Dear Franz,
Good evening Franz,
firstly: many thanks for your competent and useful response!
So trying to summarize up things a bit, I think it could make sense to
separate two cases:
If you are in the sweet spot (which includes binaural decoding) or at
least being close, you can have very good results using 3rd order HOA
(TOA).
If using HOA as < area format > (most people technically being out of
the sweet spot), we will agree that 4th order and 5th order will be
still and noticeably better than TOA. (4th order has been tested as
cinema audio format, for example.)
You wrote (refering to 5th order):
“It is as practical as 3-stereo:
We have 30° degrees of angular separation, so very-close-to-stereo
channel separation. At the moment the best we have for the practice
with large audiences and studio. The question is where the practice
should focus to. “
Comment: For 5th order reproduction we would need a minimum of 12
loudspeakers, ideally in a regular layout in 30° steps.
(For simplification: let us restrict the discussion to the 2D case.)
The speaker separation of 30° means that the maximum angular error of
into the SF encoded sources/positions should not exceed 15°, during
reproduction. ( Because the closest speaker can’t be further away. )
So maybe this would imply that the < average > error should be around
7.5°, if we are talking about higher frequencies?
For TOA the analogue values should be 22.5° and 11.25°, respectively.
In practice things might work actually quite a bit better, especially
in or near the sweet spot.
On the other hand figures above could be too optimistic if getting
close to the speakers, of course.
You wrote further:
“The newest binaural decoders (BinauralDecoder) show that 5th order
is typically not different to 7th, for dense audio scenes it 3rd is
also as good as 7th (Lee 2019, AES IAA). Sidelobe suppression does
not appear relevant in the sweet spot of 3rd order and above, as we
could prove by listening experiments in loudspeaker playback (Frank
2017, AES Conv.).”
Are these “newest binaural decoders” somehow available for being used
in software and/or CE projects?
I know that you are already offering a plugin:
https://plugins.iem.at/docs/plugindescriptions/#binauraldecoder
For the ones of us who enjoy reading papers with plenty of formulas 😉,
here a link to the corresponding “paper”:
https://asa.scitation.org/doi/10.1121/1.5040489
So this TAC method (binaural decoding of ambisonics, realized for
orders 1 to 5) seems to be “quite interesting”, but maybe also to be
used in some wider form...
Best regards
Stefan
- - -
Citando Franz Zotter <zot...@iem.at>:
Dear Stefan, dear list,
Am 06.03.2019 um 21:16 schrieb Stefan Schreiber:
I don’t want to come over as being too negative about new and
progressive projects, but “still” have some problems with “5th
order recordings”.
Many practical HOA decoders in the wild are currently restricted
to 3rd order. 4th order is seen as Master Format for VR.
(Guidelines of VR Industry Forum.)
Well, it is e.g. a good idea to encode either close mic recordings
or spaced main microphone arrays in 5th order, like ESMA, which goes
without a lot of trouble. The free-and-open plugins under
matthiaskronlachner.com, plugins.iem.at, or
research.spa.aalto.fi/projects/sparta_vsts[1] (i.e. multiencoder or
ambix_encode_i6_o5 or AmbiENC) would allow you to do so.
Would we really need 5th order if the decoders would be a bit more
optimized, by the way?
For large audience area I think it is necessary to get enough
angular separation (in level) in order to somehow preserve the
directions.
We listened to the awarded pieces in a system for 300 seats in Cologne(!).
Nothing personal again, but I think you are using rE decoders. This
might be acceptable in the loudspeaker case and using HOA, but
applying headphones and binaural decoders I still would recommend
dual-band decoders. Because you aredefinitively in the sweet spot...
Yes, but no... The newest binaural decoders (BinauralDecoder) show
that 5th order is typically not different to 7th, for dense audio
scenes it 3rd is also as good as 7th (Lee 2019, AES IAA). Sidelobe
suppression does not appear relevant in the sweet spot of 3rd order
and above, as we could prove by listening experiments in loudspeaker
playback (Frank 2017, AES Conv.).
It is anyway easy to try out: the binaural decoder comes without rE
weights and appeared to sound best this way, but there are no formal
proofs done, as far as I know.
(And maybe also NFC, as Aaron Heller might insist.)
If the audience is large, NFC is rather destroying things elsewhere,
while maybe 5% of the seats around the center might get improvements.
Anyway, for higher orders this is better done by bass management.
Of course you could use the 5th order mix as master format, and
less in practice. (The only problem I have with this approach is
that even the eigenmike can’t record 5th order. Of course you can
still mix your HOA main microphone and spot mikes.)
