other SPARTA plugins too, with 3rd order settings I get xruns in the audio
only 2nd is stable, it's better when I close the GUI.

Thanks for making those!

Florian

On Sun, Jan 20, 2019, 12:00 
mailto:sursound-requ...@music.vt.edu> wrote:

Send Sursound mailing list submissions to
   sursound@music.vt.edu<mailto:sursound@music.vt.edu>

To subscribe or unsubscribe via the World Wide Web, visit
   https://mail.music.vt.edu/mailman/listinfo/sursound
or, via email, send a message with subject or body 'help' to
   sursound-requ...@music.vt.edu<mailto:sursound-requ...@music.vt.edu>

You can reach the person managing the list at
   sursound-ow...@music.vt.edu<mailto:sursound-ow...@music.vt.edu>

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Sursound digest..."


WHEN REPLYING EDIT THE SUBJECT LINE

ALSO EDIT THE MESSAGE BODY

You are receiving the digest so when replying, please remember to edit
your Subject line to that of the original message you are replying to, so
it is more specific than "Re: Contents of Sursound-list digest…" the
subject should match the post you are replying to.

Also, please EDIT the quoted post so that it is not the entire digest, but
just the post you are replying to - this will keep the archive useful and
not polluted with extraneous posts.

This is the responsibility of digest subscribers. the community and list
subscribers care about the integrity of the threads and archives so this is
important.

Today's Topics:

  1. Re: Impressive transaural demo (Marc Lavall?e)
  2. Re: Impressive transaural demo (Ralph Glasgal)
  3. Announcing COMPASS decoder VST plugin (Politis Archontis)


--

Message: 1
Date: Sat, 19 Jan 2019 13:16:19 -0500
From: Marc Lavall?e mailto:m...@hacklava.net>>
To: sursound@music.vt.edu<mailto:sursound@music.vt.edu>
Subject: Re: [Sursound] Impressive transaural demo
Message-ID: 
<9fd04f6e-b589-2034-2498-675e6901f...@hacklava.net<mailto:9fd04f6e-b589-2034-2498-675e6901f...@hacklava.net>>
Content-Type: text/plain; charset=utf-8; format=flowed

I'm not aware of an online demo.

For an overview of the hardware requirements, check:
https://www.theoretica.us/bacch4mac/

I have the software (version 5.2) running on a laptop. I don't have a
fancy USB audio module or the super-fancy BACCH-BM microphone; I use my
cheap binaural microphone for the calibration, and it works quite well
with any stereo recordings, even better with binaural recordings. My
loudspeakers are quite ordinary and the result is already very
impressive, so I suspect that it would sound amazing with the
recommended hardware and audiophile loudspeakers.

Marc


Le 2019-01-19 ? 11:09 a.m., Augustine Leudar a ?crit?:
This looks really interesting Marc - are there any demos I could hear
online ? or does it need specific hardware to work ?

On Fri, 18 Jan 2019 at 21:24, Marc Lavall?e  wrote:

This helicopter demo does work! But it works better when controlling
first reflections.

Something else that works *really* well is BACCH-dSP.

Marc

Le 2019-01-18 ? 1:42 p.m., Augustine Leudar a ?crit :
anyone heard any other transaural demos that are work, or kinda work ?

On Fri, 18 Jan 2019 at 18:40, Augustine Leudar <
augustineleu...@gmail.com>
wrote:

ps - on speakers not headphones

On Fri, 18 Jan 2019 at 18:38, Augustine Leudar <
augustineleu...@gmail.com>
wrote:

This actually suprised me by working :




http://wavearts.com/sounds/PanoramaDemos/Helicopter%20Circling%20(speakers).mp3

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
edit account or options, view archives and so on.




--

Message: 2
Date: Sat, 19 Jan 2019 15:31:06 -0500
From: "Ralph Glasgal" 
To: "'Surround Sound discussion group'" 
Subject: Re: [Sursound] Impressive transaural demo
Message-ID: <072101d4b035$e7724d10$b656e730$@ambiophonics.org>
Content-Type: text/plain;   charset="us-ascii"

Glad you are willing to try this.  Maybe you can move back from the
speakers
to get a narrower angle.  The trick is not to involve your head shadow and
get as better pinna angle.  It works great with a little sound bar like the
Soundmatters or Jambox.  I have nothing to do with this except the inventor
uses my RACE equations in his program and understands about keepi

[Sursound] Announcing COMPASS decoder VST plugin

[Politis Archontis]

Hello,

we made available today a VST plugin implementation of a parametric ambisonic 
decoder based on the COMPASS model, published last year at IEEE ICASSP2018
(https://ieeexplore.ieee.org/abstract/document/8462608, also on ResearchGate).

The decoder works with first-, second-, and third-order material, and applies a 
parametric model of multiple sound sources and an ambient component.
More available channels (higher-orders) allow additional sound source 
components to be analyzed. The decoder can sound sharper than a traditional 
linear ambisonic decoder, especially at first-, and second-order material. The 
user can manipulate the balance between the analyzed source and ambient 
signals, and control the percentage of “parametric”-vs-“linear” decoding.

Furthermore, as with the previous SPARTA suite of loudspeaker-based plugins 
released by the Acoustics Lab in Aalto University, the plugin allows definition 
of custom loudspeaker setups, and headphone monitoring using generic HRTFs, or 
user-provided personalized ones in the SOFA format.

You can find the plugins, more info and download links here:

http://research.spa.aalto.fi/projects/compass_vsts

There is also a video demo in the web page demonstrating functionality over a 
range of first-order and third-order material.

Best regards,

Archontis Politis
Post-doctoral Researcher
Department of Signal Processing and Acoustics
Aalto University
Finland

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Publishing music in ambisonics

[Politis Archontis]

> On 13 Jan 2019, at 22:08, Bo-Erik Sandholm  wrote:
> 
> Angelo Farina has created a metadata "injector" to make ambisonic work in
> VLC.
> I think he described it it a posting in the ambisonic 360 VR audio facebook
> group

Actually, the standard “spatial media metadata” injector provided by Google 
(the one for preparing Youtube 360 videos) worked for me with FOA ambiX in VLC. 
For 3rd-order files however there is a modified version of Google’s one around, 
which I’m pretty sure Prof. Farina has.

Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Publishing music in ambisonics

[Politis Archontis]

Hi Marc,

I believe it expects the “spatial media” metadata to be written in the file for 
VLC to interpret it as ambisonic.
It worked for me in the past. Yes, it’s a shame that it is completely 
undocumented, especially since it was advertised as one of the cool new 
features of VLC 3.

The decoder is also possible to playback 3rd-order AmbiX files, with a more 
convoluted process. With a bit of digging I realized VLC uses a fixed decoder 
with 27 loudspeakers, the 20 on the vertices of a dodecahedron, and 7 more on 
the horizontal plane. I was able to replace the stock HRTF filters for these 
directions coming with VLC, with a hand-crafted SOFA file of own HRTFs, with 
good results.

Archontis Politis


On 13 Jan 2019, at 20:00, Marc Lavallée 
mailto:m...@hacklava.net>> wrote:

Le 2019-01-13 à 12:58 p.m., Marc Lavallée a écrit :

I installed VLC (3.0.6 Veritani) on my Linux laptop

Oops, I meant "Vetinari"...

Marc

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Publishing music in ambisonics

[Politis Archontis]

Hi,

In JSAmbisonics online demos we have an example that decodes to 8 speakers, 
working in Chrome with success. I remember we had some problem getting output 
for more channels out at that time, but that was a year ago and things may have 
changed in Web-Audio.

In general, the library supports one to provide and use their own decoding 
matrices for 1st-15th order for 2D, and 1st-4th order for 3D, hence one could 
write their preferred optimized decoding matrices to decode to 5.1 or 7.1, or 
binaural, and give the option in a player to switch from one to the other. The 
library can also compute internally a decoding matrix, based on ALLRAD, which 
should work ok in most cases.

I haven’t used Omnitone, but it may be providing the same functionality now. 
Since these libraries take care of the processing and decoding, and provide 
examples for playing B-format files, FOA and HOA, in any connvention, wrapping 
them in a web-based audio player should be trivial.

Best regards,
Archontis Politis



On 13 Jan 2019, at 12:35, Bo-Erik Sandholm 
mailto:bosses...@gmail.com>> wrote:

https://developer.mozilla.org/en-US/docs/Web/API/AudioDestinationNode

Destination channel numbers can be as many as soundcard outputs !
So yes there is nothing stopping us from having multichannel outputs from a
web audio implemented in JavaScript..

Bosse


On Sun, 13 Jan 2019 09:58 Bo-Erik Sandholm 
mailto:bosses...@gmail.com> wrote:


As far as I understand there is nothing stopping us to access a
multichannel soundcard from the browser.

To Marc, there is nothing else than the html competence and time that
stops us from using Omnitone or JSAmbisonics to produce websites with
Ambisonic content.

I think with the www.ohti.xyz, this is produced with faint 
memories of a
html course in the end of the eighties and googling shows this.

The thing that I have not been able to figure out is how to create a
interface for creating playlist or a GUI for a directory at files for click
and play.

The VLC supports ambix coded soundfield, there is somwhere a multimedia
version in the works that will make it possible to use headtracking
according to the VLC main developer.

Bosse




- -

Decoding in the browser would be for casual use, mostly for binaural
listening, but decoding to speaker arrays would be nice, for exemple
with 5.1 system (as a 4.1 system with a square or rectangular setup).

- - -

Why would you not be able to decode ambisonics to a speaker array (for
example 4.0 or 5.1), <  from a browser >?


Maybe this is no common option yet today, but why not in the future?

If a browser is able to support 5.1 (stereophonic surround), decoding
of ambisonics to some 5.1 system should actually be no problem. (via
WebAudio)

The usual way to support decoding of ambisonics (only) to binaural is
because the normal application case of ambisonics is nowadays to be
some “ audio track” for VR or 360° video.

Right?!

I don’t think this is just an abstract discussion, by the way. Maybe
such decoding functions could added to Omnitone, for example?


-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Soundfield by Rode plugin

[Politis Archontis]

Hi,

There are various ways to do it, one is to assume a sound-field model (e.g. 
direct & diffuse sound), estimate the model parameters in each frequency bin, 
and then use them to align the signals in a coincident way, such a parametric 
approach was presented by us in AES Convention in Milan last year for similar 
applications and higher-order recording.

Another very sensible approach was presented by Cristoff Faller and Illusonics 
in the same conference, in a simpler adaptive filter is used to align the 
microphone signals to the phase of one of the capsules, making them again in 
essence coincident.

Parametric approaches for playback and recording (see e.g. DirAC, HARPEX, 
countless speech enhancement methods) have gone a long way in the last decades 
in terms of quality and they’re definitely not “snake oil” :-).

Best regards,
Archontis Politis
Post-doctoral Researcher
Department of Signal Processing and Acoustics
Aalto University
Finland






On 17 Dec 2018, at 11:39, Dave Hunt 
mailto:davehuntau...@btinternet.com>> wrote:

Hi,

There is slightly more description of their A to B-format processing
(but not much) in Rode's blog:



>From that web page.

"The SoundField by RØDE plug-in uses a new time-frequency adaptive approach for 
A to B-format conversion. This complex mathematical process means the phase 
between the A-format channels are aligned prior to application of the 
conversion matrix – essentially correcting for the non-coincidence of the 
capsules prior to any further processing."

How might they phase/time align the capsules ??

This must indeed be highly complex, as it is frequency dependent (low 
frequencies have smaller phase differences than high frequencies) as well as 
source directionally (across multiple blind sources) dependent.

Ciao,

Dave Hunt



On 16 Dec 2018, at 17:00, 
sursound-requ...@music.vt.edu wrote:

Send Sursound mailing list submissions to
sursound@music.vt.edu

To subscribe or unsubscribe via the World Wide Web, visit
https://mail.music.vt.edu/mailman/listinfo/sursound
or, via email, send a message with subject or body 'help' to
sursound-requ...@music.vt.edu

You can reach the person managing the list at
sursound-ow...@music.vt.edu

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Sursound digest..."
WHEN REPLYING EDIT THE SUBJECT LINE

ALSO EDIT THE MESSAGE BODY

You are receiving the digest so when replying, please remember to edit your 
Subject line to that of the original message you are replying to, so it is more 
specific than "Re: Contents of Sursound-list digest…" the subject should match 
the post you are replying to.

Also, please EDIT the quoted post so that it is not the entire digest, but just 
the post you are replying to - this will keep the archive useful and not 
polluted with extraneous posts.

This is the responsibility of digest subscribers. the community and list 
subscribers care about the integrity of the threads and archives so this is 
important.Today's Topics:

 1. Re: Soundfield by Rode plugin (Gary Gallagher)

From: Gary Gallagher 
Subject: Re: [Sursound] Soundfield by Rode plugin
Date: 16 December 2018 04:36:11 GMT
To: Surround Sound discussion group 


Thanks for that reference. I guess we'll just have to wait for more
information to filter out.

On Sun, Dec 16, 2018, 01:36 Paul Hodges  wrote:

There is slightly more description of their A to B-format processing
(but not much) in Rode's blog:



Paul

--
Paul Hodges

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
edit account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 



___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound

WHEN REPLYING EDIT THE SUBJECT LINE

ALSO EDIT THE MESSAGE BODY

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Looking for JavaScript compatible AmbiX / FuMa player

[Politis Archontis]

Hi Matt,

We have the open-source JSAmbisonics library that supports first and 
higher-order Ambisonics, Fuma or ambiX, and you can find in the code examples 
one that polls your phone’s orientation and uses it to rotate 360 video and the 
ambisonic soundfield.

It may be what you are looking for - you can find it at 
https://github.com/polarch/JSAmbisonics, with online examples at
https://cdn.rawgit.com/polarch/JSAmbisonics/e28e15b384f2442a66fadc0035439c64ed65fa4d/index.html.
 You can also import it from NPM to your project.

Best regards,
Archontis Politis

On 16 Nov 2018, at 22:36, Matt Soson 
mailto:mattso...@gmail.com>> wrote:

Hi Sursound community,

I'm making an app in React Native (uses javascript) that will need to play
AmbiX / FuMa spatial audio, and am searching for a player that works in
javascript for the AmbiX/FuMa (or other ambisonic format if necessary)
files, and is set up to talk to the phone's sensors for directional
orientation among the spatial sound.  Any help or being pointed in a
direction hugely appreciated!  Can be contacted at 
mat...@crooked-grin.com
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] RIR measuring, how to capture a higher order Ambisonic room responce?

[Politis Archontis]

By the way one could actually use only one microphone, measure the RIR, rotate 
it to another point, repeat the measurement, and make a virtual array of 
hundreds or thousands of points for high-order RIR recording. This has actually 
been done and the work published by Boaz Rafaely and his research group.

Regards,
Archontis Politis



> On 23 Apr 2018, at 20:31, Bo-Erik Sandholm  wrote:
> 
> As I see it to capture the signals for the upper layers of octomic with a
> tetra mic you rotate the mic 90 degrees between the takes..
> To capture the lower octomic  elements layer go back to initial position,
> then rotate Tetra mic 45 degrees and then 90 degrees for 2  recordings.
> 
> So 4 rotation direction to place the tetra mic elements in same positions
> as the 8 Octomic capsules.
> 
> select the 8 A signals that corresponds to the octomic positions...
> 
> Then get our hands on the octomic software if possible and hopefully
> translate the tetramic calibration file in to a octomic calibration file,
> might be possible.
> 
> This has only a chance to work for IR measurements and if the rotation of
> the tetramic is done without moving the center point of the mic head.
> 
> I hope this is possible, it should be a great new use of a tetramic to be
> able with a little work to create second order room Impulse responses.
> 
> Bo-Erik
> 
> 2018-04-23 19:10 GMT+02:00 Fernando Lopez-Lezcano 
> :
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] RIR measuring, how to capture a higher order Ambisonic room responce?

[Politis Archontis]

Hi Bo-Erik,

if you use Matlab or Octave, you could also try the SDM method to upsample from 
first-order RIRs to 2nd, 3rd, or any order you want basically. SDM stands for 
the Spatial Decomposition Method from my colleague Sakari Tervo, which has been 
used quite a lot for auralization and visualization of spatial room IRs. You 
can find the toolbox available online.

HARPEX could potentially do it too, but since it is made for 
reproduction/playback and most likely block processing, I don’t know if it 
would cope well with the fine temporal structure of the RIR.

Regards,
Archontis Politis



On 23 Apr 2018, at 14:25, jack reynolds 
mailto:jackreynolds...@gmail.com>> wrote:

or yes, take a first order B-Format IR and upsample using Harpex is another
possibility.

J

On 23 April 2018 at 12:24, jack reynolds 
mailto:jackreynolds...@gmail.com>> wrote:

The new coresound octomic is based on two tetramics, with one rotated 45
degrees from the other, so if you could work out how the second order
B-format is extracted from the octomic array, you could potentially take an
A-format reponse with your tetramic, rotate the mic 45 degrees and capture
another, then process all eight channels?
Just a thought.

Jack

On 23 April 2018 at 11:51, Bo-Erik Sandholm 
mailto:bosses...@gmail.com>> wrote:

Have never used Max, I need just 2, but probably 4 sound source positions
and one listening position.
But the listening position should have full spherical  ambisonic
soundfield.
But the result should be ambisonic IR's for these 4 sources.

The question is could upsampling be used ?

BR Bo-Erik



2018-04-23 10:47 GMT+02:00 Hyunkook Lee 
mailto:h@hud.ac.uk>>:

Indeed HIRT is the best IR capture package for Max.  There is also
HAART,
which is a standalone Max application we developed using HIRT. This
software is all in one box for multichannel IR capture (24 mics x 24
sources), acoustic parameter analysis and binauralisation. The analysis
part is still under development, but the IR capture and binauralisation
parts are fully working. You can download it here

http://eprints.hud.ac.uk/id/eprint/24579/

Also as Pierre mentioned, we captured over 2000 IRs of 39 multichannel
mic
array configurations from stereo to 9ch 3D using HAART. The library
comes
with a Max renderer where you can convolve dry sources or signals fed
from
DAW with the mic array IRs for simultaneous comparisons between
techniques.

https://github.com/APL-Huddersfield/MAIR-Library-and-Renderer

Best,
Hyunkook
=
Dr Hyunkook Lee, BMus(Tonmeister), PhD, MAES, FHEA
Senior Lecturer in Music Technology
Leader of the Applied Psychoacoustics Laboratory (APL)
http://www.hud.ac.uk/apl
http://www.hyunkooklee.com
Phone: +44 (0)1484 471893
Email: h@hud.ac.uk
Office: CE 2 /14a
School of Computing and Engineering
University of Huddersfield
Huddersfield
HD1 3DH
United Kingdom


From: Sursound [sursound-boun...@music.vt.edu] on behalf of Pierre
Alexandre Tremblay [tremb...@gmail.com]
Sent: 23 April 2018 09:29
To: Surround Sound discussion group
Subject: Re: [Sursound] RIR measuring, how to capture a higher order
Ambisonic room responce?

If you use Max, try the very versatile HIRT.

2nd order with a tetramic is not possible as far as I am aware through…
we
have done (mega)multimic IRs (24 channels of inputs, of which a 1st
order
ambisonic) of 3 different spaces with our kit, and it was fun and
productive to train the ear on difference of multichannel mic techniques
(Hyunkook Lee has a cool setup and papers on them, and I was mostly
interested in DPA LCR omni vs coincident vs MS)

We did many stage positions too. I can investigate if I can share the
files if that interests anyone.

p


On 23 Apr 2018, at 08:37, Bo-Erik Sandholm 
mailto:bosses...@gmail.com>>
wrote:

I want to measure the RIR of a medium size good listening room at
least
up
to second order Ambisonic RIR.

The IR result is to incorporate the responce and reflections of the
speakers and their positions.

I have a tetramic.

Can several measurements and rotation of the tetra mic between them be
combined to create the measurements that comes closer to a second
order
mic?

I know the basics of using Audacity, and a audio sweep and creating a
IR
from this.



Bo-Erik Sandholm
Stockholm
-- next part --
An HTML attachment was scrubbed...
URL: 
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe
here,
edit account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
edit account or options, view archives and s

Re: [Sursound] Massless speakers

[Politis Archontis]

Hi,

here in Aalto university there has been quite a lot of research using lasers to 
generate precise omnidirectional impulses at various focusing points for 
impulse response measurements of enclosures.

I also remember during a lecture in ISVR, UK, the lecturer mentioning these old 
plasma tweeters, saying they did not get very far cause they were ionizing the 
air producing hazardous levels of ozone :-D .

Archontis Politis



On 25 Mar 2018, at 14:34, David Pickett 
mailto:d...@fugato.com>> wrote:

Are these not the same as the Ionophone speakers, sold as Ionofane tweeters in 
the 1960s? They were flat from 10kHz to oo!

