Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Ralf Mardorf]

PPS:
On Sat, 30 Jun 2018 09:55:41 +0200, Ralf Mardorf wrote:
>You neither described why it didn't work

If you don't know the reason, you at least could mention that you build
something this or by another way and that you run it this or that way,
without getting any messages.

-- 
pacman -Q linux{,-rt{,-securityink,-cornflower,-pussytoes}}|cut -d\  -f2
4.17.3-1
4.16.15_rt7-1
4.16.12_rt5-1
4.16.8_rt3-1
4.14.34_rt27-1
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Ralf Mardorf]

On Sat, 30 Jun 2018 09:49:48 +0200, Ralf Mardorf wrote:
>To get something working that didn't work in the first place, while it
>likely should work in the first place, you "hack"ed it in a way that it
>still doesn't work.

PS:

You neither described why it didn't work, nor what "hack"s you are
using, that it still doesn't work.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Ralf Mardorf]

Hi,

I'm not really able to help you with this specific SC issue, however,
you should describe the issues you experienced more detailed.

An/the SC package of what Suse release is broken? In what way the
package is broken? Just imagine usage of a third party repository
and e.g. a soname issues.

To get something working that didn't work in the first place, while it
likely should work in the first place, you "hack"ed it in a way that it
still doesn't work. I guess it's self-explaining that you can't expect
much help to solve this issue, by providing entirely ambiguous
information.

Regards,
Ralf
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Tim]

On 06/26/2018 04:25 PM, Tim wrote:



On 06/26/2018 03:55 PM, Hans Wilmers wrote:

On 06/26/2018 08:32 PM, Spencer Jackson wrote:



I don't know of anyone really working on polyphonic pitch recognition in
the open source world. I think Bayesian filtering of some kind though
would be compelling. Perhaps some of the work from ISSE
(http://isse.sourceforge.net/) could be used and made realtime.



There is a SuperCollider plugin by Nick Collins called PolyPitch, which
does what the name suggests.
The source is GPL, and available here:
https://composerprogrammer.com/code.html

/ Hans


 From Klapuri, "Multipitch analysis of polyphonic music and
  speech signals using an auditory model", from PolyPitch:

"The  method  consists  of  a computational model of the
  human auditory periphery, followed by a periodicity analysis
  mechanism where fundamental frequencies are iteratively
  detected and canceled from the mixture signal."

Wow. That seems much different than all the other papers I read.
Wonder how well it works, especially if applied to guitar.

It sort of reminds me of how I once was part of Sony's rollout of
  Surround Retrieval System technology.
It was TV surround speakers modeled based on human hearing,
  to make one pair of these speakers simulate a truer surround.

Tim.


Of the dozens of papers referenced by that, this one caught my eye:

"Fundamental frequency estimation of musical signals using a
 two-way mismatch procedure"

https://pdfs.semanticscholar.org/c94b/56f21f32b3b7a9575ced317e3de9b2ad9081.pdf

IIUC this may be more in line with how this real product works.


I'm having a bit of a tough time with SC. Never used it before.
First the packages in SUSE were broken, so I built from source.
Now I can run the SC IDE and so on without errors. OK.
Next, I had to really hack the PolyPitch cmake file to
 get it to install. OK.

But now what? I can't seem to run any examples. Some complain of
 unexpected/unrecognized words, some complain of asking for
 Mac Cocoa. Seems some stuff was built on Mac, like PolyPitch.

Any quick tips appreciated before I might pester the SC crowd.

Thanks.
Tim.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Ralf Mardorf]

On Tue, 26 Jun 2018 15:03:04 -0400, Tim wrote:
>If I understand correctly the theory goes something like this:
>If you are looking for a dog in a picture, far better to compare
>  with real pictures of dogs already stored than to only have
>  a rough mathematical idea of what a dog should look like.

Yes, they seemingly solved one issue of several issues.

What happens when playing the following chord

A7#9
  e a d g b e
6 | | | o | |
7 | | o | | |
8 | | | | o o

and while holding the chord bending the b and e string at the 8th fret?

Keep in mind that using a divided pickup it's possible to e.g.
use modeling for the e, a and d string, e.g. a neck pickup of a
Stratocaster, with a Drop D tuning (while the guitar is a LesPaul not
tuned to a Drop D tuning), while the g, b and e string send MIDI
messages to 3 different MIDI channels.

