audio analysis

[diemoschwarz]

• simple: 1 segmentation, pipo.descr
• full: 3 segmentations, pipo.descr
• mosaicing: MFCC

Extra points: only 1 mubu.process

[diemoschwarz]

• check a better way to manage several choices of descriptors:
  • dynamically loaded sub/bpatchers with mubu.process in them?
  • set the container refered by mubu.process to a dummy mubu to disable it?

[einbond]

• added catart-by-mubu-analysis.maxpat : allows choice of three segmentation modes based on loading abstractions from bpatcher

• so far three analysis modules, camu.process.descr-chop, camu.process.descr-gate, camu.process.descr-onseg

[diemoschwarz]

Great! need to bundle this into camu.analysis for reuse TODO/TBD:

• convert to mono before analysis: sum:descr:onseg
• allow choice for Std, default off

[einbond]

This issue is done except for the MFCC part, and no extra points for one pipo. Should we close it and split MFCC into a separate issue?

[diemoschwarz]

+1!

[diemoschwarz]

almost everything is done since ages, remaining points are in #28, #49, except:

• [ ] bundle this into camu.analysis for reuse (see #116)
• [ ] convert audio to mono before analysis: sum:descr:onseg