David

At 12:41 25-03-18, Bearcat Åzándor wrote:
>
>Accapella Audio Arts has been making horns with plasma tweeters for a while
>now. They are some of the largest, and most expensive speakers made.
>Awesome in thier rediculousness, i'd be afraid that someone would get drunk
>at my home and mistake them for a sink or a urinal.
>
>http://www.acapella.de/en/
>
>On Sun, Mar 25, 2018 at 1:23 AM, Augustine Leudar 
>mailto:augustineleu...@gmail.com>
>> wrote:
>
>> Ive been wondering for a while whether it was possible to use
>> electromagnetic waves interferance patterns to generate sound in 3D space
>> - the conclusions my physisit friends had was yes - possible - but not
>> without killing anyone that heard it .Plamsa is a bit different of course
>> to an EM wave
>
>> On 25 March 2018 at 07:45, Gary Gallagher 
>> mailto:g.null.dev...@gmail.com>> wrote:
>
>> > Never heard of this. Using laser plasma as driver for a speaker. Here as
>> > weapon - but it would be interesting to hear/see a  refined version of
>> > this. I'm curious about the spinning disk set up at the end of the video
>> > there appears to be some sort of modulation in the sound.
>> >
>> > US MILITARY DEVELOPING LASER PLASMA SPEAKERS
>> > >  > 
>> > laser-plasma-speakers/>
>> >
>> > Gary
>> > -- next part --
>> > An HTML attachment was scrubbed...
>> > URL: >  > 
>> > attachments/20180325/2167d7e3/attachment.html>
>> > ___
>> > Sursound mailing list
>> > Sursound@music.vt.edu
>> > https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
>> > edit account or options, view archives and so on.
>> >
>
>
>
>> --
>> Dr. Augustine Leudar
>> Artistic Director Magik Door LTD
>> Company Number : NI635217
>> Registered 63 Ballycoan rd,
>> Belfast BT88LL
>> www.magikdoor.net
>> -- next part --
>> An HTML attachment was scrubbed...
>> URL: > 
>> attachments/20180325/df4cbd39/attachment.html>
>> ___
>> Sursound mailing list
>> Sursound@music.vt.edu
>> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
>> edit account or options, view archives and so on.
>
>
>
>
>--
>Bearcat M. Åzándor
>Feline Soul Systems LLC
>Voice: 872.CAT.SOUL (872.228.7685)
>Fax: 406.235.7070
>-- next part --
>An HTML attachment was scrubbed...
>URL:
>25/e64fcf3d/attachment.html>
>___
>Sursound mailing list
>Sursound@music.vt.edu
>https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe
>here, edit account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Mosca: GUI assisted ambisonics quark v0.2 for SuperCollider

[Politis Archontis]

Hi Bo-Erik,

Just a small correction :-) - this is Archontis from Ville Pulkki’s group here 
in Aalto University, I think you meant Antti Vanne from IDA for personalized 
HRTFs (unless there is another spatial audio Pulkki here in Finland with the 
name Antti, offering also personalized HRTFs :-).

BR,

Archontis Politis
Post-doctoral Researcher
Department of Signal Processing and Acoustics
Aalto University
Finland

On 14 Feb 2018, at 20:43, Bo-Erik Sandholm 
mailto:bosses...@gmail.com>> wrote:

I currently have a personal SOFA file created for me by Antti Pulkki's
company.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Individualized HRTFs (Augustine Leudar)

[Politis Archontis]

I had my head & shoulders scanned in the past and my HRTFs produced by IDA have 
worked very well for me.
Unless they have changed their approach, and as far as I know, they do not 
measure features or try to match anthropometric dimensions to HRTFs in a 
database (which is still problematic since matching of specific anthropometric 
features to the fine structure of the HRTF is still an open research question), 
instead they do a full computational acoustic simulation of the HRTFs that 
takes into account the whole scan as it is.

Regards,
Archontis Politis

> On 23 Nov 2017, at 12:00, Augustine Leudar  wrote:
> 
> Does it measure ITDs and ILDs as well or just the Pinna ? I would like to
> try it anyway - but retain slight skepticism for now..
> 
> On 23 November 2017 at 09:50, Politis Archontis 
> wrote:
> 
>> Hi Gernot,
>> 
>> I haven’t used the toolkit, but I don’t see the reason for HRTF
>> interpolation in a predefined grid - if you generate the spherical harmonic
>> filters for whatever HRTF set you have without interpolation, then all you
>> need in AmbiX is a binaural decoding configuration file with a (N+1)^2 x
>> (N+1)^2 decoding matrix that is a unity matrix (zeros everywhere and ones
>> in the diagonal), along with the (N+1)^2 stereo WAV files of the filters
>> you generated (where N is the order you are going for).
>> 
>> Regards,
>> Archontis Politis
>> 
>> 
>> On 23 Nov 2017, at 01:13, Gernot von Schultzendorff <
>> g.schultzendo...@saalakustik.de<mailto:g.schultzendo...@saalakustik.de>>
>> wrote:
>> 
>> Looking through IDA informations there seem to be two different ways to
>> perform the individual scans:
>> - "Using their smartphone’s camera, users will be able to scan themselves,
>> gathering sufficient data for IDA Audio and Genelec to accurately 3D model
>> and then create the unique HRTF filter set for personal rendering of 3D
>> audio." [https://www.genelec.com/genelec-and-ida-audio-
>> redefine-immersive-3d-audio-professional-headphone-users <
>> https://www.genelec.com/genelec-and-ida-audio-redefine-immersive-3d-audio-
>> professional-headphone-users>]
>> - "An IDA 3D scan can be performed at a licensed IDA ScanStation on a
>> variety of scanners, including structured light, laser, and photogrammetry
>> scanners." [http://idaaudio.com/professionals/ <http://idaaudio.com/
>> professionals/>]
>> 
>> Regarding the import of SOFA files into the ambiX binaural player the
>> Princeton University 3-D Audio & Applied Acoustics Lab has developed the
>> SABRE toolkit specifically for this task (https://www.princeton.edu/
>> 3D3A/SABREToolkit.html <https://www.princeton.edu/3D3A/SABREToolkit.html>).
>> Requires MATLAB knowledge and decisions about interpolation "when measured
>> HRTFs are not available at the desired grid positions". Anyone having
>> experiences with this toolkit?
>> 
>> --
>> Gernot
>> 
>> 
>> 
>> On Sun, 5 Nov 2017 09:59:36 +
>> Augustine Leudar mailto:augustineleudar@gmail.
>> com> <mailto:augustineleu...@gmail.com>> wrote:
>> 
>> Let us know how you get on with height  I also wonder how much ground
>> and shoulder reflections will count and if subjective visual/cognitive
>> assessments of environment affect localisation in a significant way -
>> interesting to see. If anthropometrics alone are enough.
>> 
>> On Sunday, 5 November 2017, Marc Lavall?e mailto:marc
>> @hacklava.net> <mailto:m...@hacklava.net>> wrote:
>> 
>> 
>> I had mine done last week, after a few attempts at shooting a suitable
>> video file (it looks easy to do, but it's not). The resulting 3D model
>> is also not as good as the ones shown in the promotional video, so I
>> will probably try again because I suspect that a better video would
>> help to produce a better 3D model, therefore a better HRTF profile.
>> 
>> I'm still in the process of evaluating the SOFA file I received (in my
>> free time, something I should continue tomorrow). So far the results
>> are very interesting; it does work. I will try to compare it with some
>> generic HRTFs, and I will report here.
>> 
>> One of my problem is the lack of software to use a SOFA file. My
>> evaluation tool is now JSAmbisonics, because it's free (as in speech).
>> I guess I could also use the ambiX binaural player, but that would be
>> more work.
>> 
>> --
>> Marc
>> 
>> On Sun, 5 Nov 2017 00:07:59 +
>> Augustine Leudar 

Re: [Sursound] Individualized HRTFs (Augustine Leudar)

[Politis Archontis]

Hi Gernot,

I haven’t used the toolkit, but I don’t see the reason for HRTF interpolation 
in a predefined grid - if you generate the spherical harmonic filters for 
whatever HRTF set you have without interpolation, then all you need in AmbiX is 
a binaural decoding configuration file with a (N+1)^2 x (N+1)^2 decoding matrix 
that is a unity matrix (zeros everywhere and ones in the diagonal), along with 
the (N+1)^2 stereo WAV files of the filters you generated (where N is the order 
you are going for).

Regards,
Archontis Politis


On 23 Nov 2017, at 01:13, Gernot von Schultzendorff 
mailto:g.schultzendo...@saalakustik.de>> wrote:

Looking through IDA informations there seem to be two different ways to perform 
the individual scans:
- "Using their smartphone’s camera, users will be able to scan themselves, 
gathering sufficient data for IDA Audio and Genelec to accurately 3D model and 
then create the unique HRTF filter set for personal rendering of 3D audio." 
[https://www.genelec.com/genelec-and-ida-audio-redefine-immersive-3d-audio-professional-headphone-users
 
]
- "An IDA 3D scan can be performed at a licensed IDA ScanStation on a variety 
of scanners, including structured light, laser, and photogrammetry scanners." 
[http://idaaudio.com/professionals/ ]

Regarding the import of SOFA files into the ambiX binaural player the Princeton 
University 3-D Audio & Applied Acoustics Lab has developed the SABRE toolkit 
specifically for this task (https://www.princeton.edu/3D3A/SABREToolkit.html 
). Requires MATLAB knowledge 
and decisions about interpolation "when measured HRTFs are not available at the 
desired grid positions". Anyone having experiences with this toolkit?

--
Gernot



On Sun, 5 Nov 2017 09:59:36 +
Augustine Leudar mailto:augustineleu...@gmail.com> 
> wrote:

Let us know how you get on with height  I also wonder how much ground
and shoulder reflections will count and if subjective visual/cognitive
assessments of environment affect localisation in a significant way -
interesting to see. If anthropometrics alone are enough.

On Sunday, 5 November 2017, Marc Lavall?e 
mailto:m...@hacklava.net> > wrote:


I had mine done last week, after a few attempts at shooting a suitable
video file (it looks easy to do, but it's not). The resulting 3D model
is also not as good as the ones shown in the promotional video, so I
will probably try again because I suspect that a better video would
help to produce a better 3D model, therefore a better HRTF profile.

I'm still in the process of evaluating the SOFA file I received (in my
free time, something I should continue tomorrow). So far the results
are very interesting; it does work. I will try to compare it with some
generic HRTFs, and I will report here.

One of my problem is the lack of software to use a SOFA file. My
evaluation tool is now JSAmbisonics, because it's free (as in speech).
I guess I could also use the ambiX binaural player, but that would be
more work.

--
Marc

On Sun, 5 Nov 2017 00:07:59 +
Augustine Leudar mailto:augustineleu...@gmail.com> 
 > wrote:

Looks interesting - anthropometric approach.

On Saturday, 4 November 2017, Len Moskowitz
mailto:lenmoskow...@optonline.net> 
 > wrote:

Augustine Leudar wrote:


great sound design but no cigar on the binaural front - thats
really not
going to get solved until a quick and convenient way of measuring
HRTFs is designed - I have several in case any millionaires out
there are interested



At the recent AES Exhibition in NYC, IDA was shown:


   idaaudio.com 
   https://www.youtube.com/watch?v=pxf28tinxZg 



Individualized HRTFs, using a smartphone, for under $500.


Someone I trust had his done, and he says it finally allowed him to
disambiguate between front and rear.





Len Moskowitz (mosko...@core-sound.com 
 )
Core Sound LLC
www.core-sound.com 
Home of TetraMic and OctoMic

___
Sursound mailing list
Sursound@music.vt.edu 
 
https://mail.music.vt.edu/mailman/listinfo/sursound 
 - unsubscribe
here, edit account or options, view archives and so on.




___
Sursound mailing list
Sursound@music.vt.edu 
 
https://mail.music.vt.edu/mailman/listinfo/sursound 

Re: [Sursound] Virtual Microphone Processors for HOA?

[Politis Archontis]

Hi Stefan,

On 12 Oct 2017, at 02:50, Stefan Schreiber 
mailto:st...@mail.telepac.pt>> wrote:

Omnitone offers ambisonic decoding and binaural rendering of:

   * First-order-ambisonic stream
   * High-order-ambisonic stream: 2nd and 3rd order.

The input audio stream can be either an HTMLMediaElement ( or  
tag) or a multichannel AudioBufferSourceNode.

I may be wrong, these things change quite quickly, but the HTMLMediaElement, 
which could be used for live streaming does not support more than 8ch at the 
moment, hence it would work with FOA only, while the AudioBufferSourceNode can 
handle up to 32 (hence 4th-order) but works only with loading audiofiles.

BR,
Archontis Politis





-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Virtual Microphone Processors for HOA?

[Politis Archontis]

Hi Len,

you can find a higher-order virtual microphone implementation in our web-audio 
ambisonics library, implementing higher-order cardioids, supercardioids, and 
hypercardioid patterns. The library is open-source and generates the patterns 
for any order, even though Web Audio supports up to 4th-order signals (without 
tricks) at that point. You can find a real-time 3rd-order demo here:

https://cdn.rawgit.com/polarch/JSAmbisonics/e28e15b384f2442a66fadc0035439c64ed65fa4d/examples/hoa-virtual-mic.html

If you manage to stream your 9 second-order signals somehow in the browser, 
then you can process them live there, but unfortunately I think HTML5 supports 
up to 8 channels for streaming audio, so at the moment HOA material works only 
with wav/ogg files.

Regards,
Archontis Politis
Post-doctoral Researcher
Department of Signal Processing and Acoustics
Aalto University
Finland


On 11 Oct 2017, at 19:43, len moskowitz 
mailto:lenmoskow...@optonline.net>> wrote:

We're introducing OctoMic next week at AES (booth 315). It's a 2nd-order 
ambisonic microphone.

Does anyone else have a virtual microphone processor for 2nd-order (or higher) 
B-format?



Len Moskowitz (mosko...@core-sound.com)
Core Sound LLC
www.core-sound.com
Home of TetraMic
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Songbird

[Politis Archontis]

Hi Martin,

Interesting! Taking a closer look it seems like it’s an intermediate library to 
ease integrating ambisonic encoding to VR, or game scenes, and applying also 
some environmental effects (reverberation). The actual ambisonic processing is 
still handled by their Omnitone library.

Taking a look at the source code, Songbird has up to 3rd-order processing. 
Curiously, the Omnitone main branch is still at first-order processing only, 
but there is a development HOA branch, so I guess it is almost ready.
Some of the methods that we have used in our JSambisonics web audio library 
have also been incorporated, e.g. they are using the Ivanic and Ruedenberg 
rotation method for real spherical harmonics, which we have also found to be 
the fastest.

That’s a great development I believe, cause the FOA Omnitone binaural decoder, 
even though fast, was quite “blurry" spatially.

Regards,
Archontis Politis


On 01 Sep 2017, at 19:12, Martin Leese 
mailto:martin.le...@stanfordalumni.org>> wrote:

Hi,

Google's Songbird is a JavaScript library that
can render HOA to stereo and the Web; visit:

https://opensource.googleblog.com/2017/08/bringing-real-time-spatial-audio-to-web-with-songbird.html

https://google.github.io/songbird/

Regards,
Martin
--
Martin J Leese
E-mail: martin.leese  stanfordalumni.org
Web: http://members.tripod.com/martin_leese/
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] multichannel VST recorder os x

[Politis Archontis]

Hi Oliver,

A bit curious here, I’m not aware of any such VST, but isn’t the multichannel 
recording usually a task for the host DAW or sample editor etc. ? For example 
Reaper does super easy multichannel recording on OSX for up to 64ch on a single 
bus, if I remember correctly.

Regards,
Archontis





On 21 Jun 2017, at 00:14, Oliver Larkin 
mailto:olilar...@googlemail.com>> wrote:

hello,

does anyone know of a VST plug-in that will record a valid 16 channel wav file 
on os x? would rather not join mono files manually

thanks,

oli
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Sonicam

[Politis Archontis]

Hi Steven,

(hurried reply on the go) the numbers were for 17cm diameter (8.5cm radius). 
You also get around 15-18dB SNR improvement just by combining that many 
microphones, that’s still included in that simulation.

I wasn’t saying that the device is not good or useful, apart from the video, 
you can do a lot of useful stuff with that many microphones, binaural 
rendering, beamforming etc.. I was saying that it’s not a very good HOA 
microphone with these specs.

Best,
Archontis Politis






On 30 May 2017, at 16:26, Steven Boardman 
mailto:boardroomout...@gmail.com>> wrote:

Hi Archontis

The size is 170mm diameter not radius.
They also state the capsules are in arrays of 16, so 4 arrays of 16 capsules. 
Looking at the camera though, I can’t work it out….

The also said:

'The noise and frequency respond i quote is for individual mic. We have noise 
reduction algorithm to improve 15db in total compared to each mic.
For frequency respond, we are using the EQL process，reach to the “flat".'


I knew that it wouldn’t be well behaved in the entire range, and as a 
consequence could not be relied on for a quality full bandwidth recording.
This isn’t always the point though. As it could still be used as a reliable 
reference for the most crucial of frequency ranges, the voice.
As long as it can record dialogue, one can reconstruct with ADR. It’s just 
wether it is intelligible, and accurate enough spatially.
90% of audio for films is done in post anyway, especially dialogue.

It includes 6 camera’s (and the footage isn’t too bad), so to me it seems like 
a bargain. I just need to hear the audio first. Otherwise it maybe better just 
to spend it all on a camera.
What has been mentioned by Fons, and yourself is obviously completely true, but 
we all know in practice it is probably worse. Even more reason to hear their 
3rd order render. Of course 5th is a dream…..

While we are on the maths;
For 3rd order, and with the minimum amount of capsules, how small would the 
sphere need to be, for a range up to say 16kHz, or maybe even 18kHz?

Am I correct in assuming that the distance between the capsules needs to be 
less than half the highest wavelength represented?

Best

Steve




On 27 May 2017, at 23:18, Politis Archontis 
mailto:archontis.poli...@aalto.fi>> wrote:

Hi Steven,

as Fons mentioned before such an array will have trouble to deliver proper HOA 
components at a useful range. Running a simple simulation here for R=17cm and 
assuming perfect matched microphones and perfectly (uniformly) arranged, 
spatial aliasing starts to become serious at 4.5kHz. Assuming that the encoding 
filters allow some noise amplification, eg 15dB, to get some more usable low 
frequencies:
2nd order is well behaved at 200 Hz - 5.5 kHz
3rd order at 600 Hz - 5kHz
4th order at 1   kHz - 4.5 kHz
5th order at 1.5kHz - 4kHz

Since the microphones will have some unidealities, and their placement is not 
uniform as far as I can see, the actual performance can be worse than that, 
depending how much the encoding filters are optimized for the setup or not. And 
again that’s assuming 15dB of more noise (frequency-dependent) at the HOA 
signals than the microphones, which may be a problem at many recordings. Less 
permitted noise on the other hand means that all the low limits for all orders 
go up, hence, even smaller usable ranges.

If one wishes to capture spatial sound in a HOA format, instead of having a 
device that tries to do everything, it would make more sense to have a separate 
camera, and a dedicated HOA microphone since it requires careful optimization 
for that purpose.

Regards,
Archontis Politis



-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20170530/43b87df5/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu<mailto:Sursound@music.vt.edu>
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20170530/b80de9d5/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Sonicam

[Politis Archontis]

Hi Steven,

as Fons mentioned before such an array will have trouble to deliver proper HOA 
components at a useful range. Running a simple simulation here for R=17cm and 
assuming perfect matched microphones and perfectly (uniformly) arranged, 
spatial aliasing starts to become serious at 4.5kHz. Assuming that the encoding 
filters allow some noise amplification, eg 15dB, to get some more usable low 
frequencies:
2nd order is well behaved at200 Hz - 5.5 kHz
3rd order at600 Hz - 5kHz
4th order at1   kHz - 4.5 kHz
5th order at1.5kHz - 4kHz
 
Since the microphones will have some unidealities, and their placement is not 
uniform as far as I can see, the actual performance can be worse than that, 
depending how much the encoding filters are optimized for the setup or not. And 
again that’s assuming 15dB of more noise (frequency-dependent) at the HOA 
signals than the microphones, which may be a problem at many recordings. Less 
permitted noise on the other hand means that all the low limits for all orders 
go up, hence, even smaller usable ranges.

If one wishes to capture spatial sound in a HOA format, instead of having a 
device that tries to do everything, it would make more sense to have a separate 
camera, and a dedicated HOA microphone since it requires careful optimization 
for that purpose.

Regards,
Archontis Politis


> On 27 May 2017, at 23:47, Steven Boardman  wrote:
> 
> Surprise suprise, they can't provide a 3rd order sample at this point in
> time
> 
> Here is the message.
> 
> "We talked with our R&D team regarding to the schedule of audio. As our
> timeline is to achieve the 3rd order Ambisonic in first shipment in
> October, from now to October we still have a lot of work to do. The 3rd
> order Ambisonic is not ready yet.
> 
> Right now our priority is the KickStarter, frankly speaking, we do not have
> more personnel to optimize the recording. However, we have officially put
> forward the needs of the 3rd order Ambisonic, and R&D team will prioritize
> it.
> 
> We will try our best to give you a demo before the KS ends. However, if we
> can't do it, we will update you as soon as it comes out."
> 
> On 26 May 2017 20:15, "Fons Adriaensen"  wrote:
> 
>> On Fri, May 26, 2017 at 04:29:02PM +0100, Steven Boardman wrote:
>> 
>>> ‘Yes, 3rd Ambisonic will be in the first shipment, but 5th order
>>> Ambisonic rendering will be provided in the early next year, by
>>> upgrading the software.’
>> 
>> Anyone promising usable 5th order from such a system is either
>> ignorant or deliberately misleading potential customers. Typical
>> kickstarter vaporware.
>> 
>>> A sphere of 17cm can’t be that great at high freq?
>> 
>> Depends on the distance between the capsules. In this case the
>> diameter is twice that of the Eigenmic, and the number of capsules
>> is doubled, so polar patterns will start to break down at around
>> 6 to 7 kHz or so (assuming optimum capsules placement).
>> 
>>> Especially as the spec for each mems capsule is: noise level 67db,
>> 
>> Let's hope they mean a S/N ratio 67 dB, which is a self noise level
>> of 27 dB :-)
>> 
>> Ciao,
>> 
>> --
>> FA
>> 
>> A world of exhaustive, reliable metadata would be an utopia.
>> It's also a pipe-dream, founded on self-delusion, nerd hubris
>> and hysterically inflated market opportunities. (Cory Doctorow)
>> 
>> ___
>> Sursound mailing list
>> Sursound@music.vt.edu
>> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
>> edit account or options, view archives and so on.
>> 
> -- next part --
> An HTML attachment was scrubbed...
> URL: 
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

[Sursound] VBAP and Ambisonics [was: The BBC & Quadrophony in 1973]

[Politis Archontis]

Hi Sampo,

> On 09 Jan 2017, at 06:27, Sampo Syreeni  wrote:
> 
> The critique I'd have for such panning laws is that they don't really respect 
> the ambisonic/Gerzon theory, especially at the low frequencies. In essence, 
> they work, and necessarily would *have* to work in the high frequency, 
> (ambisonically speaking) high order,sparse array limit. Which is why they 
> mostly work for common music and speech signals.