It is as practical as 3-stereo:
We have 30° degrees of angular separation, so very-close-to-stereo
channel separation. At the moment the best we have for the practice
with large audiences and studio. The question is where the practice
should focus to. In my opinion it should include.
For home and studio applications, 3rd order would of course be
pretty fine as well.
And for higher-order spherical microphones 3rd order is pretty fine as well.
All better than first order without slight directional
postprocessing (which can be a static decode and higher-order
re-encode of the source directions! - if the adaptive/parametric
solutions don't do their job well enough).
So maybe all things are just fine, and why not being
forward-looking... (How many 5th order capable playback systems do
we have? I mean: on this planet? I would be surprised if we should
“count” more than 100 installations for 22.2, as well... Which
probably is used as master format, accordingly.)
"Forward-looking" would to me mean 7th order, as your P.S.1 :D
- But anyway in our practice in the IEM CUBE and from small tours
around with some of the productions, 5th order is clearly and
noticeably better than 3rd. And I guess it will be satisfactory for
a while.
I am in a tough place here, because many people will now show up
and < insist > that FOA recordings would be just fine. “Perfect
spatial sound forever”, so to speak.
Well... not so well for the large-audience situation if used as
puristic-and-only means of recording with little/no post production.
Then, I am afraid, most people will only hear the loudspeakers
closest to them as directions, but not really pronounced.
For some ambiences it works, though, but you can't rely on 1st order
when playing back to 100-200 people. It's a pain, from my experience.
Of course the effort of multi-band for first-order is appreciated,
but it too restrictive in applications. It would be difficult to
recognize the intended qualities to a larger audience, therefore
somewhat odd to play back in an award ceremony.
The guideline saying 5th order however does not mean that one could
not try otherwise. One should be aware then that unprocessed and
pure 1st order might get poorly rated doe to its huge directional
blur.
Compass is a hint, but I would first check HARPEX which is probably
easier to use.
Best regards
Franz
(Running away to get some bite, as our British colleagues say...)
Best regards
Stefan
P.S.: 7th order recordings/mixes would have been even better! 😎
P.S. 2: Being more constructive:
Maybe using COMPASS could help to achieve some true 5th order
result. (As long as you can reproduce this...)
Further work (for next year): 7th order?! (Losing patience...)
- - - -
Citando Franz Zotter <zot...@iem.at>:
Dear colleagues & friends,
We are happy to announce that the 5th Int. Conference on Spatial
Audio will host Europe’s Third Student 3D Audio Production
Competition, Sept. 26th to 28th 2019 in Ilmenau, Germany,
https://vdt-icsa.de/en/2019/ .
Call: Students who are interested in 3D-spatialized sound and work
on spatial music, spatial audio productions or recordings are
invited to participate by submitting short works (4min/11min) to
the third Student 3D Audio Production Competition.
There are three submission categories:
(1) contemporary / computer music (11min)
(2) audio drama / documentary / soundscapes (4min)
(3) music recording / studio production (4min);
For the first time, category (1) permits more common
contemporary/academic composition formats of eleven minutes.
The format requirements for preparing a submission and tools to
check are given here https://iaem.at/ambisonics/s3dapc/2019 and of
course you can get in touch if there are questions: <s3d...@iem.at>
The submission portal
https://easychair.org/conferences/?conf=s3dapc2019
is open until the end of June 2019.
An international jury will assess the submissions concerning the
creative and technical quality, providing feedback to all the
participants of the competition. Based on the resulting ranking,
the finalists will be selected for presentation at 5th ICSA. The
finalists will be notified and requested to attend the finals at
ICSA in Ilmenau in order to receive their award, the jury’s
comments, and say a few words about their submissions.
The finalists' work will be presented during the finals and awards
celebration, played back with powerful hemispherical loudspeaker
setup with about 30 loudspeakers.
Best regards,
Franz Zotter
Matthias Frank
Daniel Rudrich
_______________________________________________
Sursound mailing list
surso...@music.vt.eduhttps://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/20190306/841ebad3/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.
--
Ass.Prof. Dr. Franz Zotter
http://iaem.at/Members/zotter
Institut für Elektronische Musik und Akustik
Universität für Musik und darstellende Kunst
Tel.: +43(0)316/389 - 3382
Mob.:+43(0)650/9688373
Inffeldgasse 10
A - 8010 Graz
_______________________________________________
Sursound mailing list
surso...@music.vt.eduhttps://mail.music.vt.edu/mailman/listinfo/sursound -
unsubscribe here, edit account or options, view archives and so on.
----- Fim de mensagem reenviada -----
Ligações:
---------
[1] http://research.spa.aalto.fi/projects/sparta_vsts
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20190308/dddedfc9/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.