Hi,
Has anyone done any audio segmentation work in PD, along the lines of what is provided by BBCut for SuperCollider?
The sort of thing I am looking for is the ability to take a monophonic input (e.g. a flute), and perform on-the-fly segmentation of the audio into a 'database' of notes or phrases depending on pre-determined rules. The segmented 'events' would then be available for recall in an arbitrary order.
I have a few ideas of how to implement this PD, but I don't want to reinvent the wheel, so if anyone has already done it and wants to share...
Jamie
Hi Jamie, If this doesn't exist, I'd love to hear what you work out, as I'm really interested in using this approach myself. It wouldn't be to hard to, for instance, build a buffer-glitch type instrument on such a foundation... ~David
On 10/26/06, Jamie Bullock jamie@postlude.co.uk wrote:
Hi,
Has anyone done any audio segmentation work in PD, along the lines of what is provided by BBCut for SuperCollider?
The sort of thing I am looking for is the ability to take a monophonic input (e.g. a flute), and perform on-the-fly segmentation of the audio into a 'database' of notes or phrases depending on pre-determined rules. The segmented 'events' would then be available for recall in an arbitrary order.
I have a few ideas of how to implement this PD, but I don't want to reinvent the wheel, so if anyone has already done it and wants to share...
Jamie
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Hello,
In 2004 I had a live improvisation patch that did something similar (it was used during a bass/keyboard/computer concert with two musicians):
- first you had to record into a 30 sec buffer (fretless electric bass at that
time); recording was trigged manually at special moments
- during the recording, attacks were detected with [bonk~] and the times at
which they occured were stored in a [table] (there was a kind of chronometer that started at the same time as the recording)
- after the recording was finished, you could play the buffer with a MIDI
keyboard; each key played the sound of the buffer from a specific time taken in the [table] mentionned above (there was also a trapeze enveloppe for avoiding clicks)
- actually, the sound was not played as recorded but by a phase vocoder similar
to the one that you can find in the pd.fft.examples, in order to timestretch the sound with a default precession speed of 10% or such (MIDI aftertouch was then used to alter the precession)
With a flute, you would have to replace the [bonk~] by a different "rule" (unless you only play tongue rams and key percussion)
On 10/26/06, Jamie Bullock jamie@postlude.co.uk wrote:
Hi,
Has anyone done any audio segmentation work in PD, along the lines of what is provided by BBCut for SuperCollider?
The sort of thing I am looking for is the ability to take a monophonic input (e.g. a flute), and perform on-the-fly segmentation of the audio into a 'database' of notes or phrases depending on pre-determined rules. The segmented 'events' would then be available for recall in an arbitrary order.
I have a few ideas of how to implement this PD, but I don't want to reinvent the wheel, so if anyone has already done it and wants to share...
Jamie
PD-list@iem.at mailing list UNSUBSCRIBE and account-management ->
http://lists.puredata.info/listinfo/pd-list
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Cool. Have you still got the code for that? Some dudes on the forum have been tring to make a beat slicer like Freecycle using that exact method.
On Thu, 26 Oct 2006 17:45:57 +0200 julien breval julienbreval@free.fr wrote:
Hello,
In 2004 I had a live improvisation patch that did something similar (it was used during a bass/keyboard/computer concert with two musicians):
- first you had to record into a 30 sec buffer (fretless electric bass at that
time); recording was trigged manually at special moments
- during the recording, attacks were detected with [bonk~] and the times at
which they occured were stored in a [table] (there was a kind of chronometer that started at the same time as the recording)
- after the recording was finished, you could play the buffer with a MIDI
keyboard; each key played the sound of the buffer from a specific time taken in the [table] mentionned above (there was also a trapeze enveloppe for avoiding clicks)
- actually, the sound was not played as recorded but by a phase vocoder similar
to the one that you can find in the pd.fft.examples, in order to timestretch the sound with a default precession speed of 10% or such (MIDI aftertouch was then used to alter the precession)
With a flute, you would have to replace the [bonk~] by a different "rule" (unless you only play tongue rams and key percussion)
On 10/26/06, Jamie Bullock jamie@postlude.co.uk wrote:
Hi,
Has anyone done any audio segmentation work in PD, along the lines of what is provided by BBCut for SuperCollider?
The sort of thing I am looking for is the ability to take a monophonic input (e.g. a flute), and perform on-the-fly segmentation of the audio into a 'database' of notes or phrases depending on pre-determined rules. The segmented 'events' would then be available for recall in an arbitrary order.
I have a few ideas of how to implement this PD, but I don't want to reinvent the wheel, so if anyone has already done it and wants to share...