I am a bit baffled by the idea that VBAP is not compatible with Ambisonics 
theory (?) Thinking in terms of velocity and energy vectors, as far as I 
understand, VBAP with the (classic) amplitude panning formulation has zero 
angular error for the (Makita) velocity vectors for all directions. If you take 
the energy formulation of VBAP for high frequencies (solving for energies 
instead of amplitudes) then it results in the maximum (Gerzon) energy vectors 
that the setup can achieve with zero directional error again. Of course at low 
frequencies you cannot achieve the “perfect” pressure reconstruction that a 
mode-matching decoder can achieve, but then you see what are the gains that 
such a decoder imposes on not ideal regular setups to realize that perfect 
reconstruction should be compromised anyway with some more practical solution.


> However, they fail to work general speaker arrays fully. Especially at the 
> lower frequencies. Ambisonically speaking, where we'd go with a holistic, 
> whole array, directionally averaged velocity decode.

Again I think it depends how you mean it - VBAP will just work for any speaker 
array with a performance limited by the setup in a quite intuitive 
understandable way (large spread for large triangle apertures, full 
concentration at a speaker direction, nothing for regions outside a partial 
setup etc..). Ambisonic decoding for any array is not designed as easily as 
computing VBAP gains, and it seems for irregular setups, one of the most 
straightforward and practical ways to do it is to combine the properties of 
VBAP and Ambisonic decoding (as the work of Zotter, Batke, and Epain have 
shown). Considering panning specifically, I think it depends on the application 
what works best, for VR or interactive-audio stuff for example, where normally 
sound objects would be rendered with maximum sharpness VBAP would work better. 
If however some and more even directional spreading is preferred, then 
ambisonic panning should be better, or some VBAP variant with spreading as has 
been presented by Ville and others.

So I find Augustine's comments reasonable on panning sounds, but not in 
general: VBAP vs Ambisonics.

> On 09 Jan 2017, at 12:33, Augustine Leudar  wrote:
> 
> Yes i just mean - when making a 3D sound installation you can use various
> types of panning round a sphere (or whatever of speaker array). You seemed
> to be saying ambisonics had a clear advantage over other types of panning
> for 3D audio - I was just wondering what you saw as ambisonics' advantages
> over VBAP. I've actually found Ambisonics to be worse compared to VBAP in
> many situations and better in others - but generally I use Vbap or Dbap .
> The only real advantage I can see of ambisonics is having one file that can
> be up or down mixed - but you can do that to a degree with Vbap files as
> well.

(What is a VBAP file?)

That’s if you have actually access to the sound objects with their parametric 
information, in which case sure you can pan them however you like, you can even 
switch between different panners on the fly and pick the one you prefer. 
However, the generality of Ambisonics becomes clear if you have real 
sound-scene recordings, or you don’t have access to the objects due to 
bandwidth limitations, and it makes sense to downmix them to a format that 
preserves their directional properties as good as possible. This last case 
becomes especially important if decoding of some HOA channels (or even FOA with 
parametric decoding) becomes perceptually indistinguishable with respect to 
spatializing many of sound objects separately..

Regards,
Archontis

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] [allowed] brahma

[Politis Archontis]

Hello and a happy new year to the list,

just to mention some more resources on this topic: part of the open-source 
Matlab libraries I made public last year deals with this matter: deriving the 
matrix of filters form the microphone to the B-format signals, FOA or HOA.

The filters can be computed from a description of the array (e.g. 6 
super-cardioids at some desired radius and with desired orientation, or 16 
omnis on a hard sphere), or from a set of anechoic measured responses of the 
array. The solution can be chosen from a few different published approaches, 
based either on a regularized LMS (similar to Farina’s solution) or a 
thresholding inversion approach.

There is not extensive documentation at the moment, one needs to study the 
functions a bit, but you can see the results in action for the Eigenmike in 
these demos, along with the code that produces them:

http://research.spa.aalto.fi/projects/spharrayproc-lib/spharrayproc.html#3

where the filters are computed and evaluated based on a “theoretical” perfect 
Eigenmike, and then based on actual Eigenmike measurements. The evaluation is 
based on spatial correlation of the filtered responses, showing how similar the 
directivity shape across frequency is to the ideal one, (e.g. how closely the X 
channel resembles the perfect dipole), and level difference between the actual 
and the ideal component, taken as the RMS across all directions (diffuse level).

A report on these may follow soon - in the meantime if somebody’s interested to 
use these, let me know for additional help.

Regards,
Archontis


On 06 Jan 2017, at 03:54, David McGriffy 
mailto:da...@mcgriffy.com>> wrote:

I recently developed a MatLab script to create calibrations based on
Angelo's method.  Set up equations for each measurement's expected results,
find the LMS solution in each frequency bin, then build the resulting
filter matrix.  A little engineering around the low and high frequencies
and it seems to work pretty well.

I'd also like to point out that VVEncode (alas not VVTetraVST or VVMic)
will process anyone's 4x4 matrix of filters if they are put in wav files
like Angelo and Brahma do.  I also have matrix calibration processing in my
command line tool and I've done it in pure MatLab.  Contact me if you want
to try this and have any questions.

I mention all this mostly because I'd be interesting in processing anyone's
custom measurements.  For me this is test data for my scripts.  So far I've
only run data out of one lab and with one type of
excitation/reference/etc.  I'd like to make sure I'm not dependent on those
things.  And I can return a working calibration, I hope.

David
VVAudio

On Thu, Jan 5, 2017 at 1:21 PM, Fernando Lopez-Lezcano <
na...@ccrma.stanford.edu> wrote:

On 01/05/2017 02:19 AM, umashankar manthravadi wrote:

Indeed. I have an (if memory serves) emm6 "reference microphone" which is
not very expensive but comes with its own calibration curve. Flatness is
important as that will define how true is the frequency response of the
calibration. For the excitation I use a single driver small speaker, so far
it has been fine. I record is our rather small concert hall (the Stage) and
I managed to get about 5 mSecs of useful data before the first reflection
arrives. That seems to be enough. Completely agree with sealing the bass
port if you have one like that, that's what Eric Benjamin recommended doing.

I'll probably be offline for a few days, sorry (vacation, no internet,
bliss)...
Good luck!
-- Fernando



Sent from Mail for
Windows 10

From: Bo-Erik Sandholm
Sent: Thursday, January 5, 2017 3:31 PM
To: sursound
Subject: Re: [Sursound] [allowed] brahma

To get a good calibration, I expect I will need at least one
known/reference source or sensor?

The reference for me will probably be my coresound Tetramic.

I will probably use Kef eggs as speakers as they are coaxial transducers.
Only at low frequency will The port be offerter.

Have I understood The basics?
I will probably measure outside a calibration Day orborrow a anechoic room
at swedish radio.

Bo-Erik

Den 5 jan. 2017 11:53 fm skrev "umashankar manthravadi" <
umasha...@hotmail.com>:

Dear fernando



I forwarded your link to marc Lavallée when you first posted them. My
intent is that we should have multiple compatible systems for
calibrating A
format mics



Sent from my Windows 10 phone



From: Fernando Lopez-Lezcano
Sent: 05 January 2017 00:28
To: Surround Sound discussion group;
glard...@iol.ie
Subject: Re: [Sursound] [allowed] brahma



On 01/02/2017 09:59 PM, Bo-Erik Sandholm wrote:

Good luck in your continued effekt.

By the way, i have accuired one of your Tetra mic 3d shapeway for 14 mm
capsules.
I have not yet 

[Sursound] Thesis announcement and public spatial audio Matlab code

[Politis Archontis]

Hi!

My PhD is coming to an end and my dissertation is finally public in the 
University’s website.
For those interested, the topic is parametric processing of recorded spatial 
sound, with connections to Directional Audio Coding (DirAC) and Ambisonics.
It is article-based, meaning that there is an overview of the field, theory, 
literature and latest developments (introduction), followed by a number of 
published conference and journal articles (contributions), and can be found at:

https://aaltodoc.aalto.fi/handle/123456789/22499

Related to this, I recently made public another bunch of code developed and 
used throughout the thesis, packaged as a Matlab library on spherical array 
processing. It is tangential to Ambisonics, but with some useful things such as 
deriving filters from microphones to B-format/ SH signals for arbitrary arrays, 
deriving coefficients for various types of first- and higher-order virtual 
microphones, and other stuff mostly related to parametric analysis of sound 
fields. It can be found at:

https://github.com/polarch/Spherical-Array-Processing

with a showcase here:

http://research.spa.aalto.fi/projects/spharrayproc-lib/spharrayproc.html

Comments or feedback always highly appreciated!

Best regards,
Archontis Politis

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: JSAmbisonics library update

[Politis Archontis]

Hi Martin, 

Yes you are right, we stuck with that for our examples cause they are fairly 
popular, and there are many tools that implement exporting to .ogg. Of course 
if somebody uses the library, they can go with whatever multichannel format 
they prefer. 

One has to be though aware of the limitations of each browser - I’ll try to 
include a table for the that on the library’s documentation. 
For example, all of them handle multichannel WAVE files, but not all can decode 
them  for high number of channels, that’s the reason of the helper functions, 
loading 8ch files that are then concatenated into larger HOA buffers.

Then .ogg for example is supported by the Firefox/Google alliance, but Chrome 
itself messes up the channel order when decoding an 8ch .ogg file with Web 
Audio. A channel remapper is implemented inside to fix that if Chrome is 
detected (Firefox does it fine). 

Then there’s AAC, pretty nice perceptual coding, but not free, and there’s 
still the problem that if you’re trying to do Web VR stuff, then cutting edge 
development happens mostly in Chromium, which doesn’t support it at all.

So it seems like most of them have their advantages and disadvantages, and one 
has to choose depending on what is their target.

Best,
Archontis

> On 07 Oct 2016, at 22:23, Martin Leese  
> wrote:
> 
> Politis Archontis wrote:
> ...
>> Safari and iOS are partially supported (no support for multichannel .ogg
>> files at the moment, but otherwise mostly functional)
> 
> Note that the file extension ".ogg" has been
> deprecated for all but Vorbis I files (and Vorbis
> has been superseded by Opus).  Visit:
> https://wiki.xiph.org/MIME_Types_and_File_Extensions
> 
> Regards,
> Martin
> -- 
> Martin J Leese
> E-mail: martin.leese  stanfordalumni.org
> Web: http://members.tripod.com/martin_leese/
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: JSAmbisonics library update

[Politis Archontis]

Hi Marc, 

Thanks for the notice, very nice framework. I’ll try to check it out and create 
some examples for the library in the near future when I find some time..
About iOS, the library itself  seems to be working actually quite ok, it’s that 
multichannel file support varies a lot from browser to browser, and our 
examples won’t all play on Safari at the moment due to that.

Best,
Archontis


> On 07 Oct 2016, at 17:38, Marc Lavallée  wrote:
> 
> Great!
> 
> About support for iOS, I would suggest to use Cordova with the Crosswalk 
> “webview”; 
> it is a simple method to create cross-platform applications based on the 
> Chromium engine, 
> that would be fully compatible with JSAmbisonics.
> —
> Marc
> 
> 
>> On Oct 7, 2016, at 10:27 AM, Politis Archontis  
>> wrote:
>> 
>> Hello,
>> 
>> for those who are interested in ambisonic processing on the web (outside of 
>> Facebook and Youtube 360 playback),
>> 
>> this is an update on the JSAmbisonics library of Web Audio objects for 
>> first- (FOA) and higher-order (HOA) processing:
>> 
>> https://github.com/polarch/JSAmbisonics
>> 
>> Compared to the first early summer release, the examples have been updated 
>> with better decoding filters, and some more functionality; you can check 
>> them on your browser (Chrome/Firefox) or mobile (Android/Chrome) here:
>> 
>> https://cdn.rawgit.com/polarch/JSAmbisonics/1ccae3a6f0a60a690f5eb4bb5bbb21b58a5d5993/index.html
>> 
>> There was also a recent presentation and publication on the library in the 
>> Interactive Audio Systems Symposium, York, UK. You can find a description of 
>> the internals of the library on that publication here:
>> 
>> https://www.researchgate.net/publication/308761825_JSAmbisonics_A_Web_Audio_library_for_interactive_spatial_sound_processing_on_the_web
>> 
>> For people interested to integrate spatial sound on their applications, it 
>> seems to me perfectly doable to do many of the apps that pop up recently 
>> with all the VR boom, directly on the browser and without getting tied to a 
>> certain platform. Examples can be HOA ambisonic players with head-tracking, 
>> simple HOA mixing tools and manipulations with a GUI etc, acoustic 
>> visualization tools etc..
>> In the online examples, the mobile-phone player one is a quick hack we 
>> cooked that tries to demonstrate that. It is intended for Android phones 
>> (maybe will work on iPhones too) that have a gyro, and renders a spherical 
>> video of a small part from a recording here at Helsinki concert hall, in 
>> split-screen, Google-cardboard style, with FOA playback, and rotation based 
>> on the mobile’s sensors. It has worked on most phones I tried it around ( if 
>> you see the video on the screen, you have to click anywhere to get it 
>> started ).
>> 
>> On new features, various conversion tools and ambisonic mirroring have been 
>> added, but probably the most interesting one is that we did some effort on 
>> generating ambisonic-binaural filters from HRTF files, in the SOFA format, 
>> directly on the browser for an arbitrary order.
>> So that people can select HRTFs from a database and get a personalized 
>> experience without having to derive the filters themselves. It is still WIP 
>> but it seems robust. The SOFA example demonstrates that with two HRTF sets.
>> 
>> Safari and iOS are partially supported (no support for multichannel .ogg 
>> files at the moment, but otherwise mostly functional)
>> 
>> Again, any comments or feedback mostly welcome!
>> 
>> Regards,
>> Archontis Politis
>> -- next part --
>> An HTML attachment was scrubbed...
>> URL: 
>> <https://mail.music.vt.edu/mailman/private/sursound/attachments/20161007/1c33aca4/attachment.html>
>> ___
>> Sursound mailing list
>> Sursound@music.vt.edu
>> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
>> account or options, view archives and so on.
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: JSAmbisonics library update

[Politis Archontis]

Demo effects always :-). 

If anyone tried the mobile demo, the link was broken for some Dropbox-related 
reason. 
Now all the demo files have been moved to the library’s Github repository, and 
it should open properly (that doesn’t mean it will work 100% but there is a 
chance at least :-).

Best,
Archontis

> On 07 Oct 2016, at 17:27, Politis Archontis  
> wrote:
> 
> In the online examples, the mobile-phone player one is a quick hack we cooked 
> that tries to demonstrate that. It is intended for Android phones (maybe will 
> work on iPhones too) that have a gyro, and renders a spherical video of a 
> small part from a recording here at Helsinki concert hall, in split-screen, 
> Google-cardboard style, with FOA playback, and rotation based on the mobile’s 
> sensors. It has worked on most phones I tried it around ( if you see the 
> video on the screen, you have to click anywhere to get it started ).

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

[Sursound] Ambisonics on the web pt.1: JSAmbisonics library update

[Politis Archontis]

Hello,

for those who are interested in ambisonic processing on the web (outside of 
Facebook and Youtube 360 playback),

this is an update on the JSAmbisonics library of Web Audio objects for first- 
(FOA) and higher-order (HOA) processing:

https://github.com/polarch/JSAmbisonics

Compared to the first early summer release, the examples have been updated with 
better decoding filters, and some more functionality; you can check them on 
your browser (Chrome/Firefox) or mobile (Android/Chrome) here:

https://cdn.rawgit.com/polarch/JSAmbisonics/1ccae3a6f0a60a690f5eb4bb5bbb21b58a5d5993/index.html

There was also a recent presentation and publication on the library in the 
Interactive Audio Systems Symposium, York, UK. You can find a description of 
the internals of the library on that publication here:

https://www.researchgate.net/publication/308761825_JSAmbisonics_A_Web_Audio_library_for_interactive_spatial_sound_processing_on_the_web

For people interested to integrate spatial sound on their applications, it 
seems to me perfectly doable to do many of the apps that pop up recently with 
all the VR boom, directly on the browser and without getting tied to a certain 
platform. Examples can be HOA ambisonic players with head-tracking, simple HOA 
mixing tools and manipulations with a GUI etc, acoustic visualization tools 
etc..
In the online examples, the mobile-phone player one is a quick hack we cooked 
that tries to demonstrate that. It is intended for Android phones (maybe will 
work on iPhones too) that have a gyro, and renders a spherical video of a small 
part from a recording here at Helsinki concert hall, in split-screen, 
Google-cardboard style, with FOA playback, and rotation based on the mobile’s 
sensors. It has worked on most phones I tried it around ( if you see the video 
on the screen, you have to click anywhere to get it started ).

On new features, various conversion tools and ambisonic mirroring have been 
added, but probably the most interesting one is that we did some effort on 
generating ambisonic-binaural filters from HRTF files, in the SOFA format, 
directly on the browser for an arbitrary order.
So that people can select HRTFs from a database and get a personalized 
experience without having to derive the filters themselves. It is still WIP but 
it seems robust. The SOFA example demonstrates that with two HRTF sets.

Safari and iOS are partially supported (no support for multichannel .ogg files 
at the moment, but otherwise mostly functional)

Again, any comments or feedback mostly welcome!

Regards,
Archontis Politis
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Blue Ripple Sound & SN3D?

[Politis Archontis]

+1 for (S)N3D / ACN. To us newbies, as Fons mentioned :-), FuMa makes sense 
only for 1st-order, and then it is trivial to convert to the above if needed.

Best,
Archontis

On Aug 10, 2016, at 1:53 PM, Richard Furse  wrote:

> Hi there!
> 
> 
> 
> Blue Ripple Sound are wondering about changing their TOA plugins from FuMa
> to SN3D (ACN), see http://www.blueripplesound.com/story/consultation-sn3d.
> Does anyone have strong feelings on this?
> 
> 
> 
> Opinions, rants and raves appreciated, but I'd like to avoid kicking off
> another format debate on this list - perhaps any interested folk could
> respond using the email link on the page?
> 
> 
> 
> Many thanks,
> 
> 
> 
> --Richard
> 
> -- next part --
> An HTML attachment was scrubbed...
> URL: 
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Merging impulse responces, is it workable?

[Politis Archontis]

Hi Bo-Erik,

you should first convolve each sound source sample with its respective B-format 
RIR, then sum the resulting B-format streams. Do not merge the RIRs before, it 
won't give you the result you are looking for.

What you can do however is, if you have an indication of mixing time of the 
room or by trial and error, split the B-format RIRs into early parts, and late 
reverberation. As the late reverb tail will most likely not vary much with 
position, you can keep only one set of them. Then you should still convolve 
each source with the respective early part, but for the late part you can sum 
the source signals first and convolve them with the single late reverb RIR, 
assuming that they're going through the same "diffuse" reverberation. You save 
some computing this way with shorter convolutions. If that is not the issue 
though, you can go with the full RIRs which will I'd assume will give you the 
give the most natural impression.

Best regards,
Archontis

On Aug 11, 2016, at 2:21 AM, Bo-Erik Sandholm 
 wrote:

> I want to introduce the impulse responce of a real room to a foa or toa
> file with panned mono sources in fixed positions.
> I have a few questions about how to create the best result in the simplest
> way.
> 
> If I have the recorded bformat impulse responces of several sources placed
> in the same positions where the panned sources will be located.
> 
> Questions on how to get most natural result:
> 
> Should each positioned sound source be convolved with positional specific
> mono room impulse responce or a positional specific bformat RIR?
> 
> Can all separately measured bformat positional RIRs be merged as one
> bformat RIR in any way, if so how?
> 
> Can the convolving be done in bformat after mixing to a single bformat file
> or should it be done for each before mixing to bformat?
> Should that be done with mono or bformat RIR.
> 
> How many convoling instances would be best to use in a setup?
> 
> Bo-Erik Sandholm
> Stockholm
> -- next part --
> An HTML attachment was scrubbed...
> URL: 
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Conversion from FOA to TOA ? How to and why or not?

[Politis Archontis]

Hi Stefan,

No it is not only you :-), I thought I was clear that these references are 
totally synthetic. We just try to make sure that they are reproducible, they do 
sound natural, they are physically-based and they do not rely on any perceptual 
spatialization methods themselves (ambisonic, panning or anything else). Then 
we consider that as a plausible reference. As I said we do not have access to 
an original soundfield, so synthetic is our next best call. 

This relates to the difficult question, I believe, of what is the best way to 
assess transparency in a reproduction method of spatial recordings (compared, 
for example, to transparency of spatial audio coding with playback of 5.0 
material and its spatially compressed version, which is a much easier task 
since there is a clear reference). For most cases transparency is not of 
interest, and an overall perceptual quality is more important. However we have 
done these comparisons in the way I described, published the results and 
somebody interested can extract their own conclusions. And if they're good for 
DirAC decoding, then maybe they're good for other decoding approaches.