>There is talk of this software obsoleting using special pickups.
>I would tend to agree, it's pretty darn good.

Perhaps if the purpose is sending MIDI events only, but a guitar synth
provides more. You individually could change the volume and tuning of
each strings output, you could change the velocity curve. Some sounds
such as a lead synth allow pitch bend, while bending a guitar string,
other sounds, such as a grand piano don't allow this. In addition you
could mix it with all kinds of modeling.

>> My new guitar additionally has got a Sustaniac driver.  
>
>Ah, just looked that up.
>Similar to the famous e-bow hand-held sustainer?

Yes, but it could add endless sustain to a note and by a three position
switch fade to 2 different kinds of harmonics to simulate feedback.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Hans Wilmers]

On 06/26/2018 10:25 PM, Tim wrote:
> 
> 
> On 06/26/2018 03:55 PM, Hans Wilmers wrote:
>> On 06/26/2018 08:32 PM, Spencer Jackson wrote:
>>>
>>>
>>> I don't know of anyone really working on polyphonic pitch recognition in
>>> the open source world. I think Bayesian filtering of some kind though
>>> would be compelling. Perhaps some of the work from ISSE
>>> (http://isse.sourceforge.net/) could be used and made realtime.
>>>
>>
>> There is a SuperCollider plugin by Nick Collins called PolyPitch, which
>> does what the name suggests.
>> The source is GPL, and available here:
>> https://composerprogrammer.com/code.html
>>
>> / Hans
> 
> From Klapuri, "Multipitch analysis of polyphonic music and
>  speech signals using an auditory model", from PolyPitch:
> 
> "The  method  consists  of  a computational model of the
>  human auditory periphery, followed by a periodicity analysis
>  mechanism where fundamental frequencies are iteratively
>  detected and canceled from the mixture signal."
> 
> Wow. That seems much different than all the other papers I read.
> Wonder how well it works, especially if applied to guitar.
> 
> It sort of reminds me of how I once was part of Sony's rollout of
>  Surround Retrieval System technology.
> It was TV surround speakers modeled based on human hearing,
>  to make one pair of these speakers simulate a truer surround.
> 
> Tim.

I have used PolyPitch for resynthesis of violin sounds. It does detect
double stops, though there were also false positives.
I have no idea how it would perform with a guitar, but if you ask Nick
Collins, he will give you a hint.

/ Hans




---
Hans Wilmers
Notam
Sandakerveien 24 D, bygg F3
N-0473 Oslo Norway
tlf.: +47 22358060
mob.: +47 92459361
http://www.notam02.no
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Greg]

On Tue, Jun 26, 2018 at 04:25:27PM -0400, Tim wrote:

"The  method  consists  of  a computational model of the
human auditory periphery, followed by a periodicity analysis
mechanism where fundamental frequencies are iteratively
detected and canceled from the mixture signal."


Neat. Sounds like the Progressive Interference Cancellation my team patented
for CDMA years ago. Hear a signal, reconstruct it based on what you think you
heard, subtract the synthesized signal from the input, drop the noise floor.
Lather, rinse, repeat. (We got ~3x channel capacity from it in a CDMA 
context...)

Of course, it's a bit easier to reconstruct a CDMA signal than an analog
guitar signal, but we'll get there some day. I can't wait! ;)

-g
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Tim]

On 06/26/2018 03:55 PM, Hans Wilmers wrote:

On 06/26/2018 08:32 PM, Spencer Jackson wrote:



I don't know of anyone really working on polyphonic pitch recognition in
the open source world. I think Bayesian filtering of some kind though
would be compelling. Perhaps some of the work from ISSE
(http://isse.sourceforge.net/) could be used and made realtime.



There is a SuperCollider plugin by Nick Collins called PolyPitch, which
does what the name suggests.
The source is GPL, and available here:
https://composerprogrammer.com/code.html

/ Hans


From Klapuri, "Multipitch analysis of polyphonic music and
 speech signals using an auditory model", from PolyPitch:

"The  method  consists  of  a computational model of the
 human auditory periphery, followed by a periodicity analysis
 mechanism where fundamental frequencies are iteratively
 detected and canceled from the mixture signal."