Jamie
PD-list@iem.at mailing list UNSUBSCRIBE and account-management ->
http://lists.puredata.info/listinfo/pd-list
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Hallo, Jamie Bullock hat gesagt: // Jamie Bullock wrote:
Has anyone done any audio segmentation work in PD, along the lines of what is provided by BBCut for SuperCollider?
The sort of thing I am looking for is the ability to take a monophonic input (e.g. a flute), and perform on-the-fly segmentation of the audio into a 'database' of notes or phrases depending on pre-determined rules. The segmented 'events' would then be available for recall in an arbitrary order.
You may want to check out aubio for this: http://aubio.piem.org/
Aubio is a bit like improved bonk~ and fiddle~ and also works outside of Pd.
Ciao
Frank Barknecht _ ______footils.org_ __goto10.org__
Hallo, Frank Barknecht hat gesagt: // Frank Barknecht wrote:
Hallo, Jamie Bullock hat gesagt: // Jamie Bullock wrote:
Has anyone done any audio segmentation work in PD, along the lines of what is provided by BBCut for SuperCollider?
The sort of thing I am looking for is the ability to take a monophonic input (e.g. a flute), and perform on-the-fly segmentation of the audio into a 'database' of notes or phrases depending on pre-determined rules. The segmented 'events' would then be available for recall in an arbitrary order.
You may want to check out aubio for this: http://aubio.piem.org/
Okay, I hacked together a little patch which uses aubioonset~ to analyse a sample and then randomly playes back the slices. The "-help" file is the main patch.
Ciao
Frank Barknecht _ ______footils.org_ __goto10.org__
On Thu, 2006-10-26 at 22:49 +0200, Frank Barknecht wrote:
Okay, I hacked together a little patch which uses aubioonset~ to analyse a sample and then randomly playes back the slices. The "-help" file is the main patch.
Thanks Frank, that's useful, in fact it is pretty much where I was thinking of starting. I am already familiar with the aubio library (the onset detector is very impressive even for non-percussive sounds), and I will definitely use it.
I'll take your patch as a starting point and report back to the list when I've done some further work. I'll probably use my own libxtract library to 'tag' the event associated with each onset with additional features (fundamental frequency, centroid, zcr etc.), to assist with algorithmic recombination of events. It would also be good to do some tempo estimation to create possibilities for timings relative to the original source.
best,
Jamie
Do you ever sleep, Frank? ;-) ~Kyle
On 10/26/06, Frank Barknecht fbar@footils.org wrote:
Hallo, Frank Barknecht hat gesagt: // Frank Barknecht wrote:
Hallo, Jamie Bullock hat gesagt: // Jamie Bullock wrote:
Has anyone done any audio segmentation work in PD, along the lines of what is provided by BBCut for SuperCollider?
The sort of thing I am looking for is the ability to take a monophonic input (e.g. a flute), and perform on-the-fly segmentation of the audio into a 'database' of notes or phrases depending on pre-determined
rules.
The segmented 'events' would then be available for recall in an arbitrary order.
You may want to check out aubio for this: http://aubio.piem.org/
Okay, I hacked together a little patch which uses aubioonset~ to analyse a sample and then randomly playes back the slices. The "-help" file is the main patch.
Ciao
Frank Barknecht _ ______footils.org_ __goto10.org__
PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Do you mind if I repost this on the pd-forum? It will make some happy musicians I think.
On Thu, 26 Oct 2006 22:49:33 +0200 Frank Barknecht fbar@footils.org wrote:
Hallo, Frank Barknecht hat gesagt: // Frank Barknecht wrote:
Hallo, Jamie Bullock hat gesagt: // Jamie Bullock wrote:
Has anyone done any audio segmentation work in PD, along the lines of what is provided by BBCut for SuperCollider?
The sort of thing I am looking for is the ability to take a monophonic input (e.g. a flute), and perform on-the-fly segmentation of the audio into a 'database' of notes or phrases depending on pre-determined rules. The segmented 'events' would then be available for recall in an arbitrary order.
You may want to check out aubio for this: http://aubio.piem.org/
Okay, I hacked together a little patch which uses aubioonset~ to analyse a sample and then randomly playes back the slices. The "-help" file is the main patch.
Ciao
Frank Barknecht _ ______footils.org_ __goto10.org__
Hallo, padawan12 hat gesagt: // padawan12 wrote:
Do you mind if I repost this on the pd-forum? It will make some happy musicians I think.
No, feel free to do so. Unless specified otherwise I consider every patch I post to the list public domain. (Credits are nice though.)
Ciao
Frank Barknecht _ ______footils.org_ __goto10.org__
-
David Powers -
Frank Barknecht -
Jamie Bullock -
julien breval -
Kyle Klipowicz -
padawan12