Regards,
Archontis



> On 05 Jul 2016, at 21:23, Stefan Schreiber  wrote:
> 
> Politis Archontis wrote:
> 
>> 
>> We start by setting up a large dense 3D loudspeaker setup in a fully 
>> anechoic chamber (usually between 25~35 speakers at a distance of ~2.5m), so 
>> that there is no additional room effect at reproduction. Then we decide on 
>> the composition of the sound scene (e.g. band, speakers, environmental 
>> sources), their directions of arrival and the surrounding room 
>> specifications. We then generate room impulse responses (RIR) using a 
>> physical room simulator for the specified room and source positions. We end 
>> up with one RIR for each speaker and for each source in the scene. 
>> Convolving these with our tests signals and combining the results we end up 
>> with an auralization of the intended scene. This part uses no spatial sound 
>> method at all, no panning for example - if a reflection falls between 
>> loudspeakers it is quantized to the closest one. The final loudspeaker 
>> signals we consider as the reference case (after listening to it and 
>> checking if it sounds ok).
>> 
> 
> Is it only me to notice that these "original scenes" look highly synthetical?
> 
> Maybe good for DirAC encoding/decoding, but a natural recording this is not...
> 
> BR
> 
> Stefan
> 
> P.S.: (Richard Lee )
> 
>> Some good examples of 'natural' soundfield recordings with loadsa stuff
>> happening from all round are Paul Doombusch's Hampi, JH Roy's schoolyard &
>> John Leonard's Aran music.
>> 
> 
> --
> 
> 
>> Then we generate our recordings from that reference. either by encoding 
>> directly to ambisonic signals, by simulating a microphone array recording, 
>> or by putting a Soundfield or other microphone at the listening spot and 
>> re-recording the playback. These have been dependent on the study.
>> 
>> Finally the recordings are processed, and decoded back to the loudspeakers, 
>> usually to a subset of the full setup (e.g. horizontal, discrete surround, 
>> small 3D setup), or even to the full setup. That allows us to switch 
>> playback between the reference and the method.
>> 
>> The tests have been usually MUSHRA style, where the listeners are asked to 
>> judge perceived distance from the reference and various randomized playback 
>> methods (including a hidden reference and a low quality anchor, used to 
>> normalize the perceptual scale for each subject). The criteria are a 
>> combination of timbral distance/colouration, spatial distance, and artifacts 
>> if any.
>> 
>> I’ve left out various details from the above, but this is the general idea. 
>> Some publications that have used this approach are:
>> 
>> 
>> Vilkamo, J., Lokki, T., & Pulkki, V. (2009). Directional Audio Coding: 
>> Virtual Microphone-Based Synthesis and Subjective Evaluation. Journal of the 
>> Audio Engineering Society, 57(9), 709–724.
>> 
>> Politis, A., Vilkamo, J., & Pulkki, V. (2015). Sector-Based Parametric Sound 
>> Field Reproduction in the Spherical Harmonic Domain. IEEE Journal of 
>> Selected Topics in Signal Processing, 9(5), 852–866.
>> 
>> Politis, A., Laitinen, MV., Ahonen, A., Pulkki, V. (2015). Parametric 
>> Spatial Audio Processing of Spaced Microphone Array Recordings for 
>> Multichannel Reproduction. Journal of the Audio Engineering Society 63 (4), 
>>

Re: [Sursound] Conversion from FOA to TOA ? How to and why or not?

[Politis Archontis]

Hi Richard,

Let me clarify a bit what I mean by “closer to the original” and how the 
listening tests were made. First of all, this is not a claim on perceived 
quality, which would be a different test and with different results I believe. 
And I’m not saying that first-order material cannot sound good or natural 
without parametric decoding.

Comparing reproduction against a reference is problematic in recorded cases, 
cause we cannot teleport from, say, a concert hall directly to our listening 
room to switch and compare between the two. And our auditory memory is that of 
a goldfish.
The way we have constructed these tests is not ideal but I believe it’s 
adequate to give us some indication of perceived distance from a reference.

We start by setting up a large dense 3D loudspeaker setup in a fully anechoic 
chamber (usually between 25~35 speakers at a distance of ~2.5m), so that there 
is no additional room effect at reproduction. Then we decide on the composition 
of the sound scene (e.g. band, speakers, environmental sources), their 
directions of arrival and the surrounding room specifications. We then generate 
room impulse responses (RIR) using a physical room simulator for the specified 
room and source positions. We end up with one RIR for each speaker and for each 
source in the scene. Convolving these with our tests signals and combining the 
results we end up with an auralization of the intended scene. This part uses no 
spatial sound method at all, no panning for example - if a reflection falls 
between loudspeakers it is quantized to the closest one. The final loudspeaker 
signals we consider as the reference case (after listening to it and checking 
if it sounds ok).

Then we generate our recordings from that reference. either by encoding 
directly to ambisonic signals, by simulating a microphone array recording, or 
by putting a Soundfield or other microphone at the listening spot and 
re-recording the playback. These have been dependent on the study.

Finally the recordings are processed, and decoded back to the loudspeakers, 
usually to a subset of the full setup (e.g. horizontal, discrete surround, 
small 3D setup), or even to the full setup. That allows us to switch playback 
between the reference and the method.

The tests have been usually MUSHRA style, where the listeners are asked to 
judge perceived distance from the reference and various randomized playback 
methods (including a hidden reference and a low quality anchor, used to 
normalize the perceptual scale for each subject). The criteria are a 
combination of timbral distance/colouration, spatial distance, and artifacts if 
any.

I’ve left out various details from the above, but this is the general idea. 
Some publications that have used this approach are:


Vilkamo, J., Lokki, T., & Pulkki, V. (2009). Directional Audio Coding: Virtual 
Microphone-Based Synthesis and Subjective Evaluation. Journal of the Audio 
Engineering Society, 57(9), 709–724.

Politis, A., Vilkamo, J., & Pulkki, V. (2015). Sector-Based Parametric Sound 
Field Reproduction in the Spherical Harmonic Domain. IEEE Journal of Selected 
Topics in Signal Processing, 9(5), 852–866.

Politis, A., Laitinen, MV., Ahonen, A., Pulkki, V. (2015). Parametric Spatial 
Audio Processing of Spaced Microphone Array Recordings for Multichannel 
Reproduction. Journal of the Audio Engineering Society 63 (4), 216-227

Vilkamo, J., & Pulkki, V. (2014). Adaptive Optimization of Interchannel 
Coherence. Journal of the Audio Engineering Society, 62(12), 861–869.

Getting the listening test samples and generating recordings or virtual 
recordings from the references would be a lot of work for the time being.

What is easier and I can definitely do is process one or some of the recordings 
you mentioned for your speaker setup, and send you the results for   listening. 
There is no reference in this case, but you can compare against your preferred 
decoding method. And it would be interesting for me to hear you feedback too.

Best regards,
Archontis

On 05 Jul 2016, at 09:32, Richard Lee 
mailto:rica...@justnet.com.au>> wrote:

Can you give us more detail about these tests and perhaps put some of these
natural recordings on ambisonia.com?

The type of soundfield microphone used .. and particularly the accuracy of
its calibration ... makes a HUGE difference to the 'naturalness' of a
soundfield recording.

Some good examples of 'natural' soundfield recordings with loadsa stuff
happening from all round are Paul Doombusch's Hampi, JH Roy's schoolyard &
John Leonard's Aran music.  Musical examples include John Leonards Orfeo
Trio, Paul Hodges "It was a lover and his lass" and Aaron Heller's (AJH)
"Pulcinella".  The latter has individual soloists popping up in the
soundfield .. not pasted on, but in a very natural and delicious fashion
... as Stravinsky intended.

Also to my experience, and that doesn?t seem to be a very popular view
yet in ambisonic community, these

Re: [Sursound] Conversion from FOA to TOA ? How to and why or not?

[Politis Archontis]

> On 01 Jul 2016, at 18:50, Justin Bennett  wrote:
> 
> 
> that’s interesting to hear, Trond, I was also wondering about how upsampling 
> would affect the reproduction of field recordings.
> 
> best, Justin
> 

Hi Justin,

To my experience, parametric methods such as Harpex or DirAC deal greatly with 
field recordings, since there is always enough ‘activity’ and natural 
variability in the sound scene that is analyzed and reproduced reasonably by 
the method’s underlying model. This is in contrast for example to synthetic 
material, in which you can generate unnatural cases that can “confuse” the 
model, e.g. six anechoic saw-tooth waves coming from various angles 
simultaneously with the same fundamental frequency.

Also to my experience, and that doesn’t seem to be a very popular view yet in 
ambisonic community, these parametric methods do not only upsample or sharpen 
the image compared to direct first-order decoding, but they actually reproduce 
the natural recording in a way that is closer perceptually to how the original 
sounded, both spatially and in timbre. Or at least that’s what our listening 
tests have shown in a number of cases and recordings. And the directional 
sharpening is one effect, but also the higher spatial decorrelation that they 
achieve (or lower inter-aural coherence) in reverberant recordings is equally 
important.

By the way, I have always considered the term upsampling a bit inaccurate for 
this parametric FOA-to-HOA mapping., and has no relation to upsampling from a 
signal processing POV. Upmixing would be more appropriate, since this is what 
the methods are essentially doing internally, not dissimilar to the older 
parametric upmixing methods from, e.g., stereo to surround.

Regards,
Archontis

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA

[Politis Archontis]

Hi Justin,

Thanks for the feedback.

> Hi Archontis, I think your ears must be somewhat different to mine… 

Not a problem. You can check the IRCAM HRTF set, listen to one the examples 
from different subjects here:

http://recherche.ircam.fr/equipes/salles/listen/sounds.html

and find a subject number that sounds convincing to you. Then I can generate 
the decoding filters using that specific subject's HRTFs for you (or you can 
use the included Matlab scripts in the library to do it yourself, if you have 
access to Matlab and you are familiar with it). It’s in the plans to implement 
that directly in the browser too in the near future..

> Thanks for the work, I have the feeling this could be very useful to me in 
> the future.

Glad to hear that.

Regards,
Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

Hi, 

just a note that I replaced the binaural decoding filters in the examples with 
others that fix most of the decoding colouration issues that were very strong 
in the first version. 
I‘ve also added some Matlab routines that can show how these filters can be 
computes from an HRTF set, with two different approaches, a virtual decoding 
and a “direct” approach.

I’d like to get some feedback (does it sound better?) from people that are 
interested in this. (You can find and listen to the examples at the bottom of 
the webpage https://github.com/polarch/JSAmbisonics).

The library will soon be packaged as a more formal javascript library, with 
NodeJS - thanks to David Poirier-Quinot from IRCAM for the massive work!

Best regards,
Archontis

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

Hi Albert,

> 
> Most of the first order samples reproduce nothing above 10k and a lot of
> them sound severely saturated/clipped.


I re-normalized the decoding filters in the examples, could you confirm if that 
fixes the clipping issues you mentioned on your browser?

Best regards,
Archontis 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

(Sorry for the re-posting, sursound seems to destroy indentation from my mail 
manager, making the previous one hard to read..)

Hi Steven,

There seems to be a misunderstanding, your seem to address the considerations 
raised by Stefan here on the list , not me :-) …

I am quite a fun of the microphone and I have gotten excellent results 
rendering to large loudspeaker arrays (>25ch) using both ambisonic decoders and 
the parametric decoders that we are developing here in Aalto University.

>On 22 Jun 2016, at 00:31, Steven Backer 
>mailto:s...@mhacoustics.com>> wrote:

>Nice work!  It’s great someone is taking the time to put together the 
>infrastructure for web-based ambisonics.
>I enjoyed watching your Intensity Analyzer.
Thanks a lot! The intensity analyser it’s a bit incomplete at the moment, cause 
it’s showing the broadband intensity vector, so it’s hard to show multiple 
different sources from different directions. I’ll soon add a band-passed 
version.

>How did you encode the raw Eigenmike signals?
>Using some of our software (EigenUnits-Encoder or EigenStudio), or something 
>else?
All the examples use the Encoder plugin provided by you. We have created our 
own encoding filters too here, based on dense anechoic measurements of the EM, 
which get some decent 4th-order components with a range of 2kHz~7kHz, but they 
lack some bass on orders 2&3 compared to your Encoder (I guess you’re using a 
thresholding approach at LF?)

>In fact, we’ve found that some decoders, specifically for HOA, do present some 
>challenges that can degrade parts of the spatial image and spectral response.
>There will be a paper presented on this topic at the upcoming AES Sound Field 
>Control conference in Guildford.
This is an interesting point you raise cause ambisonic research itself has 
avoided it pretty much. I think because it breaks the useful separation between 
microphone encoding and loudspeaker decoding. Hence, decoding becomes 
frequency-dependent with respect to the array size, mics etc. We have used both 
an order-limited approach here (a n-th order decoding matrix per band that the 
EM delivers properly the n-th order signals), and a decoding filter matrix 
approach. But we have gotten good results one way or another so I know that EM 
can deliver :-).

>In the past, we had control of both ends of the signal chain (i.e. the 
>beamforming in EigenStudio), and we could “do the right thing”.
>Now without knowing a-priori what the decoder is doing it becomes more 
>difficult.  This could explain some of the effects discussed earlier in this 
>thread.
>We will soon be releasing some software updates that will hopefully address 
>some of these issues.  So perhaps reserve some judgement for a later date ;-)
That sounds great! (and no judgment from my side..)

>>An Eigenmike has some aliasing limit frequency
>You are correct that the Eigenbeams will spatially alias around ~8kHz.
>We do implement a workaround for traditional beamforming, but for ambisonics 
>applications we just let it alias.
>There is still usable signal up to (temporal) Nyquist and I’ve heard plenty of 
>material that sounds just fine up there.
Let aliasing above that HF is reasonable. Another approach that may improve 
slightly compared to that, and something we have used in the past for 
parametric decoding, is to point your alias-suppression HF-beamformers towards 
the speakers.

>>The eigenmike could be a nice tool for VR/AR recordings
>Absolutely!  We think so, too.
>We’ve actually been pretty busy recently making quite a few recordings (some 
>with video!), and have some great material we are going to share publicly very 
>soon.
Please do! The 3-4 recordings you have at the moment on the website do not do 
justice to the EM potential!

Best regards,
Archontis Politis
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

Hi Steven,

There seems to be a misunderstanding, your seem to address the considerations 
raised by Stefan here on the list , not me :-) …

I am quite a fun of the microphone and I have gotten excellent results 
rendering to large loudspeaker arrays (>25ch) using both ambisonic decoders and 
the parametric decoders that we are developing here in Aalto University.

On 22 Jun 2016, at 00:31, Steven Backer 
mailto:s...@mhacoustics.com>> wrote:

Nice work!  It’s great someone is taking the time to put together the 
infrastructure for web-based ambisonics.  I enjoyed watching your Intensity 
Analyzer.
Thanks a lot! The intensity analyser it’s a bit incomplete at the moment, cause 
it’s showing the broadband intensity vector, so it’s hard to show multiple 
different sources from different directions. I’ll soon add a band-passed 
version.

How did you encode the raw Eigenmike signals?  Using some of our software 
(EigenUnits-Encoder or EigenStudio), or something else?
All the examples use the Encoder plugin provided by you. We have created our 
own encoding filters too here, based on dense anechoic measurements of the EM, 
which get some decent 4th-order components with a range of 2kHz~7kHz, but they 
lack some bass on orders 2&3 compared to your Encoder (I guess you’re using a 
thresholding approach at LF?)

In fact, we’ve found that some decoders, specifically for HOA, do present some 
challenges that can degrade parts of the spatial image and spectral response.  
There will be a paper presented on this topic at the upcoming AES Sound Field 
Control conference in Guildford.
This is an interesting point you raise cause ambisonic research itself has 
avoided it pretty much. I think because it breaks the useful separation between 
microphone encoding and loudspeaker decoding. Hence, decoding becomes 
frequency-dependent with respect to the array size, mics etc. We have used both 
an order-limited approach here (a n-th order decoding matrix per band that the 
EM delivers properly the n-th order signals), and a decoding filter matrix 
approach. But we have gotten good results one way or another so I know that EM 
can deliver :-).

In the past, we had control of both ends of the signal chain (i.e. the 
beamforming in EigenStudio), and we could “do the right thing”.  Now without 
knowing a-priori what the decoder is doing it becomes more difficult.  This 
could explain some of the effects discussed earlier in this thread.  We will 
soon be releasing some software updates that will hopefully address some of 
these issues.  So perhaps reserve some judgement for a later date ;-)
That sounds great! (and no judgment from my side..)

An Eigenmike has some aliasing limit frequency
You are correct that the Eigenbeams will spatially alias around ~8kHz.  We do 
implement a workaround for traditional beamforming, but for ambisonics 
applications we just let it alias.  There is still usable signal up to 
(temporal) Nyquist and I’ve heard plenty of material that sounds just fine up 
there.
Let aliasing above that HF is reasonable. Another approach that may improve 
slightly compared to that, and something we have used in the past for 
parametric decoding, is to point your alias-suppression HF-beamformers towards 
the speakers.

The eigenmike could be a nice tool for VR/AR recordings
Absolutely!  We think so, too.  We’ve actually been pretty busy recently making 
quite a few recordings (some with video!), and have some great material we are 
going to share publicly very soon.
Please do! The 3-4 recordings you have at the moment on the website do not do 
justice to the EM potential!

Best regards,
Archontis Politis

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

Hi Michael,

On 21 Jun 2016, at 10:35, Michael Chapman 
mailto:s...@mchapman.com>> wrote:

No reason specifically,
I always thought that most people associate B-format with the traditional
1st-order specification, and maybe the FuMa definition up to 3rd-order. I
wasn’t sure if that’s the common term for general HOA signals.

So, you think something like HOA B-format is better (if I don’t have to
go into the details, channels, normalization etc.) ?


IMHO each of A-, B-, C-, D- and -formats are generic terms.

So B-format is " the spherical harmonics stage in Ambisonics" regardless
of channels present or what order they are in.

As for (your) "HOA B-format", no criticism, but my take is that unless
context implies/demands otherwise then 'an ambisonic file' (or 'ambisonic
signal set') is implicitly B-format. (To be pedantic there is a reasonable
presumption that it is B-format.).
But, there must be much to be said for instructions to novices being
explicit (if only once).

I agree with that, I’ll just add explicit definitions in the documentation, SH 
conventions and channel ordering for the first-order and high-order processing, 
and leave to people to pick the term they prefer. In the end they all descrbe 
the same thing..

BR,
Archontis
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

Hi Albert, 

Thanks for the feedback!

> On 21 Jun 2016, at 07:38, Albert Leusink  wrote:
> 
> Well done Archontis!. I was hoping this could be possible, somebody just had
> to put the work into it...thanks so much for this.
> 
> The examples are (sort of) working for me but sound very low res and dull. 
> 
> When I switch between 1,2,3 and 4th order in the HOA decoder, it seems that
> each time when you increase the order, it extends the frequency range
> upwards by half an octave…

The low-pass effect at the moment is the effect of order truncation. HRTFs at 
higher frequencies have too much directional variability, and that means that 
these high-frequencies are reflected to the higher-order filters. Dropping the 
HO generates the low-pass effect. Note that this is also due to the basic 
implementation of the decoding filter examples. You can correct (on average) 
this high-frequency loss in a systematic way, and generally you can improve the 
colouration by tuning the decoder, which is something that these filters do not 
have. You cannot correct however the spatial spreading or blurring of the lower 
orders (or decreased externalization and elevation effects).

The idea of the examples was to show that you can include your own decoding 
HRTF-based filters. If somebody has their own optimized binaural decoding 
filters, I can provide the same examples with their own filters (or gradually 
write some proper documentation).

If filters are not provided, then the decoder just creates two opposing 
cardioids, which have much less colouration (and worse binaural effects) - I’ll 
try to add this mode on the examples too for comparison.

> 
> Most of the first order samples reproduce nothing above 10k and a lot of
> them sound severely saturated/clipped.

Hm, that seems to be my mistake, I noticed it too at some other browser 
recently. 
I think it’s due to having normalized the decoding filters to peak unity, which 
may result in clipped output after all processing. I’ll check it when I can.

> 
> You note on your Github that you used an Eigenmike for the recordings, could
> this be the issue? I've never heard anything musical sounding coming out of
> that microphone, of course the localization is stellarbut it seems to get 
> duller and duller the more orders you truncate

In a sense the Eigenmike provides a much better first-order B-format. Using 
proper encoding filters the range of the dipoles and omni are close to the 
ideal ones for up to ~9kHz, for the Soundfield the patterns start to deviate 
from ideal at lower frequencies. But you are right, I also haven’t managed to 
get from the Eigenmike the great sound quality I’ve heard from Soundfield 
recordings.

Regards,
Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

No reason specifically, 
I always thought that most people associate B-format with the traditional 
1st-order specification, and maybe the FuMa definition up to 3rd-order. I 
wasn’t sure if that’s the common term for general HOA signals. 

So, you think something like HOA B-format is better (if I don’t have to go into 
the details, channels, normalization etc.) ?

Archontis


> On 21 Jun 2016, at 00:14, Courville, Daniel  wrote:
> 
> Politis Archontis wrote:
> 
>> Good point, I?ll add some clarification in the dcumentation that whenever I 
>> mention B-format I mean first-order B-format (which should be unambiguous).
> 
> Thanks. But I'm curious: any reason why you don't want to use the term 
> "B-Format" in HOA?
> 
> People working with HOA in the last ten years or so have kept the B-Format 
> moniker for the spherical harmonics stage in Ambisonics.
> 
> - Daniel
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

Hi Daniel,
> On 19 Jun 2016, at 19:21, Courville, Daniel  wrote:
> 
> Politis Archontis wrote:
> 
>> https://github.com/polarch/JSAmbisonics
>> - WebAudio_FOA.js: Implements B-format encoding, rotations, virtual 
>> microphones, acoustic intensity analysis, and binaural decoding
>> - WebAudio_HOA.js: Implements HOA encoding, rotations, virtual microphones 
>> and binaural decoding for a user-specified order
> 
> This looks great. My only concern is regarding terminology: you seem to use 
> the term "B-Format" only for 1st order (and I presume it's of the FuMa kind), 
> and you drop it for HOA.
> 
> So, on the GitHub page, we read:
> 
> • HOA_bf2acn: converts a B-format stream to an ACN/N3D HOA stream
> • HOA_acn2bf: converts the first-order channels of a HOA stream to B-format
> 
> This could be confusing for ambisonic newbies, I think...

Good point, I’ll add some clarification in the dcumentation that whenever I 
mention B-format I mean first-order B-format (which should be unambiguous).

Were the examples working for you at all?

BR,
Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

[Sursound] Ambisonics on the web pt.1: Web Audio FOA/HOA ambisonic objects

[Politis Archontis]

Hi,

I thought this may be of interest to some - I finally found some time to 
organize some code for real-time ambisonic spatialization and binaural decoding 
on the browser, using the Web Audio API and JavaScript. I just published the 
code in Github, you can find it at:

https://github.com/polarch/JSAmbisonics

The library is split into 3 parts:

- WebAudio_FOA.js: Implements B-format encoding, rotations, virtual 
microphones, acoustic intensity analysis, and binaural decoding

- WebAudio_HOA.js: Implements HOA encoding, rotations, virtual microphones and 
binaural decoding for a user-specified order

- JSHlib.js: A (non-audio) set of functions computing spherical harmonics, 
Legendre polynomials, rotation matrices for spherical harmonics, and the 
forward and inverse spherical harmonic transform. It is required by the 
WebAudio_HOA.js objects.