Wow. That seems much different than all the other papers I read.
Wonder how well it works, especially if applied to guitar.

It sort of reminds me of how I once was part of Sony's rollout of
 Surround Retrieval System technology.
It was TV surround speakers modeled based on human hearing,
 to make one pair of these speakers simulate a truer surround.

Tim.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Hans Wilmers]

On 06/26/2018 08:32 PM, Spencer Jackson wrote:
> 
> 
> I don't know of anyone really working on polyphonic pitch recognition in
> the open source world. I think Bayesian filtering of some kind though
> would be compelling. Perhaps some of the work from ISSE
> (http://isse.sourceforge.net/) could be used and made realtime.
> 

There is a SuperCollider plugin by Nick Collins called PolyPitch, which
does what the name suggests.
The source is GPL, and available here:
https://composerprogrammer.com/code.html

/ Hans
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Tim]

On 06/26/2018 02:16 PM, Ralf Mardorf wrote:

On Mon, 25 Jun 2018 18:31:12 -0400, Tim wrote:

Then I stumbled across this product, MIDI-Guitar from Jam Origins.


For quite some time now, the free version is installed on my iPad, but
I never tested it and meanwhile I've got two electric guitars with
Roland GK-3 PUs and a Roland GR-55. Keep in mind that even the best
polyphonic tracking done by software for Apple and Microsoft based
computers, for usage with averaged guitar pickups, at least suffers from
the additional audio device latency.


What is truly astounding is the product's low latency.
The patent mentions an ideal FFT data block of around 95mS
 or 4096 samples at 44100Hz samplerate.
But, I know what a 4096 sample delay feels like. It's long.
The latency seems /much/ less than that in this product.
It sure feels like magic.

I've been playing it for two nights now and it really is
 fast and accurate.

If I understand correctly the theory goes something like this:
If you are looking for a dog in a picture, far better to compare
 with real pictures of dogs already stored than to only have
 a rough mathematical idea of what a dog should look like.


Before I bought the Roland gear, I watched videos and read reviews.
Roland seems to be the best solution at the moment, even combinations
of a Roland GR-55 with piezo bridge pickups, instead of the Roland GK-3
seems to be less reliable.

However, for several reasons it indeed would be nice, if the polyphonic
tracking software, for usage with averaged guitar pickups would make
progress.


I highly advise to try the software. I found it amazing.
There is talk of this software obsoleting using special pickups.
I would tend to agree, it's pretty darn good.


OTOH some kind of divided pickup build into modern guitars
has got an advantage, too, since modeling makes a lot of progress. My
new guitar additionally has got a Sustaniac driver.


Ah, just looked that up.
Similar to the famous e-bow hand-held sustainer?

I use a novel approach:
I 'sing' into my guitar neck, the neck touching my throat,
 which resonates the strings infinitely for as long as I
 can sing a note.


IOW adding special
pickups to guitars is not necessarily a disadvantage. Perhaps
polyphonic tracking software, for usage with averaged guitar pickups,
in combination with special pickups is interesting in the near future.


Yeah that would be neat. Maybe someday all guitars will come with
 such pickups and standardized connections.
From there, the sky's the limit on modeling and recognition.

Tim.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Spencer Jackson]

On Mon, Jun 25, 2018 at 4:31 PM, Tim  wrote:

> I read they use more than just spectral stuff.
> Like AI used in speech recognition and so on.
>
> Amazing what DSP audio and image coding can do these days.
> Any thoughts on coding techniques? I've read a lot of papers!
> Some say using FFTs + auto-correlation comparisons.
> Some say non-negative matrix.
> My head spins, but this team definitely deserves praise.
> Can open source come up with something?
>


Had I sufficient time I'd like to investigate using hidden markov models
and turbo coding the way some speech recognition algorithms do. This could
give you estimates of the key you are playing in and use that data to
calculate the likelihood of various pitches etc etc. I suspect you could
get some interesting results.

I don't know of anyone really working on polyphonic pitch recognition in
the open source world. I think Bayesian filtering of some kind though would
be compelling. Perhaps some of the work from ISSE (
http://isse.sourceforge.net/) could be used and made realtime.

