[Csnd] Audio Descriptors

Hello lists.
Anybody working on Audio Descriptors with Csound?
Opcodes, UDOs, Plugins, and even scores could be useful in my search.
Thanks in advance.

Hi

This is something I’ve been working with on-and-off for some time but is not really ready for public release yet… However within a few days I could pull together into a legible state and share some bits, but they would be in a very beta state, at the most. There are two approaches I’ve been using, mainly for concatenative resynthesis/similar:

Using MFCC natively with the mfb Csound opcode.
Various descriptors using LibXtract in a custom plugin opcode.

Hopefully could be of some interest!

Richard

I would be most interested in this too. What you do with LibXract sounds interesting.

I’ve done some feature extraction too, for use in the crossadaptive processing, the analyzer module could easily be untangled and used in other contexts:
https://github.com/Oeyvind/featexmod

all best
Øyvind

tor. 11. mar. 2021 kl. 22:49 skrev Richard Knight <richard@1bpm.net>:

Hi

This is something I’ve been working with on-and-off for some time but is not really ready for public release yet… However within a few days I could pull together into a legible state and share some bits, but they would be in a very beta state, at the most. There are two approaches I’ve been using, mainly for concatenative resynthesis/similar:

Using MFCC natively with the mfb Csound opcode.
Various descriptors using LibXtract in a custom plugin opcode.

Hopefully could be of some interest!

Richard

Hello lists.
Anybody working on Audio Descriptors with Csound?
Opcodes, UDOs, Plugins, and even scores could be useful in my search.
Thanks in advance.

Dr. Oscar Pablo Di Liscia
Profesor Titular
Director Programa de Investigación “Sistemas Temporales y Síntesis Espacial en el Arte Sonoro”
http://stseas.web.unq.edu.ar/
Director Colección Editorial “Música y Ciencia”
Escuela Universitaria de Artes
Universidad Nacional de Quilmes
Argentina

Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Many thanks to both, Richard and Oeyvind, I will take a look to the resources you pointed at.
Best.
Pablo

Hi

Thanks Øyvind, that’s really interesting and I’ll have a proper play shortly, I hadn’t considered the possibility of such detailed feature extraction natively in Csound.

My LibXtract opcodes are currently quite experimental - I need to refactor some bits, but they work to get some usable features into Csound.
http://git.1bpm.net/csound-xtract/about/ , latest snapshot here: http://git.1bpm.net/csound-xtract/snapshot/csound-xtract-master.zip

Happy to have any suggestions or help if any issues compiling etc. Have tested on Windows and Linux.

Also related, here’s my native Csound concatenative resynthesis demo, using the mfb opcode to extract/store/match MFCC analysis:
http://git.1bpm.net/csd-mfccmatch/about/ , latest snapshot here: http://git.1bpm.net/csd-mfccmatch/snapshot/csd-mfccmatch-master.zip

all the best,
Richard