The objects are easy to set-up, it needs a few lines of code to initialize and 
connect the FOA/HOA blocks. The documentation at the moment is sketchy, but you 
can see some real-time examples at the bottom of the webpage, and check their 
page source code:
https://github.com/polarch/JSAmbisonics#examples
(If the play-button is not enabled, it means that the page is still loading the 
soundfile..)

Web Audio is still WIP, and different browsers implement things differently. I 
expect it to work on Firefox and Chrome (most likely on Android too), but not 
on other browsers. Any feedback, comments or recommendations are very welcome.

Best regards,
Archontis Politis
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] OSSIC Headphones (was Re: Using Ambisonic for a live streaming VR project)

[Politis Archontis]

Hi John, 

I am wondering if you have any information on their personalization approach? I 
have worked a bit on that problem, and I know that you can either do it by some 
visual/scanning input (e.g. using a camera and extracting morphological 
dimensions from images), or some acoustic input (e.g. taking some reference 
measurements and try to infer the shape of the HRTFs at all directions from 
these), or tune parametric HRTF filters by giving the user some active 
listening task (e.g. a localization game).

I could see OSSIC being in the second category, but they do not advertise any 
microphones inside the casing (and even if they had, from that distance you can 
measure mostly the canal response rather than the far-field HRTF useful for 
binaural listening). Probably you can get a rough estimate of the distance 
between the two ears without any of these approaches, to tune some rough ITDs 
based on a sphere, but this is not exactly the HRTF personalization advertized 
in their website? I am thinking of trying them, and any additional info before 
buying would be welcome.

Regards,
Archontis

> On 08 Jun 2016, at 22:34, John Leonard  wrote:
> 
> Thanks for all the replies: I need to sort out exactly how the discount offer 
> works: I think I have to sign up to buy the headphones (I’m a Kickstarter 
> backer for these) and then apply the discount. If that’s the case, you’ll 
> have to send me the money upfront and you must understand that, as a 
> Kickstarter project, there’s no guarantee when they’ll be ready or when 
> they’ll arrive. One other headphone system that I’ve backed is many months 
> overdue for various reasons not wholly under the control of the producer.
> 
> I’ll do a bit more research and contact those who’ve expressed and interest 
> in due course.
> 
> Thanks,
> 
> John
> 
> Please note new email address & direct line phone number
> email: j...@johnleonard.uk
> phone +44 (0)20 3286 5942
> 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Using Ambisonic for a live streaming VR project

[Politis Archontis]

Hi Stefan,


On 07 Jun 2016, at 04:35, Stefan Schreiber 
mailto:st...@mail.telepac.pt>> wrote:

Politis Archontis wrote:

But instead of combining all microphones to generate the binaural directivities 
(as in ambisonics), it interpolates only between the two adjacent microphones 
that should be closest to the listener’s ears. Otherwise, it does not capture 
pinna cues or cues from a non-spherical/assymetrical head.
Any source  for this explanation?

I actually dare to question your view... How will you receive any binaural cues 
via interpolation between two relatively closely spaced omni mikes (fixed on a 
sphere)?

As you even write, this doesn't seem to give any head and pinna cues. (It's 
called MTB. So I guess they would aim to provide several binaural perspectives, 
including head and pinna cues?)

The source is the AES paper describing the method:

Algazi, R. V., Duda, R. O., & Thompson, D. M. (2004). Motion-Tracked Binaural 
Sound. In 116th AES Convention. Berlin, Germany.

It does give head-related cues, that of a spherical head without pinna. If you 
put an omni on a rigid sphere, it is not an omni anymore, it has a 
frequency-dependent directionality, if you put two of them at opposite sides, 
they have opposite directionalities and introduce inter-channel level 
differences. Depending on the size of the sphere, the two signals have a 
direction-dependent phase-difference too. If the size of the sphere is 
approximately the size of a head, then you can assume that the level and time 
differences are close to the binaural ones. This is the infamous spherical head 
model, and its ITDs and ILDs are known analytically. It captures the cues for 
lateralization, but not for a pinna (that it doesn’t have) or for head 
assymetries.

If instead of two omnis, you put many of them on the horizontal plane, then you 
can track the listener’s head yaw rotation and use the two omnis that are 
closer to their ears - or interpolate for a smoother transition. That’s what 
Algazi and Duda are doing in their paper, and they compare various 
interpolation schemes.

Regards,
Archontis
-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20160607/4d3112ff/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Using Ambisonic for a live streaming VR project

[Politis Archontis]

Hi Dave,

just a clarification, Algazi & Duda’s system does not use binaural recordings, 
it approximates them using a spherical microphone array, not dissimilar to ones 
used in ambisonics. 
But instead of combining all microphones to generate the binaural directivities 
(as in ambisonics), it interpolates only between the two adjacent microphones 
that should be closest to the listener’s ears. Otherwise, it does not capture 
pinna cues or cues from a non-spherical/assymetrical head. 

To be honest, the ambisonic recording and decoding provides a better framework 
for imposing individualized or non-individulized HRTFs on the recordings (it is 
harder to do that in their method), but you need high-orders to capture the 
high-frequency cues.

Regards,
Archontis

> On 05 Jun 2016, at 20:14, Dave Hunt  wrote:
> 
> Hi,
> 
> It is hardly surprising that "I hear directional information and head 
> tracking effects, but have never experienced the externalization and 
> verisimilitude that direct dummy head or Algazi and Duda's motion-tracked 
> binaural recordings can produce."
> 
> Even software direct binaural encoding seems more 'accurate' than B-Format 
> ambisonics recoded to binaural. Then Algazi and Duda's system uses binaural 
> recordings, and they know what they're doing.
> 
> Decorrelation, and software reverb,  can help with a sense of 
> externalisation, though you can go too far.
> 
> Ciao,
> 
> Dave Hunt
> 
> 
>> From: Aaron Heller 
>> Date: 4 June 2016 20:53:09 BDT
>> To: Surround Sound discussion group 
>> Subject: Re: [Sursound] Using Ambisonic for a live streaming VR project
>> 
>> 
>> My experience with FOA-to-binaural rendering is pretty much the same as
>> what Acrhontis says.   I hear directional information and head tracking
>> effects, but have never experienced the externalization and verisimilitude
>> that direct dummy head or Algazi and Duda.'s motion-tracked binaural
>> recordings can produce.
>> 
>> Aaron (hel...@ai.sri.com)
>> Menlo Park, CA

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Using Ambisonic for a live streaming VR project

[Politis Archontis]

Hi Peter,

I have heard 7th-order real-time decoding with headtracking, and that’s from a 
real microphone array. There was no perceptible latency.

And I think the AmbiX plugins can handle the rotations and (N+1)^2 short 
convolutions for the same order without problem (head-tracking performance 
though inside a DAW can be laggy I guess..).

"Achieved so far” depends I guess, if you want to capture the full 
high-frequency variability of HRTFs, you need about 15th order HOA signals. We 
cannot record that high with any practical microphone array. But if 7th-order 
decoding has a very small imperceptible difference compared to 15th, or if 
parametric decoders can achieve the same result with the first few orders for 
95% of the sound scenes, then we don’t need to go that much. There is some 
literature that shows the maximum required order for HRTFs (from a physical 
perspective), and the effect of the order on the perceptual perspective.

On the other hand, for mixing of synthetic scenes a recent PC can probably 
handle the encoding to 256 HOA channels for 15th order, the matrix 
multiplications for rotations and the 256 short convolutions, using most of the 
computer’s resources, but I don’t believe that’s a smart way to go.

Best regards,
Archontis




> On 06 Jun 2016, at 11:49, Peter Lennox  wrote:
> 
> That reminds me - what's the highest order ambisonic-to-binaural encoding 
> achieved so far, in combination with head-tracking (and with what latency)? - 
> anyone know?
> Cheers
> ppl
> 
> Dr. Peter Lennox
> Senior Lecturer in Perception
> College of Arts
> University of Derby, UK
> e: p.len...@derby.ac.uk
> t: 01332 593155
> https://derby.academia.edu/peterlennox
> https://www.researchgate.net/profile/Peter_Lennox

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Using Ambisonic for a live streaming VR project

[Politis Archontis]

Hi Jörn,

On 03 Jun 2016, at 15:27, Jörn Nettingsmeier 
mailto:netti...@stackingdwarves.net>> wrote:

Note however that while the quality of first-order to binaural is quite good 
because the listener is by definition always in the sweet spot, first-order 
over speakers can be difficult for multiple listeners when they're far outside 
the center.


This is by no means meant to provoke, but I have never managed to hear a 
convincing B-format to binaural rendering, or to produce one myself. Could you 
possibly share some info on the decoding approach that you used that results in 
a good example?

In my experience, no matter how much tweaking in the decoding, there is severe 
localization blur, due to the large inherent spreading of directional sounds, 
and low envelopment due to the wrong (high) coherence in the binaural signals 
with reverberant sound, compared to the actual binaural coherence. And there is 
also serious colouration, with a loss of high-frequencies, that seems 
direction-dependent. The fact that everything is on the ideal sweet spot under 
free-field conditions doesn’t seem to improve much, it actually seems to do 
more harm (I believe that a small amount of natural added decorrelation from a 
room and tiny misalignments from speakers etc. seem to improve binaural 
coherence and the perceptual quality somewhat of loudspeaker B-format 
reproduction).

Listening to a binaural rendering from a real B-format recording is not so bad, 
there is no reference for comparison, but for VR the difference I’ve heard 
between using directly HRTFs and B-format rendering is huge. And as many of 
these applications rely on sharp directional rendering with accurate 
localization of multiple sound events, traditional B-format decoding seems 
unsuitable to me. The performance improves somewhat with 2nd-order rendering, 
and significantly with 3rd and 4th-order rendering. Also it improves 
dramatically using plain B-format with a well-implemented parametric active 
decoder, such as HARPEX or DirAC.
So my guidelines for VR till now have been,
a) if bandwidth is not an issue go for HOA rendering (at least 3rd-order),
b) if it is an issue, like the streaming application of the OP, stream B-format 
and use an active decoder at the client side.

But I’d like to hear many opinions on this too, and any counter examples!

Regards,
Archontis Politis

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Using Ambisonic for a live streaming VR project

[Politis Archontis]

Hi Antoine,

I don't know of any out-of-the-box solutions for that, maybe other people on 
the list do, but a DiY solution is to use one of the audio programming 
environments that can stream audio, and they have modules for ambisonic 
decoding and rotation.

Puredata (Pd) has such objects from where you can stream the audio, and then on 
the client side, have a predate patch that takes head tracking input, rotates 
the B-format, and does the binaural decoding. Objects from IEM, Graz, do all 
these operations, and it is easy to integrate head-tracking in Pd at present 
(e.g. using the Razor IMU).

This is a solution that needs though some development from your side, using 
Pd's visual programming style.

One correction, you do not need to convert to stereo if you are using head 
tracking and headphones, but to binaural audio, so you have to get some HRTFs 
involved at the last stage.

Best regards,
Archontis Politis


From: Sursound [sursound-boun...@music.vt.edu] on behalf of Antoine Simon 
[anto...@oxygenstream.fr]
Sent: 02 June 2016 19:13
To: sursound@music.vt.edu
Subject: [Sursound] Using Ambisonic for a live streaming VR project

Hello,

We would like to use an ambisonic microphone + a 360° camera for live
streaming VR projects.

I know how to stream the 360° video in live for a VR headset like GearVR or
> a Cardboard.
> But I would like to understand more about how to stream the ambisonic
> audio.


*This is what I understood so far (maybe I'm wrong):*

   - Ambisonic audio equals to 4 channels in "B" format (W, X, Y and Z)
   - The player (client side) need to interpret those 4 channels in "B"
   format into a stereo, depending on the VR Headset position into the 3d
   space.


*My basic Question:* Do you know an exemple or a way to stream that
ambisonic audio + a player to interpret the stereo depending on the VR
Headset position in realtime?

*My further questions:*

   - Can we transform those 4 channels in "B" format into a 4.0 AAC for the
   streaming without loosing too much data compared to a WAV PCM?
   - How to interpret those 4 channels in "B" format into a stereo,
   depending on the VR Headset position into the 3d space?


Thanks in advance for your answers...
Best regards,


*Antoine SIMON*
Live Events Manager, OxygenStream
Mob: +33 698 171 875 | Tel: +33 954 826 133
anto...@oxygenstream.fr | http://oxygenstream.fr

[image: OxygenStream Website] 
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Reduced set of B-format (or HOA) to binaural filters

[Politis Archontis]

Thanks for the clarification, this makes perfect sense..

Archontis

On 26 May 2016, at 10:06, Fons Adriaensen 
mailto:f...@linuxaudio.org>> wrote:

On Wed, May 25, 2016 at 10:23:23PM +, Politis Archontis wrote:

So is this how it is usually done? Convolving all (N+1)^2 HOA
channels of order N with the set of binaural filters and summing
for the left ear, and then doing the same for the right but with
inverted polarity for the (N^2+N)/2 HOA channels of m<0 ?

There's no need to store all the convolution results and do the
summing twice: let

M = sum of components with m >= 0
S = sum of compomemts with m < 0
L = M + S
R = M - S

Ciao,

--
FA

A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)

___
Sursound mailing list
Sursound@music.vt.edu<mailto:Sursound@music.vt.edu>
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20160526/a816f391/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Reduced set of B-format (or HOA) to binaural filters

[Politis Archontis]

On 26 May 2016, at 01:04, Eric Benjamin 
mailto:eb...@pacbell.net>> wrote:

I have used that symmetry in the past. I assume that everyone does.


Hi Eric,

good to know that it’s used by everybody and I wasn’t totally off in my 
“revelations” :-). It wasn’t mentioned before in the discussion about 
generating HOA to binaural filters (even though nothing about symmetry was 
mentioned either).

So is this how it is usually done? Convolving all (N+1)^2 HOA channels of order 
N with the set of binaural filters and summing for the left ear, and then doing 
the same for the right but with inverted polarity for the (N^2+N)/2 HOA 
channels of m<0 ?

Best regards,
Archontis

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Reduced set of B-format (or HOA) to binaural filters

[Politis Archontis]

Hi Fons,

thanks for the info, it is clear to me now that for any order of HOA to 
binaural decoding, only (N+1)^2 filters are required for left-right symmetric 
HRTFs, with a subsequent polarity inversion of N channels corresponding to the 
left-right antisymmetric SHs (the sine ones) in order to get one of the two ear 
signals.

Best,
Archontis

On 25 May 2016, at 21:07, Fons Adriaensen 
mailto:f...@linuxaudio.org>> wrote:

On Wed, May 25, 2016 at 03:37:34PM +, Politis Archontis wrote:

My question to any of the decoder developers on the list is if you
have seen that anywhere analyzed or mentioned in the context of
binaural decoding??

Haven't seen it in any publications, but the 2nd and 3rd order
amb2bin apps I wrote recently work as you suggest. The convolutions
do AMB -> M/S, then M/S -> L/R is trivial.

Ciao,

--
FA

A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)

___
Sursound mailing list
Sursound@music.vt.edu<mailto:Sursound@music.vt.edu>
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20160525/00f4fe1e/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Reduced set of B-format (or HOA) to binaural filters

[Politis Archontis]

Thanks a lot Bruce for the pointers,
I will check your work!

Regards,
Archontis

On 25 May 2016, at 19:42, Bruce Wiggins 
mailto:bruce.wigg...@gmail.com>> wrote:

Definately mentioned and looked at in
Wiggins, B. Paterson-Stephens, I., Schillebeeckx, P. (2001) The analysis of
multi-channel sound reproduction algorithms using HRTF data. 19th
International AES Surround Sound Convention, Germany, p. 111-123.

Wiggins, B. (2004) An Investigation into the Real-time Manipulation and
Control of Three-dimensional Sound Fields. PhD thesis, University of Derby,
Derby, UK. p. 103

See http://www.brucewiggins.co.uk/ for my look at the you tube
implementation which also assumes symmetry..

Cheers

Bruce

On Wed, 25 May 2016 16:38 Politis Archontis, 
mailto:archontis.poli...@aalto.fi>>
wrote:

Hi,

There has been some discussion before on B-format or HOA to binaural
filters (HRTF-based), which no matter if it goes through a virtual decoder
or with a direct HRTF-to-Bformat approach, two times the number of HOA
channels of filters are needed (so 8 for B-format, 18 for 2nd-order HOA, 32
for 3rd-order HOA etc. ).

Playing around a bit I realized that if the HRTFs are made left-right
symmetric, which makes sense to always force for non-individialized ones
(and also for individualized sometimes), then a smaller set of filters is
needed due to this left-right symmetry. For example for B-format binaural
filters, the W, X and Z channel would have exactly the same filters, while
the Y channel would have the same filter with a polarity inversion for one
ear with respect to the other. Hence 4 filters instead of 8.

Thinking about it, it simply makes sense, as the spherical harmonics are
themselves either symmetric or antisymmetric with respect to the x-z
(median) plane. This is also what makes mirroring of a HOA sound scene from
left to right for example, as easy as inverting certain HOA channels.

My question to any of the decoder developers on the list is if you have
seen that anywhere analyzed or mentioned in the context of binaural
decoding??

Many thanks,
Archontis Politis





-- next part --
An HTML attachment was scrubbed...
URL: <
https://mail.music.vt.edu/mailman/private/sursound/attachments/20160525/a7b7b617/attachment.html

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
edit account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20160525/f793cacf/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu<mailto:Sursound@music.vt.edu>
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20160525/e9cb0744/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

[Sursound] Reduced set of B-format (or HOA) to binaural filters

[Politis Archontis]

Hi,

There has been some discussion before on B-format or HOA to binaural filters 
(HRTF-based), which no matter if it goes through a virtual decoder or with a 
direct HRTF-to-Bformat approach, two times the number of HOA channels of 
filters are needed (so 8 for B-format, 18 for 2nd-order HOA, 32 for 3rd-order 
HOA etc. ).

Playing around a bit I realized that if the HRTFs are made left-right 
symmetric, which makes sense to always force for non-individialized ones (and 
also for individualized sometimes), then a smaller set of filters is needed due 
to this left-right symmetry. For example for B-format binaural filters, the W, 
X and Z channel would have exactly the same filters, while the Y channel would 
have the same filter with a polarity inversion for one ear with respect to the 
other. Hence 4 filters instead of 8.

Thinking about it, it simply makes sense, as the spherical harmonics are 
themselves either symmetric or antisymmetric with respect to the x-z (median) 
plane. This is also what makes mirroring of a HOA sound scene from left to 
right for example, as easy as inverting certain HOA channels.

My question to any of the decoder developers on the list is if you have seen 
that anywhere analyzed or mentioned in the context of binaural decoding??

Many thanks,
Archontis Politis





-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Integrating binaural recordings into ambisonics

[Politis Archontis]

Hi Jamie,

I’m not sure the transaural to B-format approach would work as, if I understand 
correctly, the cross-talk cancelation filters equalize the direct paths and 
minimize the cross-talk paths
at two control points that are spaced apart (the ears). I’m not sure how you 
would capture that information with coincident recording?

One approach that could work is a parametric approach DirAC-like, but that 
would require quite some work. For that you need an estimate of diffuseness 
(based on the interaural coherence) and an estimate of the direciton of arrival 
(based on estimated interaural time and level differences). Then you could 
re-encode you binaural signals to B-format based on these parameters.

There are quite a few details missing from the above description but I think 
it’s possible, even though it needs a careful implementation and some 
experimentation.
You can also check some quite ahead of its time work from Julia Jakka in her 
master’s thesis here in Aalto university:

http://lib.tkk.fi/Dipl/2005/urn007903.pdf

Best regards,
Archontis


On 11 May 2016, at 21:10, Jamie Smith 
mailto:jamie.sm...@uhi.ac.uk>> wrote:

Hello all,
new subscriber here, so apologies if this topic has already been discussed.

Can anyone offer a way of successfully integrating a recording that
has been made binauraly into an ambisonic project?  The end product
would be delivered on headphones but, critically, must allow for head
tracking.  Clearly the two methods have compatibility issues.

As a work-around I have considered this:
binaural recording - pass through transaural processor (or similar) -
send to 2 virtual speakers - encode into b-format - listen on
headphones via ambi explorer or some VR player.

It seems convoluted and likely to degrade the spatial and spectral information.

Any thoughts?


Jamie Smith
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] YouTube now supports Ambisonics (warning....part advertisement..)

[Politis Archontis]

> That would be wonderful to have a HOA version of this. I had searched on
> your website for the tool to make it myself, but the page doesn't load
> http://mchapman.com/amb/soft/positions/
> 
> Regards,
> 
> Albert

Hi,

is that any different from opening up Reaper or any other DAW with some HOA 
plugins like the AmbiX ones, and encoding your mono test samples to any set of 
directions, and for the order that you need?

Regards,
Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Can anyone share test HOA (higher-order ambisonic) recordings?

[Politis Archontis]

Hi Anastasia,

> Dear sursounders,
> 
> Do any of you have any HOA recordings you could share for some testing? I 
> can’t quite get my hands on a proper mic. (Please nothing under license / NDA 
> / etc)

We have some orchestral and theatrical play recordings done with an Eigenmike 
here. May I ask what you would be using them for?

> Also, any recommendations for an HOA mic other than Eigenmike and RealSpace3D 
> one?

There is additionally the Zylia Audioimmersion one mentioned recently on this 
list, but I have never tried it myself.

Best regards,
Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] YouTube now supports Ambisonics (warning....part advertisement..)

[Politis Archontis]

Hi Albert,

This is interesting!
I don’t have an android and I cannot test it unfortunately. I would like to ask 
what are you doing with the B-format? decode it to headphones dynamically?