_Spencer
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Ralf Mardorf]

On Mon, 25 Jun 2018 18:31:12 -0400, Tim wrote:
>Then I stumbled across this product, MIDI-Guitar from Jam Origins.

For quite some time now, the free version is installed on my iPad, but
I never tested it and meanwhile I've got two electric guitars with
Roland GK-3 PUs and a Roland GR-55. Keep in mind that even the best
polyphonic tracking done by software for Apple and Microsoft based
computers, for usage with averaged guitar pickups, at least suffers from
the additional audio device latency.

Before I bought the Roland gear, I watched videos and read reviews.
Roland seems to be the best solution at the moment, even combinations
of a Roland GR-55 with piezo bridge pickups, instead of the Roland GK-3
seems to be less reliable.

However, for several reasons it indeed would be nice, if the polyphonic
tracking software, for usage with averaged guitar pickups would make
progress. OTOH some kind of divided pickup build into modern guitars
has got an advantage, too, since modeling makes a lot of progress. My
new guitar additionally has got a Sustaniac driver. IOW adding special
pickups to guitars is not necessarily a disadvantage. Perhaps
polyphonic tracking software, for usage with averaged guitar pickups,
in combination with special pickups is interesting in the near future.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Tim]

On 06/26/2018 02:44 AM, Bengt Gördén wrote:

Den 2018-06-26 kl. 00:31, skrev Tim:


I read they use more than just spectral stuff.
Like AI used in speech recognition and so on.





Amazing what DSP audio and image coding can do these days.
Any thoughts on coding techniques? I've read a lot of papers!
Some say using FFTs + auto-correlation comparisons.
Some say non-negative matrix.


 From the abstract of the patent [1] [2]

"An audio matching method, use of the method in a game system, an audio
matching system and a data carrier are provided, where the audio
matching method is for comparing an input audio fragment with reference
audio fragment variants, the method being an incremental search method,
including repeating the steps of: obtaining a number of reference audio
fragments variants on the basis of one or more stored audio fragments
from a reference storage; and comparing the input audio fragment against
the number of reference audio fragment variants to determine a
comparison result; whereby repetition of the steps is carried out a
predetermined number of times or as long as the comparison result improves."

SOS had a review of it 4 years ago [3]


ref.
[1]
https://patentimages.storage.googleapis.com/93/f6/7b/30b78e5067b5e5/US20120132057A1.pdf
[2] https://patents.google.com/patent/US20120132057A1/en
[3] https://www.soundonsound.com/reviews/jam-origin-midi-guitar

Cheers,



Awesome. Thank you for that.
I am always late to the party, but I can't help gushing
 over this thing since I have a special interest in it.

Tim.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Dominique Michel]

Le Mon, 25 Jun 2018 18:31:12 -0400,
Tim  a écrit :

> Hi list, some time ago a coder was asking about
>   making such an app. I think it's on github.
>

> Amazing what DSP audio and image coding can do these days.
> Any thoughts on coding techniques? I've read a lot of papers!
> Some say using FFTs + auto-correlation comparisons.
> Some say non-negative matrix.
> My head spins, but this team definitely deserves praise.
> Can open source come up with something?

To go fast and get a low enough latency, you can use can what time is
between 2 consecutive peaks of the signal. 

Be aware than you would have to smooth the signal because when playing
loud and the strings are touching the frets, the harmonic content
become so weird you can get false maximums (as seen by a signal
analysis of my electrical guitar with a memory oscilloscope), which will
give you a much higher fundamental frequency than expected.

Normal signal:

   *   *
   * * * *
  *   *   *   *
  **  **
 *  **  *
 *   *   *   *
*=*=*=*
   *   *   *
   *  **
**  *
*   *   *
 * * * *
 *   *

Fretted signal:

   *   *   *   *
   * ** *  * ** *
  *   * * *   * *
  * * * *
 *  **  *
 *   *   *   *
*=*=*=*
   *   *   *
   *  **
**  *
*   *   *
 * * * *
 *   *

These double peaks can be very huge as an electrical guitar have a huge
dynamic during the attack of the notes. Doing a repetitive arithmetic
average calculation when the sampling are coming in must be enough to
smooth them in real time, and to get ride of other high order harmonics
at the same time.