Archontis

> On 20 Apr 2016, at 18:29, Albert Leusink  wrote:
> 
> Hello all,
> 
> Some welcome news,
> 
> YouTube is now supporting ACN/SN3D FOA attached to a .mov video container as
> uncompressed PCM audio. 
> 
> This is great, as most of the world gets a lot of their content through
> YouTube and now they'll all get to experience the power of Ambisonics!.
> 
> For now, it only works on Android, using the YouTube app, but hopefully soon
> more platforms will be supported, if we create enough good content in this
> format.
> 
> To contribute to that end, we've started our own channel, 360 Performances,
> featuring music and performing arts in 360º video and recorded/mixed in
> Ambisonics:
> 
> https://www.youtube.com/channel/UC6_696Vph-306CVsenXvbvQ
> 
> Hopefully this will turn into something viable, and we can continue doing
> this for more artists and keep releasing new content in this manner.
> 
> As nowadays it's more about the number of subscribers/likes/views than the
> quality of the content, please do subscribe to the channel if you like this
> (and even if you don't like it, you can still subscribe...:-)
> 
> Regards,
> 
> Albert
> 
> (PS I don't work for YouTube)
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] YouTube now supports Ambisonics (warning....part advertisement..)

[Politis Archontis]

Hi David,

if you are asking if the ACN/SN3D FOA is the same as the B-format, yes, 
first-order is the B-format anyway you look at it. 
The B-format in ACN/SN3D convention is the same as in the traditional 
definition without the W scaling and in the ACN order:

B_ACN_SN3D = [sqrt(2)*W Y Z X]. 

Regards,
Archontis



> On 20 Apr 2016, at 19:09, David Pickett  wrote:
> 
> At 17:29 20-04-16, Albert Leusink wrote:
> 
> >YouTube is now supporting ACN/SN3D FOA attached to a .mov video container as
> >uncompressed PCM audio.
> 
> I am sure I should know, but if this is a multitrack thing, what do the 
> channels each contain?  B format or what?
> 
> David
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Are mems a good choice for ambisonic microphones?

[Politis Archontis]

Hi David,

I’m not too knowledgable in this stuff, especially from the manufacturing side, 
but just by listening I would rate what I get from the Soundfield (ST350/450 & 
SPS200) the best from the arrays I have used so far, so maybe something with an 
EIN of ~12dBA, and that’s not taking into account all other properties like 
diaphragm size, capsule type and construction.. Eigenmike claims ~15dBA I 
think, not bad either. MEMs are typically between 30~35 I think.

Best,
Archontis


> On 14 Apr 2016, at 15:29, David Pickett  wrote:
> 
> At 14:02 14-04-16, Politis Archontis wrote:
> 
> > Very high quality recording/reproduction would still
> > require proper capsules in my opinion
> 
> What do you regard a minimally "proper capsules"?
> 
> David 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Are mems a good choice for ambisonic microphones?

[Politis Archontis]

Hi,

I have listened to some very nice binaural renderings of a 7th-order spherical 
array with 64 MEMs microphones on the top of a custom hard 3D-printed sphere, a 
bit larger than the Eigenmike. 64 MEMs would result in an 18dB SNR improvement 
for the omni output (all the higher-order signals would still have a 
frequency-dependent SNR improvement). The single channel quality that it was 
giving was not as high as what I’m getting from the Eigenmike or one of the 
Soundfields, but pretty good for most applications, especially in the 
telepresence context that was used.

Another interesting fact I heard from the makers was that they got tens of the 
MEMs in a batch, and checking them they found small magnitude response 
deviations between them, +-1dB. All the electronics were also housed inside in 
a very compact FPGA getting the digital signals and converting them to a single 
convenient MADI cable. Considering all these, MEMs seem an easy choise for 
large arrays. Very high quality recording/reproduction would still require 
proper capsules in my opinion, with the trade-off a harder design and effort in 
construction.

Regards,
Archontis


-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Anyone know anything about this?

[Politis Archontis]

Just a note that direct beamforming is not necessarily worse than 
encoding/decoding, it can be used for example to approximate in an optimal 
(least-squares) sense the directional patterns of an ambisonic decoder for some 
target speaker setup.

I haven’t met personally the people in the company, but I’ve heard good things 
about them from colleagues who spoke with them in conferences, both about their 
enthusiasm and their technical knowledge. And a second higher-order microphone 
in the market other than the Eigenmike (and probably cheaper) doesn’t harm..

Regards, 
Archontis Politis

> On 09 Apr 2016, at 14:02, Paul Hodges  wrote:
> 
> --On 08 April 2016 18:14 +0200 Jörn Nettingsmeier
>  wrote:
> 
>> a casual glance over the site seems to suggest direct beamforming
>> without an intermediate b-format.
> 
> But note, in the FAQ (under VR):
> "Audio is recorded in the higher order ambisonic format which allows
> precise representation of the 3D audio scene."
> 
> Paul
> 
> -- 
> Paul Hodges
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] [Warning! advertisement] Hefio earphones

[Politis Archontis]

I guess that’s their target application, you start with a physical flat 
response, and then the user can apply a personalized target response that he 
wants. That can be perceptual equalization, stereo widening filters, hrtfs, 
etc. I think they offer some equalizatio or flitering tools for that. The 
virtual monitoring application you mentioned is another, and then there are 
more specialized applications of more audiological interest, where you need to 
control the stimuli that reaches the ear drum.

But I shouldn’t be speaking for them, if somebody’s interested you should 
contact them, I’m sure they’ll be more than happy to answer all questions.

Regards,
Archontis


> On 08 Apr 2016, at 09:52, Steven Boardman  wrote:
> 
> I kind of guessed that after posting,  but the marketing l read appeared
> just to push only the flat response. Which i suppose to most people sounds
> like a great idea on its own. The thing is, without a personalised HRTF,
> they are likely to sound worse to most people.  Well at least the frequency
> response won't sound flat.
> I personally use headphones that sound flat to me,  or where I know where
> the deficiencies are.
> Another method is to A/B with a known speaker set up,  and eq the
> headphones until they match. Unfortunately this is trial and error. So I do
> see the value of these, if one has HRTF set that matches, or has been
> learnt.
> 
> Best
> 
> Steve
> On 7 Apr 2016 10:53 pm, "Politis Archontis" 
> wrote:
> 
>> Hi Steve,
>> 
>> I guess the idea is that if you equalize the response of the
>> headphones/earphones, then you can apply the target response you need
>> without undesired modifications by the headphones, and that can be
>> individualized HRTFs if you have them, which include the effects you
>> mentioned.
>> 
>> Regards,
>> Archontis
>> 
>>> On 08 Apr 2016, at 00:32, Steven Boardman 
>> wrote:
>>> 
>>> Not sure one needs actual flat response at the ear drum.
>>> Surely it needs to sound like the torso,  head,  pinna and ear canal have
>>> filtered the sound before we think its flat?
>>> 
>>> Best
>>> 
>>> Steve
>> 
>> ___
>> Sursound mailing list
>> Sursound@music.vt.edu
>> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
>> edit account or options, view archives and so on.
>> 
> -- next part --
> An HTML attachment was scrubbed...
> URL: 
> <https://mail.music.vt.edu/mailman/private/sursound/attachments/20160408/77836259/attachment.html>
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] [Warning! advertisement] Hefio earphones

[Politis Archontis]

Hi Steve,

I guess the idea is that if you equalize the response of the 
headphones/earphones, then you can apply the target response you need without 
undesired modifications by the headphones, and that can be individualized HRTFs 
if you have them, which include the effects you mentioned.

Regards,
Archontis

> On 08 Apr 2016, at 00:32, Steven Boardman  wrote:
> 
> Not sure one needs actual flat response at the ear drum.
> Surely it needs to sound like the torso,  head,  pinna and ear canal have
> filtered the sound before we think its flat?
> 
> Best
> 
> Steve

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

[Sursound] [Warning! advertisement] Hefio earphones

[Politis Archontis]

Hi,

this is an advertisement pretty much, but I thought it could be interesting to 
people of the list.

An ex-colleague of mine here at Aalto university made the brave decision to 
start his own company couple of years ago, applying the research he was doing, 
and these days they announced their first product. The produce self-calibrating 
earphones that equalize individually the response from the transducer to the 
eardrum. The applications can vary, but HRTF/binaural reproduction can be one 
of them.

More info here:

http://www.hefio.com/

Best regards,
Archontis




-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics for Android and also Oculus..

[Politis Archontis]

Oops, that should have been “About N3D…” instead of “About ACN...”

I agree with Fons that there isn’t much reason for multiple inconsistent 
definitions, especially in the case of HOA, and then even first-order can be 
treated in the same framework. There is the intuitive ‘XYZ’ interpretation of 
first-order signals (which disappears altogether already from second order), 
which is probably useful for design of first-order systems, but that doesn’t 
mean that the signals cannot be stored in a more universal ACN/(S)N3D format, 
and convert the B-format tools to handle the ACN signals internally. It is just 
re-arranging of channels and a scaling in the end.

Regards,
Archontis

On 03 Apr 2016, at 13:29, Politis Archontis 
mailto:archontis.poli...@aalto.fi>> wrote:

The advantage of ACN is that it is a natural ordering of the HOA channels 
described with a single number q=n^2+n+m+1 (with n the order and m the degree), 
and it is aligned with pretty much any other field using spherical spectral 
analysis (graphics, signal processing, physics, etc...) , making transfer of 
knowledge from these fields to ambisonics much less confusing.

About ACN, if you are working on the more theoretical side of ambisonics, it is 
the only normalization that makes sense again since it keeps the spherical 
harmonics orthonormal and it simplifies most things on paper a lot! It can just 
as well be kept as a standard, however conversion to other normalizations (such 
as SN3D) are just a gain factor per order and very easy to do.

Regards,
Archontis

-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20160403/d3d7977d/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonics for Android and also Oculus..

[Politis Archontis]

The advantage of ACN is that it is a natural ordering of the HOA channels 
described with a single number q=n^2+n+m+1 (with n the order and m the degree), 
and it is aligned with pretty much any other field using spherical spectral 
analysis (graphics, signal processing, physics, etc...) , making transfer of 
knowledge from these fields to ambisonics much less confusing.

About ACN, if you are working on the more theoretical side of ambisonics, it is 
the only normalization that makes sense again since it keeps the spherical 
harmonics orthonormal and it simplifies most things on paper a lot! It can just 
as well be kept as a standard, however conversion to other normalizations (such 
as SN3D) are just a gain factor per order and very easy to do.

Regards,
Archontis





On 03 Apr 2016, at 10:51, Xavier Bonjour 
mailto:x.bonj...@3dsoundlabs.com>> wrote:

Hello,

And to add one more level of confusion, MPEG-H 3D Audio has gone for
ACN/N3D normalization!

Going beyond 3rd order  is also going for better quality and it makes
sense for the following reasons:

- Recording technologies make progress: we should expect multicapsules
microphone technology to improve and go beyond 3rd Order
- 3D sound can also be computer generated to any order and there is a
significant perceptual difference between the binaural rendering of 3rd
order content and content of higher orders. And this difference is even
bigger when HRTF are personalized
- Compression (MPEG-H 3D Audio,  ...)  will make HOA content manageable,
even on a smartphone.

I guess defining a normalization that covers HOA beyond 3rd order is about
getting the "format" ready for the future development.

Kind regards

-Xavier



-Original Message-
From: Sursound [mailto:sursound-boun...@music.vt.edu] On Behalf Of Albert
Leusink
Sent: dimanche 3 avril 2016 02:20
To: sursound@music.vt.edu
Subject: [Sursound] Ambisonics for Android and also Oculus..


Hello all,

Coming soon, to a phone near you !:

https://storage.googleapis.com/jump-inspector/Jump_Inspector_Quick_Start.p
df

There will definitely be many user complaints initially due to
misunderstandings as it uses ACN/SN3D ordering and 99% of all the tools
(VST plugins etc..) currently used by the VR community are FuMa / .AMB
based...

To add to that confusion, Oculus has just updated their Gear VR video
player specs to accept FuMa first order

I'm sure you have all had this discussion many times over, but what, if
any, are the advantages of 1st order ACN over FuMa?

>From what I've gathered, ACN was initially proposed as it would allow
bigger file sizes but I don't think that really applies to mobile phones
in this case...doesn't it only make sense to have ACN beyond 3rd order?



Have a good weekend,


Albert





___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
edit account or options, view archives and so on.
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] expressing HRTFs in spherical harmonics

[Politis Archontis]

> The point is that, since the direct method is equivalent to
> a decoding for the set of directions used to compute the SHT
> (which will be the set for which you have HRIR), there is
> nothing 'ideal' or special to it. It is just one specific case
> of the decoder + virtual speakers method in disguise, and one
> that is very probably using way too many speakers.

Heh, you see that's exactly the opposite of how I would call it, the virtual 
loudspeaker+decoding method would correspond to the direct method "in disguise".

The order of the HRTFs is limited by their spatial/angular variability, and 
that would be estimated correctly by orthonormalization of the measurement 
directions on the SHT stage (as you mentioned). It has nothing to do with too 
many speakers etc.. The limiting happens inherently by the number of HOA 
channels, no matter if the SH coefficients of the HRTFs are of higher order, 
since only the ones corresponding to the HOA components would be used. That 
would equivalent of decoding with the minimum number of speakers for the 
certain HOA order and then doing the binaural conversion.

I will try to show with some plots that there is no error in the direct 
approach compared to a decoding stage. But it has to wait for a bit...

BR,
Archontis

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] expressing HRTFs in spherical harmonics

[Politis Archontis]

>> So what is the benefit then of adding a decoding stage in the middle?

> The advantage of having an explicit decoder stage is that you
> can tweak the decoder for optimum results. For example it can
> be dual-band [1], or have some front preference, etc.

I see. I find that more a matter of preference though than a benefit, or 
whatever one finds more intuitive! I tend to think of these operations not on 
the decoder stage but before that. I mean it is possible to smooth, apply rE 
weighting, warp or modulate the HOA signals on the stage before decoding or 
binauralization. But I can see how it can be useful to think of it as a 
decoding operation. 

One case I can think of when it is really necessary is when one needs to 
auralize binaurally the effect of a certain decoder, e.g. due to non-uniform 
arrangement of virtual loudspeakers, compared to an ideal case..

BR,
Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] expressing HRTFs in spherical harmonics

[Politis Archontis]

> No, this is not true. The decoder and convolution matrix can be combined
> into a (N+1)^2 * 2 convolution matrix.

Ah true! by summing the terms..

> The only remaining difference is the set of directions. And it is known
> that using too many speakers for a given order is suboptimal.

So what is the benefit then of adding a decoding stage in the middle?

BR,
Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] expressing HRTFs in spherical harmonics

[Politis Archontis]

And I forgot to mention in the previous message that, while I don't see any 
benefit in the virtual loudspeaker approach, I see benefits in the direct 
approach. Doing the virtual loudspeaker decoding, you'll need some uniform 
arrangement of decoding directions that will be most likely of more points than 
the harmonics, so you'll need more convolutions compared to the minimum of the 
direct approach (N+1)^2. 

I personally find also useful the fact that storing the HRTFs in an SH format 
gives an efficient and high-resolution interpolation for non-ambisonic binaural 
panning, one that uses all the measurement data to interpolate and not only the 
2-3 surrounding data points (not relevant though to the ambisonic conversion 
question..)

BR,
Archontis


From: Sursound [sursound-boun...@music.vt.edu] on behalf of Fons Adriaensen 
[f...@linuxaudio.org]
Sent: 27 February 2016 12:14
To: sursound@music.vt.edu
Subject: Re: [Sursound] expressing HRTFs in spherical harmonics

On Thu, Feb 25, 2016 at 09:25:48PM +, Politis Archontis wrote:

> - Measure the HRIRs at Q directions around the listener
> - Take the FFT of all measurements
> - For each frequency bin perform the SHT to the complex HRTFs,
>   up to maximum order that Q directions permit (and their arrangement:
>   for equiangular measurement grids the order is N<=4*Q^2).

The ^2 probably should be on N, not Q.

>   You end up with (N+1)^2 coefficients per bin per ear.
> - Take the IFFT for each of the (N+1)^2 coefficients.
>   You end up with 2x(N+1)^2 FIR filters that can be used
>   to binauralize your HOA recordings directly.
> - To binauralize, convolve each HOA signal with the respective
>   SH coefficient filter of the HRTF, for each ear, and sum the
>   outputs per ear.

To me it looks like the FFT/IFFT can be factored out. Both FFT and SHT
are linear transforms, so their order can be swapped. With H (t,q) the
HRIR in direction q:

   IFFT (SHT (FFT (H (t,q = IFFT (FFT (SHT (H (t,q = SHT (H (t,q))

Now if Q is a set of more or less uniformly distributed directions,
the coefficients of a systematic decoder will be very near to just
the SH evaluated in the directions Q. So summing the convolution
of the decoder outputs with the HRIR is equivalent to the SHT on
the HRIR.

In other words, this method is just the same as the 'decoding to
virtual speakers' one, with Q the set of speaker directions and
using a systematic decoder.

Ciao,

> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

--
FA

A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] expressing HRTFs in spherical harmonics

[Politis Archontis]

Hi Fons,

True, slight mistake! for equiangular grids it should be (N+1)^2<4*Q.

You are absolutely correct about the linearity and the exchange of the order of 
the transforms. And the virtual loudspeakers approach should be exactly 
equivalent, and that's the main reason I don't understand why one should use 
it. In the best case, taking into account that the decoding was done properly,  
it will just give the same result as doing the conversion directly on the SH 
domain. 

The virtual loudspeaker approach essentially takes the HOA signals back to the 
spatial domain, via decoding, convolves each plane wave with the respective 
HRTF, and integrates across all directions. Which is done directly in the SH 
domain by convolving the HOA signals with the HRTF SH coefficients, and summing 
the results. I don't see the reason for this extra intermediate inverse SHT 
(decoding) in this case.

Regards,
Archontis

From: Sursound [sursound-boun...@music.vt.edu] on behalf of Fons Adriaensen 
[f...@linuxaudio.org]
Sent: 27 February 2016 12:14
To: sursound@music.vt.edu
Subject: Re: [Sursound] expressing HRTFs in spherical harmonics

On Thu, Feb 25, 2016 at 09:25:48PM +, Politis Archontis wrote:

> - Measure the HRIRs at Q directions around the listener
> - Take the FFT of all measurements
> - For each frequency bin perform the SHT to the complex HRTFs,
>   up to maximum order that Q directions permit (and their arrangement:
>   for equiangular measurement grids the order is N<=4*Q^2).

The ^2 probably should be on N, not Q.

>   You end up with (N+1)^2 coefficients per bin per ear.
> - Take the IFFT for each of the (N+1)^2 coefficients.
>   You end up with 2x(N+1)^2 FIR filters that can be used
>   to binauralize your HOA recordings directly.
> - To binauralize, convolve each HOA signal with the respective
>   SH coefficient filter of the HRTF, for each ear, and sum the
>   outputs per ear.

To me it looks like the FFT/IFFT can be factored out. Both FFT and SHT
are linear transforms, so their order can be swapped. With H (t,q) the
HRIR in direction q:

   IFFT (SHT (FFT (H (t,q = IFFT (FFT (SHT (H (t,q = SHT (H (t,q))

Now if Q is a set of more or less uniformly distributed directions,
the coefficients of a systematic decoder will be very near to just
the SH evaluated in the directions Q. So summing the convolution
of the decoder outputs with the HRIR is equivalent to the SHT on
the HRIR.

In other words, this method is just the same as the 'decoding to
virtual speakers' one, with Q the set of speaker directions and
using a systematic decoder.

Ciao,

--
FA
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] expressing HRTFs in spherical harmonics

[Politis Archontis]

Hi Jorn,

Yes, you’re right. To summarize the process:

- Measure the HRIRs at Q directions around the listener
- Take the FFT of all measurements
- For each frequency bin perform the SHT to the complex HRTFs, up to maximum 
order that Q directions permit (and their arrangement: for equiangular 
measurement grids the order is N<=4*Q^2). You end up with (N+1)^2 coefficients 
per bin per ear
- Take the IFFT for each of the (N+1)^2 coefficients. You end up with 2x(N+1)^2 
FIR filters that can be used to binauralize your HOA recordings directly.
- To binauralize, convolve each HOA signal with the respective SH coefficient 
filter of the HRTF, for each ear, and sum the outputs per ear.

Thinking about it now, it should be very easy to set-up a HOA2binaural 
conversion like that in REAPER, after getting the filters, by using the ambiX 
plugins, and setting the decoder matrix to a unity matrix, then convolving the 
outputs with the FIR filters using the multichannel convolver plugin.

If you would like to experiment, I’ll be happy to process some HRTFs and send 
you the filters.

There are various papers on expansion of HRTFs to SHs, Duraiswami, Dylan 
Menzies, Brungart, Abhayapala, Evans, are some of the names that spring to 
mind. It is a very convenient way of interpolating HRTFs too at any direction, 
so  two birds with one stone..

About the inter-aural time delay, the complex HRTFs should include that 
automatically, since they are measured at the ears, “off-centre” for the 
measuring setup. You can also do it however by the common factorization of 
splitting the HRTFs into a (directional) inter-aural time difference and a 
minimum phase filter, expand the minimum phase HRTFs, and then introduce the 
inter-aural time difference afterwards.

About inter-aural time delay and frequency dependence, I think it has been 
shown that for most practical purposes replacing it with a 
frequency-independent one does not affect much. You can also express the ITD on 
SHs, it is a very convenient representation of it, since it really approximates 
a slightly elongated dipole on the interaural axis, and hence only the first 
2-3 orders are enough to describe it well.

Regards,
Archontis




On 25 Feb 2016, at 22:48, Jörn Nettingsmeier 
mailto:netti...@stackingdwarves.net>> wrote:

On 01/27/2016 01:56 PM, Jörn Nettingsmeier wrote:
On 01/26/2016 11:05 PM, Politis Archontis wrote:
Hi Jorn,

yes that is correct. I think however that the virtual loudspeaker
stage is unnecessary. It is equivalent if you expand the left and
right HRTFs into spherical harmonics and multiply their coefficients
(in the frequency domain) directly with the coefficients of the sound
scene (which in the 1st-order case is the B-format recording). This
is simpler and more elegant I think. Taking the IFFT of each
coefficient of the HRTFs, you end up with an FIR filter that maps the
respective HOA signal to its binaural output, hence as you said it's
always 2*(HOA channels) no matter what. Arbitrary rotations can be
done on the HOA signals before the HOA-to-binaural filters, so
head-tracking is perfectly possible.

Wow. That sounds intriguing, thanks! I'll try to wrap my head around the
SH expression of an HRTF set in the coming months, hopefully with the
help of Rozenn Nicol's book.

Sorry to revive such an old thread, but the AES monograph on binaural 
technology has arrived, and I've begun to study it. Definitely a great 
resource, recommended:

http://www.aes.org/publications/monographs/

Archontis, I'm still trying to understand how to express a set of HRTFS as a SH 
series.
If I understand correctly, all HRTFS for a given ear can be expressed as a 
function on the sphere, but it would be frequency dependent. So we'd need an 
extra degree of freedom there, how does that tie in with Ambisonics? One HRTF 
"balloon" per frequency bin?
Also, how do you express the inter-aural time delay conveniently (which, as 
I've learned from Rozenn Nicol, depends not only on direction, but also on 
frequency)?

Are there papers out there that describe this in detail?

Best,


Jörn



--
Jörn Nettingsmeier
Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487

Meister für Veranstaltungstechnik (Bühne/Studio)
Tonmeister VDT

http://stackingdwarves.net

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20160225/36230812/attachment.html>
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] expressing HRTFs in spherical harmonics

[Politis Archontis]

Hi Jorn,

Yes, you’re right. To summarize the process:

- Measure the HRIRs at Q directions around the listener
- Take the FFT of all measurements
- For each frequency bin perform the SHT to the complex HRTFs, up to maximum 
order that Q directions permit (and their arrangement: for equiangular 
measurement grids the order is N<=4*Q^2). You end up with (N+1)^2 coefficients 
per bin per ear
- Take the IFFT for each of the (N+1)^2 coefficients. You end up with 2x(N+1)^2 
FIR filters that can be used to binauralize your HOA recordings directly.
- To binauralize, convolve each HOA signal with the respective SH coefficient 
filter of the HRTF, for each ear, and sum the outputs per ear.

Thinking about it now, it should be very easy to set-up a HOA2binaural 
conversion like that in REAPER, after getting the filters, by using the ambiX 
plugins, and setting the decoder matrix to a unity matrix, then convolving the 
outputs with the FIR filters using the multichannel convolver plugin.

If you would like to experiment, I’ll be happy to process some HRTFs and send 
you the filters.

There are various papers on expansion of HRTFs to SHs, Duraiswami, Dylan 
Menzies, Brungart, Abhayapala, Evans, are some of the names that spring to 
mind. It is a very convenient way of interpolating HRTFs too at any direction, 
so  two birds with one stone..

About the inter-aural time delay, the complex HRTFs should include that 
automatically, since they are measured at the ears, “off-centre” for the 
measuring setup. You can also do it however by the common factorization of 
splitting the HRTFs into a (directional) inter-aural time difference and a 
minimum phase filter, expand the minimum phase HRTFs, and then introduce the 
inter-aural time difference afterwards. 

About inter-aural time delay and frequency dependence, I think it has been 
shown that for most practical purposes replacing it with a 
frequency-independent one does not affect much. You can also express the ITD on 
SHs, it is a very convenient representation of it, since it really approximates 
a slightly elongated dipole on the interaural axis, and hence only the first 
2-3 orders are enough to describe it well.

Regards,
Archontis


From: Sursound [sursound-boun...@music.vt.edu] on behalf of Jörn Nettingsmeier 
[netti...@stackingdwarves.net]
Sent: 25 February 2016 22:48
To: sursound@music.vt.edu
Subject: [Sursound] expressing HRTFs in spherical harmonics

On 01/27/2016 01:56 PM, Jörn Nettingsmeier wrote:
> On 01/26/2016 11:05 PM, Politis Archontis wrote:
>> Hi Jorn,
>>
>> yes that is correct. I think however that the virtual loudspeaker
>> stage is unnecessary. It is equivalent if you expand the left and
>> right HRTFs into spherical harmonics and multiply their coefficients
>> (in the frequency domain) directly with the coefficients of the sound
>> scene (which in the 1st-order case is the B-format recording). This
>> is simpler and more elegant I think. Taking the IFFT of each
>> coefficient of the HRTFs, you end up with an FIR filter that maps the
>> respective HOA signal to its binaural output, hence as you said it's
>> always 2*(HOA channels) no matter what. Arbitrary rotations can be
>> done on the HOA signals before the HOA-to-binaural filters, so
>> head-tracking is perfectly possible.
>
> Wow. That sounds intriguing, thanks! I'll try to wrap my head around the
> SH expression of an HRTF set in the coming months, hopefully with the
> help of Rozenn Nicol's book.

Sorry to revive such an old thread, but the AES monograph on binaural
technology has arrived, and I've begun to study it. Definitely a great
resource, recommended:

http://www.aes.org/publications/monographs/

Archontis, I'm still trying to understand how to express a set of HRTFS
as a SH series.
If I understand correctly, all HRTFS for a given ear can be expressed as
a function on the sphere, but it would be frequency dependent. So we'd
need an extra degree of freedom there, how does that tie in with
Ambisonics? One HRTF "balloon" per frequency bin?
Also, how do you express the inter-aural time delay conveniently (which,
as I've learned from Rozenn Nicol, depends not only on direction, but
also on frequency)?

Are there papers out there that describe this in detail?

Best,


Jörn

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Inventory of Ambisonics plug-ins/tools on OS X

[Politis Archontis]

Hi Daniel,

There is also the HOA suite for Puredata from IEM Graz (IEM_ambi).

BR,
Archontis
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] HRTF tinting

[Politis Archontis]

Thanx Marc for the pointer. 

Having a quick look, it seems that tinting is what I call sound-feld shaping, 
and also what Kronlachner has included in his ambisonic plugins as “directional 
loudness”. 
Essentially it describes the application of an arbitrary directional weighting 
to the sound-field before decoding, so that for example certain angular regions 
can be attenuated or modified compared to others. I have also used it for 
acoustic analysis purposes, to analyze what’s happening (sources/reflections) 
only on a certain angular region. The operations to do that arise naturally in 
the SH theory.

Richard in his patent seems to be applying the HRTFs as spatial weighting to 
the sound-field, in order to enhance various cues including elevation, a smart 
idea! Note that this is different from a direct ambisonic to binaural 
conversion, in which you take the HOA signals and you end up with two binaural 
signals, here you take the HOA signals, you weight the sound-field with the 
HRTFs, and you end up with 2 times the HOA signals, HRTF-weighted, one set for 
each ear. These can then be used for an enhanced ambisonic to binaural 
conversion.

I too fail to see though what this has to do with the OP headphones with the 4 
drivers inside…

Regards,
Archontis

> On 24 Feb 2016, at 19:58, Marc Lavallee  wrote:
> 
> On Wed, 24 Feb 2016 14:27:36 +
> Politis Archontis  wrote:
> 
>>> On 24 Feb 2016, at 16:04, Steven Boardman
>>>  wrote:
>>> 
>>> Yes. 
>>> 
>>> Richard Furse’s Blue Ripple Sound uses HRTF tinting in their
>>> decoder, it works very well. Although I think he may have a patent
>>> on it…:)
>>> 
>>> Steve
>>> 
>> 
>> Hi Steve,
>> 
>> what is HRTF tinting? I haven’t heard the name before..
>> 
>> Archontis
> 
> There's a description in the text of the patent:
> http://www.freepatentsonline.com/y2015/0262586.html
> http://www.freepatentsonline.com/20150262586.pdf
> 
> How multiple drivers can use HRTF tinting is unclear.
> (but as most patents, they are difficult to understand).
> --
> Marc
> 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] OSSIC Kickstarter Campaign Begins

[Politis Archontis]

> On 24 Feb 2016, at 16:04, Steven Boardman  wrote:
> 
> Yes. 
> 
> Richard Furse’s Blue Ripple Sound uses HRTF tinting in their decoder, it 
> works very well. Although I think he may have a patent on it…:)
> 
> Steve
> 

Hi Steve,

what is HRTF tinting? I haven’t heard the name before..

Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonic Decoder Design Resources

[Politis Archontis]

(sorry for the double post, my mail manager is misbehaving..)
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Ambisonic Decoder Design Resources

[Politis Archontis]

Hi Richard,

In terms of literature for general HOA decoding, the parts are all around, I 
personally like the papers by Franz Zotter cause they present all the relevant 
information in a clear and usable manner. The All-round Ambisonic Panning and 
Decoding paper in JAES I think has most of the stuff you’ll need, and not only 
for the specific method.

In terms of open code there are not so much stuff around. As the rest of the 
people mentioned, one is Aaron Heller’s ambisonic toolkit, in Matlab, which is 
very extensive! 
https://bitbucket.org/ambidecodertoolbox/adt.git

I have also released a Matlab library in a more educational manner, not so much 
for efficiency as for readability, you can find it if you want to have a look 
in:
http://research.spa.aalto.fi/projects/ambi-lib/ambi.html
This implements some recent decoders too, including the ALLRAD method mentioned 
before, and another interesting one called CSAD (constant-angular spread), 
which has constant energy vector magnitude and zero angular error. You can also 
check energy/velocity vector plots for any decoder that you choose to 
implement. Could be a good starting point for making your own code.

Additionally, a powerful but more complicated is the decoder generator of 
Davide Scaini, from University of Pompeau Fabra and Barcelona Media, which 
converges on an optimal decoder by iterative numerical optimization on the 
velocity/energy vectors, in Python. You can find the implementaiton in:
https://github.com/BarcelonaMedia-Audio/idhoa
 
Good luck with your project!
Archontis Politis



From: Sursound [sursound-boun...@music.vt.edu] on behalf of Dave Malham 
[dave.mal...@york.ac.uk]
Sent: 20 February 2016 20:56
To: Surround Sound discussion group
Subject: Re: [Sursound] Ambisonic Decoder Design Resources

I would definitely recommend Aaron's decoder toolkit and, in fact, the
whole oeuvre from the Benjamin-Lee-Heller group (BLaH). This can be found
at http://www.ai.sri.com/ajh/ambisonics/

 Dave

On 20 February 2016 at 17:58, Jörn Nettingsmeier <
netti...@stackingdwarves.net> wrote:

> On 02/20/2016 06:22 PM, Richard Graham wrote:
>
>> Hi Archontis,
>>
>> I would like to design decoders for 2d and 3d arrays, 1st through
>> 3rd order (at least), both regular and irregular arrays. C code
>> examples would be incredibly helpful as I plan to develop decoders
>> for Pd and Max.
>>
>
> Fons' ambdec is GPL, and it comes with a nice set of example setups.
> It's C++, but the way Fons uses it, it reads pretty much like plain C.
> After all, a dsp loop is a dsp loop...
>
> Most importantly, I’d like to figure out how to calculate these
>> coefficients myself and I am having trouble finding literature on
>> how to do that. I have reached out to a few folks who used their own
>> programs to calculate coefficients. Essentially, I’d like to build
>> my own program in the C programming language.
>>
>
> Aaron Heller has a Matlab/Octave toolkit out that will generate matrices
> for you, and it's completely open. But it relies on quite complex
> functions of the framework... His solutions are used at CCRMA, to
> great effect. Probably your best starting point.
>
> Richard has one but keeps it proprietary, Fons has one but also doesn't
> like to part with it (although he has been very generous about
> generating custom Ambdec setups for people, me included).
>
> For the nitty-gritty, check out the papers from recent Ambisonics
> symposia and the ICSA conferences. Talk to Thomas Musil from Graz for
> the old-school, lovingly hand-optimized matrix approach, or to Zotter et
> al. for the All-Rad approach that works for arbitrary setups but is
> quite complex and kind of brute-forceish. I can dig them up for you if you
> can't find them.
>
> Shortly, I will have access to a 16-channel ring on the horizontal
>> plane and a b-format cube. This system will be modular and
>> configurable into irregular setups, too.
>>
>
> nice! but unless you really need extremely high horizontal resolution for
> research purposes or a truly humongous listening area, a better use for all
> those speakers would be to make a more or less uniform 3D rig.
> gets you a nice dodecahedron for full third-order all around.
>
>
> best,
>
>
> jörn
>
>
>
> --
> Jörn Nettingsmeier
> Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487
>
> Meister für Veranstaltungstechnik (Bühne/Studio)
> Tonmeister VDT
>
> http://stackingdwarves.net
>
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here,
> edit account or options, view archives and so on.
>



--

As of 1st October 2012, I have retired from the University.

These are my own views and may or may not be shared by the University

Dave Malham
Honorary Fellow, Department of Music
The University of York
York YO10 5DD
UK

'Ambisonics - Component Imaging for Audio'
-- next part -

Re: [Sursound] Ambisonic Decoder Design Resources

[Politis Archontis]

Hi Richard,

there are numerous articles, do you intend to use anything more specific? 2D or 
3D decoding? regular or irregular setups? Or you are looking for the most 
general case?
(when you say in C, do you mean published code examples?)
And by low-orders do you mean first-order systems mainly (b-format)? Most 
practical systems at the moment target up to 3rd-order, which are not very high 
orders either, but it makes a difference.

Regards,
Archontis


On 20 Feb 2016, at 17:25, Richard Graham 
mailto:ri...@rickygraham.net>> wrote:

Hi all,

Can anyone recommend literature on designing ambisonic decoders (in C)? 
Specifically, articles that speak to the calculation of coefficients for 
low-order decoders.

All the best,

Ricky
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Advice on practicalities of 16-speaker half-spherical arrangement

[Politis Archontis]

Hi Michael,

What I had in mind are t-designs, which exist till a very high number of points 
and they have been used by various researchers in decoder design (and spherical 
acoustic processing in general). These are perfectly uniform for all practical 
purposes. And there exist other uniform arrangements on the sphere, mainly 
distinguished by the method that generates them, such as Fliege grids and 
Lebedev grids, all used in research for ambisonic-related research.

For irregular arrangements, I mentioned already the work of Zotter, Frank etc. 
from IEM Graz, all published. There is also the work of Epain and rest from 
Sydney university, with a decoder that has constant energy vector spread, also 
suitable for irregular setups. And there is also the work of Davide Scaini from 
Barcelona media and Pompeau Fabra, which targets irregular setups using 
numerical optimisation on the velocity and energy vectors, and who has also 
released python code to produce decoders.

Regards,
Archontis

> On 10 Feb 2016, at 00:40, Michael Chapman  wrote:
> 
>> Hi Martin,
>> 
>> HOA are not limited to icosahedra or only uniform arrangements (which
>> exist also beyond the 5 platonic solids).
> 
> Details, please.
> 
> Michael
> 
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Advice on practicalities of 16-speaker half-spherical arrangement

[Politis Archontis]

Hi Martin,

HOA are not limited to icosahedra or only uniform arrangements (which exist 
also beyond the 5 platonic solids). It would be very hard to claim them 
flexible or universal if that was the case. Uniform arrangements though 
simplify decoder design significantly.

As I mentioned you would need some more recent decoder methods for a 
hemisphere, however these methods are available and not so sensitive to speaker 
placement (no need for quests for magical geometries). My two cents, start with 
a regular horizontal ring, covering the case of horizontal decoding too, and 
spread the rest of the speakers evenly at the hemisphere, covering roughly 
equal area partitions. You can also fix one straight above and work with the 
rest. I believe the system should be capable of decoding 3rd order signals to 
half-space for such an arrangement, but that needs some checking.

I mentioned two publicly available resources for decoding to such a setup. I 
forgot to mention however that there exist also professional tools for that. I 
think Blue Ripple Sound’s decoder can handle hemispheres and irregular setups 
(haven’t tried it myself) and also the latest Spat from IRCAM has implemented 
the more advanced HOA decoders I mentioned.

Best,
Archontis

> On 10 Feb 2016, at 00:16, Martin Dupras  wrote:
> 
> I have no objection to using fewer than 16 speakers; it's just the
> maximum I have available to me.
> 
> The reason I had not considered the icosahedron vertices setup is
> because, according to the wikipedia page, it's capable of 2nd order,
> not 3rd order. Is that not the case?
> 
> Again from wikipedia: "Since stacked rings are somewhat wasteful at
> higher elevations and necessarily have a hole at the zenith, they have
> been largely surpassed by hemispherical layouts since mature methods
> for decoder generation have become available. As they are difficult to
> rig and require overhead points, hemispheres are usually found either
> in permanent installations or experimental studios, where expensive
> and visually intrusive trussing is not an issue."
> 
> That's the whole reason why I was considering a hemispherical setup
> originally. It talks about "mature methods for decoder generation" so
> I (wrongly, perhaps) assumed that there were capable decoders, and
> there would likely be already available "templates" or "typical
> scenarios" to use as a starting point.
> 
> Thanks,
> 
> - martin
> 
> 
> - martin
> 
> On 9 February 2016 at 22:02, Augustine Leudar  
> wrote:
>> I know Im treading on thin ice here around all these ferocious maths
>> guys... but might it be that there is not a suitable array that uses 16
>> speakers? I know if you have 16 speakers you will want to use all of them
>> but an Icosahedron is only 12 speakers (vertices) but it might be the best
>> option.
>> Also you could try ICST ambisonics plugins in max - they let you put the
>> speaker array in and it adjusts accordingly.
>> 
>> On 9 February 2016 at 21:51, Martin Dupras  wrote:
>> 
>>> Thanks for all the responses. Much appreciated.
>>> 
>>> I'll re-phrase the question in light of some of the answers I've been
>>> given.
>>> 
>>> I will be using third-order Ambisonics. My aim mostly is to experiment
>>> to get a good sense of what is possible with Ambisonics with height. I
>>> have experimented successfully with 8-channel planar Ambisonics some
>>> time ago. My primary intent is to spatialise multiple monophonic
>>> (synthesised) sources using 3rd-order Ambisonics spatialisation, and
>>> the playback of mixed sources (spatialised monphonic and stereophonic
>>> sources as well as B-format 4-channel recordings.)
>>> 
>>> At this moment in time, I have the opportunity to deploy (next week) a
>>> 16-channel array, so I would like some advice on a configuration that
>>> would be a good start to experiment with Ambisonics with height.
>>> Someone suggested that I consult the wikipedia page on Ambisonics.
>>> That is indeed where I got the idea that an "upper hemisphere" setup
>>> might be suitable, since I only have on this occasion 16 speakers.
>>> There is however no suggestion as to what a suitable hemispherical
>>> configuration might be for a 16-speaker array, which is why I asked my
>>> original question.
>>> 
>>> So let me ask a new question. Given the constraint that I can only use
>>> 16 speakers at the moment, and that I need to deploy this next week,
>>> can somehow point me in the direction of what might be a suitable and
>>> reasonable geometric configuration to try out? It seems to me that the
>>> only really practical options here are two stacked rings (stacked
>>> octagons) or a hemisphere. I would have thought that the hemisphere
>>> would be the better choice, and in my scenario, a full lighting rig
>>> allows me theoretically speaking to have speakers at the required
>>> positions.
>>> 
>>> Again, thank you for all the responses.
>>> 
>>> - martin
>>> 
>>> 
>>> On 8 February 2016 at 15:19, Martin Dupras  wrote:
 Hi,

Re: [Sursound] Advice on practicalities of 16-speaker half-spherical arrangement

[Politis Archontis]

Hi Martin,

one note on the arrangement, as far as I know, traditional ambisonic decoding 
won’t work on hemispherical setups (due to the partial coverage of the sphere 
by the speakers). You may have to use more recent/advanced methods to get 
decoding matrices, such as the energy-preserving ambisonic decoding with 
modified basis functions (EPAD), or the hybrid ambisonic/VBAP All-round 
ambisonic decoding (ALLRAD), published by Franz Zotter (and collaborators) from 
University of Graz.

I know of two usable available implementations of such decoder matrix 
calculators, one is a compact Matlab/Octave library I made available recently:

http://research.spa.aalto.fi/projects/ambi-lib/ambi.html
https://github.com/polarch/Higher-Order-Ambisonics

the other is Aaron Heller’s ambisonic decoding toolbox, more extended than 
mine, which includes scripts to produce VST plugins from the decoding matrices 
using Faust, which may be more suitable for your workflow: 

https://bitbucket.org/ambidecodertoolbox/adt.git

Best,
Archontis


> On 08 Feb 2016, at 17:19, Martin Dupras  wrote:
> 
> Hi,
> 
> I'm intending to try setting up a 16-speaker Ambisonics array next
> week in a small TV studio. I'm trying to figure out the practical
> arrangements for setting up the speakers. I was wondering if anyone
> with experience might be able to offer some advice or point me in the
> right direction?
> 
> What I'm planning at the moment is a half-sphere arrangement which
> would likely consist of:
> 
> - 8 speakers in a circle of radius 2m at a height of approximately 1.6m
> - 6 speakers in a smaller circle at an elevation of 45 degrees
> - 2 speakers at an elevation of approximately 75 degrees
> 
> Alternatively, I would be happy with an arrangement similar to the
> first 16-speakers in this diagramme:
> http://www.matthiaskronlachner.com/wp-content/uploads/2013/10/loudspeaker-plan-observatory.jpg
> 
> I've been trying to find out if there is a convention or "most usual"
> arrangement but couldn't find anything. I'm not particularly attached
> to the actual arrangement, I just want to find an arrangement that
> will work well enough with 16 speakers. Any advice?
> 
> The other thing I would welcome is advice on how to mount the speakers
> to lighting rigs in a manner that is practical enough to offer some
> good compromise between precision and ease of setup. I believe the
> speakers we'll be using for the upper tiers will be Genelec 8060s.
> 
> Many thanks. Any advice will be greatly appreciated!
> 
> Cheers,
> 
> - martin
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Never do electronic in public.