The latency will be dependant of the fundamental frequency, something
like one period + trigger delay. 

Cheers,
Dom
  
> 
> Cheers, Tim.
> ___
> Linux-audio-dev mailing list
> Linux-audio-dev@lists.linuxaudio.org
> https://lists.linuxaudio.org/listinfo/linux-audio-dev


-- 
If you have a problem and you are not doing anything to fix it, you are
at the heart of the problem.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Bengt Gördén]

Den 2018-06-26 kl. 00:31, skrev Tim:
>
> I read they use more than just spectral stuff.
> Like AI used in speech recognition and so on.



> Amazing what DSP audio and image coding can do these days.
> Any thoughts on coding techniques? I've read a lot of papers!
> Some say using FFTs + auto-correlation comparisons.
> Some say non-negative matrix.

From the abstract of the patent [1] [2]

"An audio matching method, use of the method in a game system, an audio
matching system and a data carrier are provided, where the audio
matching method is for comparing an input audio fragment with reference
audio fragment variants, the method being an incremental search method,
including repeating the steps of: obtaining a number of reference audio
fragments variants on the basis of one or more stored audio fragments
from a reference storage; and comparing the input audio fragment against
the number of reference audio fragment variants to determine a
comparison result; whereby repetition of the steps is carried out a
predetermined number of times or as long as the comparison result improves."

SOS had a review of it 4 years ago [3]


ref.
[1]
https://patentimages.storage.googleapis.com/93/f6/7b/30b78e5067b5e5/US20120132057A1.pdf
[2] https://patents.google.com/patent/US20120132057A1/en
[3] https://www.soundonsound.com/reviews/jam-origin-midi-guitar

Cheers,

-- 
/bengan
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Tim]

On 06/25/2018 07:08 PM, Paul Davis wrote:



On Mon, Jun 25, 2018 at 6:31 PM, Tim > wrote:


​hen I stumbled across this product,
  MIDI-Guitar from Jam Origins.


​but can it handle negative harmony?​


OK I could make a TON of jokes about that.

But I looked it up and I see it's a real thing.

But "as a comic in all seriousness" (How are ya! Eugene Levy
 as Bobby Bittman) I can say the product does really well at
 separating whatever notes you throw at it on the guitar.

Cheers, Tim.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

Re: [LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Paul Davis]

On Mon, Jun 25, 2018 at 6:31 PM, Tim  wrote:

> ​hen I stumbled across this product,
>  MIDI-Guitar from Jam Origins.
>

​but can it handle negative harmony?​
​
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev

[LAD] Polyphonic normal guitar to midi: Jam Origins' MIDI-Guitar

[Tim]

Hi list, some time ago a coder was asking about
 making such an app. I think it's on github.

I replied that I made one, and then I posted my
 very old project on github. terminator356/polyguitsynth

It used FFTs, no windowing, but actually sort of worked.
Latency was of course fundamentally an issue.
I have always had my eye on this golden egg of a goal,
 but obviously it's tough, beyond my skills and time.

So I sat down this weekend to try to modernize it.
Qt, RtAudio, and especially the DSP. I looked at aubio
 but unfortunately its 'aubionotes' feature is
 (currently?) monophonic.


Then I stumbled across this product,
 MIDI-Guitar from Jam Origins.

I realize it's Windows product and this is Linux, but...

Wow! What the... how the...
The accuracy is astounding. The latency very low.

Go and grab it, ye original poster who requested this!

Very cheap price. I bought it. Worth every penny (I told them)!
It runs flawlessly under wine with wineASIO  - and - Jack2
 which some said was broken in this respect...

I read they use more than just spectral stuff.
Like AI used in speech recognition and so on.

Amazing what DSP audio and image coding can do these days.
Any thoughts on coding techniques? I've read a lot of papers!
Some say using FFTs + auto-correlation comparisons.
Some say non-negative matrix.
My head spins, but this team definitely deserves praise.
Can open source come up with something?

Cheers, Tim.
___
Linux-audio-dev mailing list
Linux-audio-dev@lists.linuxaudio.org
https://lists.linuxaudio.org/listinfo/linux-audio-dev