[Politis Archontis]

Hi David,

I agree completely with what you say apart from a single point: This dichotomy 
is the exact reason why, in my opinion, proper binaural rendering is crucial. 
And I don’t think music is the issue here anyway (actually most of the people I 
know listening to music with loudspeakers at their places, are perfectly happy 
with a mono playback, or preference to a “diffuse” room sound, since they will 
rarely sit down to listen to music specifically). The issue is that all these 
emerging VR and AR technologies, independently of if they will be successful or 
not, should be designed from the bottom-up to deliver consistent and natural 
cues, especially if you are mixing reality with the virtual. Otherwise, you end 
up with a mess of audiovisual conflicting cues, which may be exhausting and 
disorientating. 

About music listening however, there is a chance that with headphones you can 
teach all these young people some proper listening habits! since if the 
processor binauralizes and externalizes correctly the stereo or multichannel 
recording, then they won’t be able to escape the sweet spot :-).

Regards,
Archontis

> On 02 Feb 2016, at 09:37, David Pickett  wrote:
> 
> At 07:42 02-02-16, umashankar manthravadi wrote:
> 
> >For the moment, I would like video kept out of motion tracked audio on
> >headphones. I want this to be a system where I can sit in a rocking
> >chair and listen to music of many kinds on a good pair of headphones
> >and the headtracking is only to reinforce the audio image.
> 
> Although I use a pair of Sennheiser HD600 headphones almost every day for 
> making and editing recordings, I cant get at all excited about having to wear 
> headphones to listen to music for pleasure, and unlike mulitchannel 
> loudspeaker surround, I personally regard the effort to perfect binaural 
> sound on headphones as a waste of time and energy.  This may be my loss; but 
> it is my honest opinion, and others are entitled to disagree.
> 
> Most young people that I see in the city where I live seem to have earbuds 
> permanently in their ears; they are presumably listening to normal stereo 
> recordings made for loudspeakers and seem to be quite happy with that format 
> and MP3 quality.  I fear it might be rather disorienting for such people to 
> be listening to recordings that give them cues that contradict those they 
> would get from the environment where they are actually standing or walking.  
> Unlike Umashankar at home in his rocking chair, this is at least rather 
> dichotomous, or whatever the word is.
> 
> David
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Never do math in public, or my take on explaining B-format to binaural

[Politis Archontis]

> On 30 Jan 2016, at 00:05, Richard Lee  wrote:
> 
> Just to bring everyone down to earth ..
> 
> There are two easily reproduced experiments first carried out by prominent 
> members of this group which put these effects into perspective.  They are 
> the
> 
> Greene/Lee Neckbrace
> and
> Malham/Van Gogh Experiment
> ...
> 
> SPECIAL OFFER
> 
> Send $500 in used bank notes to me at Cooktown Recording and Ambisonic 
> Productions mentioning Sursound, for a sample Greene/Lee Neckbrace and 
> Diamond Encrusted Malham/Van Gogh cap. Golden Pinnae are an extra cost 
> option on the last item.  No Confederate money please.
> 

HI Richard,

man I didn’t understood almost anything from your message, but I opted for a 
Greene/Lee Neckbrace and a Diamond Encrusted Malham/Van Gogh cap at the 
aforementioned address (no Golden Pinnae necessary :-)!

What I did understand is that head-tracking is important, we know that, and 
there’s not much sense in binaural reproduction otherwise. 

What I was trying though to say I guess in my previous message is that, if we 
believe that the HRTFs capture all the information we need, then we better use 
them, it’s not that heavy nowadays, even in real-time.

And my point about first order is mainly that, if we listen to a sound from a 
certain direction through the respective HRTF with headphones, and we listen to 
its first -order encoded and decoded version, they sound quite different, 
spectrally and spatially.

Regards,
Archontis

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Never do math in public, or my take on explaining B-format to binaural

[Politis Archontis]

HI Bo-Erik,

Sorry I wasn’t very clear on my comments about the HRTF order, Jörn covered it 
much better!

You need very high orders if you want to preserve HRTFs as they are when you 
measure them, and the order that you can expand them depends on how many points 
you measure them around the head. For a typical measurement setup this days 
with a resolution of ~5deg elevation and azimuth, that corresponds to an 
effective ambisonic order of ~15, which is enough.
If you want to play a virtual source at a direction of some HRTF, you just need 
to convolve them with the respective ones, and you can do that either by using 
directly the measured (or modelled one) or do it in the ambisonic domain; this 
is not a heavy operation, it’s just a few tens to a few hundreds 
multiplications and additions per frequency band for all these orders. In any 
case, for binaural panning and spatialisation, no need to discard any of that 
information and lose accuracy. And I don’t think it’s just strictly about 
distinguishing the direction between two closely spaced sounds, it’s about 
delivering correct binaural cues in general, correct fluctuations of ITDs and 
ILDs in complex sound scenes with many sounds including early reflections and 
late reverberation, which are important for reverberation perception, apparent 
source width and externalization.

When however you want to reproduce or binauralize a recorded sound scene, the 
available orders are limited by the (physics of) the microphone array. In this 
case no matter your HRTF resolution, it’s always going to be limited by the 
order of the recording, blurring the binaural cues gradually at higher 
frequencies. That may or may not be a problem depending on the application. One 
kind of solution is given by parametric processing of the FOA/HOA signals (such 
as HARPEX or DirAC), which restore somewhat the correct cues, by “de-blurring” 
the directional components in the recording. That’s the field I am working on 
too, and even though I think they are impressively effective, they need a lot 
of care in implementation, and they have their own issues to address (that’s 
why it’s a field of current research).

But regarding VR or AR applications or virtual monitoring of other multichannel 
setups, where you want maximum sharpness (and where you can employ the full 
HRTFs), I don’t think going through a FOA representation is a good solution, as 
I find the blurring of the virtual sources to be very high compared to when 
using directly the HRTFs (basically a point source is spread in space around 
you with an almost cardioid or supercardioid shape). Another factor that is 
overlooked sometimes is that the same blurring affects the sense of 
reverberation and externalisation, as it “correlates” more than necessary the 
reverberant sound at the two ears..

Best regards,
Archontis Politis


On 28 Jan 2016, at 23:12, Bo-Erik Sandholm 
mailto:bosses...@gmail.com>> wrote:

I do understand that HOA can represent resolution of directivity in the
mathematic domain better than FOA.
But I am starting to suspect we are overworking something when we are
talking of order 8 to 15?
Is it realistic to even think of measuring individual HRTF response with
that angle resolution? And is it even neccessary when we know the
adaptability of the auditory system?

As stereo works good enough over 45 degrees with 2 speakers and correct
psycho acoustic setup and a good recording are we not aiming for a overkill
system?

As a normal guy without training in listening for direction of sound
sources I suspect I cannot really pinpoint many things in more than +-10
degrees without visual cues.

I remember old discussion results about ideal number of loudspeakers for
horizontal FOA replay being 6 speakers.

My goal is to have a device that can play through headphones a stereo or
FOA recording and give me a minimum experience of listening to a stereo
system or FOA setup with out of head sound and a stable position of the
soundstage.

I am not certain this is relevant in this discussion thread as we probably
have different views of the goals and the path to the goals.

Bo-Erik

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Never do math in public, or my take on explaining B-format to binaural

[Politis Archontis]

Hi Fons,

___
From: Sursound [sursound-boun...@music.vt.edu] on behalf of Fons Adriaensen 
[f...@linuxaudio.org]
Sent: 27 January 2016 23:58
To: sursound@music.vt.edu
Subject: Re: [Sursound] Never do math in public, or my take on explaining 
B-format to binaural

> plus independent decorators/reverberators per HOA channel,
> with different decays at different frequencies, for the
> late part. The late part filters require a further tuning
> stage though, to match the ‘sinc’ like binaural-coherence
> of left and right ear signals in diffuse sound. I have
> found that this matching improves significantly
> externalisation, and sounds more natural.

I'd be much interested to learn more about that...

Ciao,

--
FA


sure I can expand on that, is it ok if I contact you or send you some text 
off-list?

I am wondering though, as I understood, you are interested in the case of 
converting to binaural other multichannel formats with dynamic head-tracking. 
In this case since you have discrete sources you are binauralizaing, with some 
added room effect, is there a reason to go through a second order ambisonic 
system, instead of using directly HRTFs with some interpolation, plus a 
binaural late reverb? Don't you blur significantly (well, up to a second-order 
blurring) the spatial resolution of the HRTFs, considering again that we need 
orders of 8-15 for fullband reconstruction of HRTFs at mid and high frequencies?

Regards,
Archontis
___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Never do math in public, or my take on explaining B-format to binaural

[Politis Archontis]

Hi Fons,

these are great points you raise here!


On 27 Jan 2016, at 00:13, Fons Adriaensen 
mailto:f...@linuxaudio.org>> wrote:


I've been reading this thread with much interest, as it is exactly
about the topic I've been working on for the last two months.

* Most HRIR sets have an LF response that is almost certainly
not correct. Up to a few hundred Hz it should be flat. One
essential step in the preparation is to fix this. How this
is done best depends on the particular data set. If this is
done correctly you can trim the IRs to a few ms without any
adverse effect.

Something that has worked well for me is to replace the (unmeasurable) low 
frequency HRTF with the response of a spherical head model, including the phase 
response, normalised to match the rest of the measured HRTF. One can go more 
exact and find a “personalised” sphere radius, with a more realistic 
‘off-horizontal’ position of the two receivers (ears) on the sphere. This 
approach has been used too to fit a sphere on a quick Kinect head scan, for 
personalised ITD estimation, with good results:
"Hannes Gamper, Mark R. P. Thomas, and Ivan J. Tashev, Anthropometric 
Parameterisation of a Spherical Scatterer ITD Model with Arbitrary Ear 
Angles, in 
Proc. IEEE Workshop on Applications of Signal Processing to Audio and Acoustics 
(WASPAA)”

* All content derived from non-surround sources (e.g.
plain stereo or 5.1) requires some 'room sound' to work
well. Externalisation seems to depend on having early
reflections from different directions (which would allow
the brain to compare their spectra). Generating such
room sound can be done in the AMB domain. What exactly
is required and how to do that efficiently is my current
research problem.

Again something that has worked well for me is to use a few HOA-encoded point 
sources as early-reflections, plus independent decorators/reverberators per HOA 
channel, with different decays at different frequencies, for the late part. The 
late part filters require a further tuning stage though, to match the ‘sinc’ 
like binaural-coherence of left and right ear signals in diffuse sound. I have 
found that this matching improves significantly externalisation, and sounds 
more natural.

Regards,
Archontis
-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Never do math in public, or my take on explaining B-format to binaural

[Politis Archontis]

HI,

yes there are multiple works from many reearchers on spherical harmonic 
expansion of HRTFs (Evans, Duraiswami, Dylan Menzies and others), Rozenn 
Nicol’s is a very thorough one, and the last (probably) on the list are two 
nice papers from Romigh and Brungart in IEEE’s Journal of Selected Topics on 
Signal Processing: Spatial Audio issue (JSTSP, 9(5), August 2015).

I think the important facts are that in this way you get a nice continuous 
representation of HRTFs, which you can interpolate anywhere smoothly and with 
an interpolation that uses all your data points (measurements). You also get a 
natural ambisonic binauralization operation, the binaural signal is simply the 
product-and-sum of the HRTF coefficients with the HOA signals (in the frequency 
domain, or convolutions equivalently in the time domain).
You need a lot of coefficients at high orders (~15) to capture all the spatial 
variation of the HRTFs, but that is frequency dependent, at low frequencies the 
few first orders are enough to capture the shape. 

Regards,
Archontis


> On 27 Jan 2016, at 11:02, florian.came...@orf.at wrote:
> 
> Hello,
> 
> may I point you to the AES Monograph on Binaural Technology by Rozenn Nicol,
> published on 2010. Rozenn has nicely summarised most of the issues which have 
> been discussed
> here lately, and she provides an extensive list of references (more than 
> 200!). Well worth reading
> (35$ for AES members).
> 
> Best wishes, Florian
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Never do math in public, or my take on explaining B-format to binaural

[Politis Archontis]

Hi Fons,

I think I see your point, but I still believe they are equivalent in any case? 
In the end by having a lower-order sound field recording (eg B-format), than 
the expansion order of the HRTFs at some frequency, the spatial product is 
going to be limited by the sound-field order. If there is a decoding stage in 
the middle, again a fully directional sound will be spread to multiple speakers 
with an amplitude distribution determined by the order of the decoder, and will 
be convolved with multiple HRTFs from the respective virtual directions, 
essentially limiting the spatial resolution of the HRTFs in the same way and to 
the same order. Are you referring to any other effects of going through the 
virtual loudspeaker case that would affect that?

I see though the point of Jorn and Joseph’s of using dual-band (or any other 
more perceptually-tuned) decoder. But i think that can be done also in the 
spherical harmonic domain with no need for virtual decoding, as it would be 
essentially a weighting of the components of each order (or a convolution/ 
smoothing over the sphere).

Regards,
Archontis


> On 27 Jan 2016, at 00:27, Fons Adriaensen  wrote:
> 
> On Tue, Jan 26, 2016 at 10:05:07PM +, Politis Archontis wrote:
> 
>> yes that is correct. I think however that the virtual loudspeaker
>> stage is unnecessary. It is equivalent if you expand the left and
>> right HRTFs into spherical harmonics ...
> 
> True, but the problem with this is that it requires quite high
> order at HF (which is where all the binaural magic happens).
> 
> But then you could say that any small set of virtual speakers
> would be inaccurate as well as it can't capture the high order
> dependency of the HRIR. This is true of course. The only reason
> why it works is that it relies on amplitude or ambisonic panning
> between the virtual speakers, just as reproduction via real
> speakers does.
> 
> Ciao,
> 
> -- 
> FA
> 
> A world of exhaustive, reliable metadata would be an utopia.
> It's also a pipe-dream, founded on self-delusion, nerd hubris
> and hysterically inflated market opportunities. (Cory Doctorow)
> 
> ___
> Sursound mailing list
> Sursound@music.vt.edu
> https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
> account or options, view archives and so on.

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

Re: [Sursound] Never do math in public, or my take on explaining B-format to binaural

[Politis Archontis]

Hi Jorn,

yes that is correct. I think however that the virtual loudspeaker stage is 
unnecessary. It is equivalent if you expand the left and right HRTFs into 
spherical harmonics and multiply their coefficients (in the frequency domain) 
directly with the coefficients of the sound scene (which in the 1st-order case 
is the B-format recording). This is simpler and more elegant I think. Taking 
the IFFT of each coefficient of the HRTFs, you end up with an FIR filter that 
maps the respective HOA signal to its binaural output, hence as you said it's 
always 2*(HOA channels) no matter what. Arbitrary rotations can be done on the 
HOA signals before the HOA-to-binaural filters, so head-tracking is perfectly 
possible.

Best,
Archontis


From: Sursound [sursound-boun...@music.vt.edu] on behalf of Jörn Nettingsmeier 
[netti...@stackingdwarves.net]
Sent: 26 January 2016 22:52
To: sursound@music.vt.edu
Subject: [Sursound] Never do math in public, or my take on explaining B-format 
to binaural

I think the 8 impulses are used differently. I'm scared of trying to
explain something of which my own understanding is somewhat hazy, but
here it goes: please correct me ruthlessly. Even if in the end I wish
I'd never been born, there might be something to learn from the
resulting discussion :)

W goes to loudspeaker LS1, LS2, ..., LSn.
Same for X, Y, and Z.

Each LSn then goes both to left ear and right ear.

So you start with a 4 to n matrix, feeding into an n to 2 matrix. The
component-to-speaker convolutions and the speaker-to-ear convolutions
(the HRTFs) are constant.

Convolution and mixing are both linear, time-invariant operations. That
means they can be performed in any order and the result will be
identical. I guess in math terms they are transitive and associative, so
that (a # X) + (b # X) is the same as (a + b) # X, and a # b # c is the
same as a # (b # c), where "#" means convolution.

So the convolution steps can be pre-computed as follows, where DEC(N,m)
is the decoding coefficient of component N to loudspeaker m, expressed
as convolution with a dirac pulse of the appropriate value:

L = W # DEC(W,LS1) # HRTF(L,LS1) + ... + W # DEC(W,LSn) # HRTF(L,LSn)
   + X # DEC(X,LS1) # HRTF(L,LS1) + ... + X # DEC(X,LSn) # HRTF(L,LSn)
   + Y # ...
   + Z # ...

(same for R)

which can be expressed as

L = W # ( (DEC(W,LS1) # HRTF(L,LS1) + ... + DEC(W,LSn) # HRTF(L,LSn) )
   + X # ...
   + Y # ...
   + Z # ...

(same for R).

Note that everything in brackets is now constant and can be folded into
a single convolution kernel.

That means you can, for first order, reduce the problem to 8
convolutions, going from {WXYZ} to {LR} directly. The complexity is
constant no matter how many virtual loudspeakers you use.

Of course, that does not take into account dual-band decoding. But if we
express the cross-over filters as another convolution and split the
decoding matrix into a hf and lf part, we can also throw both halves of
the decoder together and do everything in one go.

For nth order, you have (n-1)² * 2 convolutions to handle.

For head-tracking, the virtual loudspeakers would move with the head (so
that we don't have to swap HRTFs), and the Ambisonic signal would be
counter-rotated accordingly. Of course that gets the torso reflections
slightly wrong as it assumes the whole upper body moves, rather than
just the neck, but I guess it's a start.



--
Jörn Nettingsmeier
Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487

Meister für Veranstaltungstechnik (Bühne/Studio)
Tonmeister VDT

http://stackingdwarves.net

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

[Sursound] HOA-related public Matlab/Octave code

[Politis Archontis]

Hi,

Please excuse the spam, but I thought that could be of interest to some people 
in the list.

My name is Archontis Politis, I am a doctoral student in the spatial sound 
research group of Aalto University in Finland, working on parametric spatial 
sound recording/reproduction techniques, and combinations with 
non-parametric/ambisonic approaches.
I’m wrapping up my thesis and in the meantime I tidied up some Matlab/Octave 
code I had, and pushed it out publicly. These include:

- A compact Higher-order ambisonics library that implements a few recent 
decoding approaches, along with performance evaluation in terms of velocity and 
energy vectors

- A Spherical Harmonic Transform library, which includes rotations, 
convolutions and coupling coefficients of spherical functions, for both real 
and complex SHs.

- An amplitude panning library that implement VBAP, VBIP, VBAP & VBIP-based 
spreading, and some variants

- An array simulator for arbitrary directional microphones and arrays of 
microphone on and around a rigid scatterer (spherical or cylindrical)

The HOA library is only for 3D layouts at the moment, and it returns decoding 
matrices for two classic decoding approaches, plus the recent energy-preserving 
decoding by (Zotter, Pomberger, Noisterning, 2012), the All-round ambisonic 
decoding by (Zotter and Frank, 2012), and the Constant Angular Spread Decoding 
by (Epain, Jin and Zotter, 2014). It also includes rotations of recordings, 
max-rE weighting and frequency-dependent decoding as options.

The array simulator can be used to obtain impulse/frequency responses of arrays 
that deliver HOA signals, and study the effect of the array to their frequency 
response/aliasing etc.

These can be found at:

https://github.com/polarch/

with respective showcase/documentation found at:

http://research.spa.aalto.fi/projects/sht-lib/sht.html
http://research.spa.aalto.fi/projects/vbap-lib/vbap.html
http://research.spa.aalto.fi/projects/ambi-lib/ambi.html
http://research.spa.aalto.fi/projects/arraysim-lib/arraysim.html

I hope something here could prove useful to you. In case you use them, any 
feedback, corrections, etc. is mostly welcome!

Best regards,

Archontis Politis
Doctoral Researcher
Department of Signal Processing and Acoustics
Aalto University, Finland






-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.

[Sursound] HOA-related public Matlab/Octave code

[Politis Archontis]

Hi,

Please excuse the spam, but I thought that could be of interest to some people 
in the list.

My name is Archontis Politis, I am a doctoral student in the spatial sound 
research group of Aalto University in Finland, working on parametric spatial 
sound recording/reproduction techniques, and combinations with 
non-parametric/ambisonic approaches.
I’m wrapping up my thesis and in the meantime I tidied up some Matlab/Octave 
code I had, and pushed it out publicly. These include:

- A compact Higher-order ambisonics library that implements a few recent 
decoding approaches, along with performance evaluation in terms of velocity and 
energy vectors

- A Spherical Harmonic Transform library, which includes rotations, 
convolutions and coupling coefficients of spherical functions, for both real 
and complex SHs.

- An amplitude panning library that implement VBAP, VBIP, VBAP & VBIP-based 
spreading, and some variants

- An array simulator for arbitrary directional microphones and arrays of 
microphone on and around a rigid scatterer (spherical or cylindrical)

The HOA library is only for 3D layouts at the moment, and it returns decoding 
matrices for two classic decoding approaches, plus the recent energy-preserving 
decoding by (Zotter, Pomberger, Noisterning, 2012), the All-round ambisonic 
decoding by (Zotter and Frank, 2012), and the Constant Angular Spread Decoding 
by (Epain, Jin and Zotter, 2014). It also includes rotations of recordings, 
max-rE weighting and frequency-dependent decoding as options.

The array simulator can be used to obtain impulse/frequency responses of arrays 
that deliver HOA signals, and study the effect of the array to their frequency 
response/aliasing etc.

These can be found at:

https://github.com/polarch/

with respective showcase/documentation found at:

http://research.spa.aalto.fi/projects/sht-lib/sht.html
http://research.spa.aalto.fi/projects/vbap-lib/vbap.html
http://research.spa.aalto.fi/projects/ambi-lib/ambi.html
http://research.spa.aalto.fi/projects/arraysim-lib/arraysim.html

I hope something here could prove useful to you. In case you use them, any 
feedback, corrections, etc. is mostly welcome!

Best regards,

Archontis Politis
Doctoral Researcher
Department of Signal Processing and Acoustics
Aalto University, Finland

-- next part --
An HTML attachment was scrubbed...
URL: 

___
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit 
account or options, view archives and so on.