Pierre Alexandre Tremblay
|
779b35cc1e
|
various compositing of the clustering format- TODO: understand overlap consequences for realz TODO: maybe another step of conditional rejection (if 2 of 3 slice, then slice)
|
5 years ago |
|
Pierre Alexandre Tremblay
|
4e321b9249
|
fix for overlap of 2. TODO: fix for all overlaps
|
5 years ago |
|
Pierre Alexandre Tremblay
|
0e1a62b11a
|
now adapted for using our folder/dataset creators
|
5 years ago |
|
Pierre Alexandre Tremblay
|
d084350de7
|
a few improvments: more iteration to kmeans, the indices are now in integer, and duplicates are removed, and there is a date stamp to the Reaper file
|
5 years ago |
|
Pierre Alexandre Tremblay
|
bb2034d5ec
|
explored less dimensions, more overlap
|
5 years ago |
|
Pierre Alexandre Tremblay
|
dc6ad1e410
|
tigher variable scope
|
5 years ago |
|
Pierre Alexandre Tremblay
|
25dce02ce8
|
now a proper recursive function!
|
5 years ago |
|
Pierre Alexandre Tremblay
|
e5223ef0e5
|
corrected the last bloc
|
5 years ago |
|
Pierre Alexandre Tremblay
|
09a2c48c50
|
added the reaper export
|
5 years ago |
|
Pierre Alexandre Tremblay
|
2b92864757
|
this condition trick is genius
|
5 years ago |
|
Pierre Alexandre Tremblay
|
2d7da671d5
|
trying the nested iterator, not conclusive at all
|
5 years ago |
|
Pierre Alexandre Tremblay
|
85ec2e624f
|
super verbose version to check assumptions of return order and content.
|
5 years ago |
|
Pierre Alexandre Tremblay
|
646d30af17
|
segmentation redux by sliding window clustering in 2 class - v1
|
5 years ago |
|
Pierre Alexandre Tremblay
|
b8af4f3125
|
example 11: an embarassing error found by @weefuzzy (thanks!)
example 10b: a small typo found by Alice on Zoom :)
|
5 years ago |
|
Pierre Alexandre Tremblay
|
0b7538e419
|
example 11: attempt at using normalisation to weigh between datasets. Not conclusive in the end for now, conceptual errors might be the problem. I need help from the wise
|
5 years ago |
|
Pierre Alexandre Tremblay
|
c6a4c649f0
|
example 11: now with proper sanitation of the pitch and duration and explanation of why results gets stranger.
|
5 years ago |
|
Pierre Alexandre Tremblay
|
6aed1ba3dd
|
example 11. make the standardisation systematic and more in line with literature.
|
5 years ago |
|
Pierre Alexandre Tremblay
|
bb53874d83
|
cleaning examples for release
|
5 years ago |
|
Pierre Alexandre Tremblay
|
86ab09d769
|
example 11: victory!
|
5 years ago |
|
Pierre Alexandre Tremblay
|
41a8bb4ae3
|
example 11 - up to a simple compositing query
|
5 years ago |
|
Pierre Alexandre Tremblay
|
8bc4cb1d8a
|
example11 - comparison for timbral reduction
|
5 years ago |
|
Pierre Alexandre Tremblay
|
8122fab3d5
|
example 11 in progress - crashing with certain datasets
|
5 years ago |
|
Pierre Alexandre Tremblay
|
a4e6413eca
|
another cheeky sentence... I will oblige and change the order one day but not yet
|
5 years ago |
|
Pierre Alexandre Tremblay
|
da54b52102
|
examples: remove any ref to Users/pa/
|
5 years ago |
|
Pierre Alexandre Tremblay
|
b21d6a5c8d
|
pushed a first version of example 10b : using pitch to go verbose on outliers and weighting and observing data. quite sexy
|
5 years ago |
|
Pierre Alexandre Tremblay
|
5da1df8ea3
|
changed the example 10 to use the sexy new resizing options.
|
5 years ago |
|
Pierre Alexandre Tremblay
|
f191e4a327
|
comparison example of weighted and/or curated MFCC stats
|
5 years ago |
|
Pierre Alexandre Tremblay
|
33a12c3a00
|
now working with scaled loudness.
|
5 years ago |
|
Pierre Alexandre Tremblay
|
9dec86ba55
|
weighted mfcc first attempt now working
|
5 years ago |
|
Pierre Alexandre Tremblay
|
743a538e0c
|
example in progress for weighted mfccs
|
5 years ago |
|
Pierre Alexandre Tremblay
|
d0f618ea9a
|
few typos
|
5 years ago |
|
Pierre Alexandre Tremblay
|
391e43ebb0
|
now with a slighty more verbose GUI
|
5 years ago |
|
Pierre Alexandre Tremblay
|
3a115910d5
|
working example 8c (AE with gui and retrieve)
|
5 years ago |
|
Pierre Alexandre Tremblay
|
454a667ae9
|
renamed the examples to match the taxonomy in Max (better shows the missing elements too) (using git mv so everyone is happy)
|
5 years ago |
|
Pierre Alexandre Tremblay
|
a03aab2ae4
|
example 9 now in SC
|
5 years ago |
|
Pierre Alexandre Tremblay
|
0a0642502f
|
updated examples to new interface
|
5 years ago |
|
Pierre Alexandre Tremblay
|
1efc9427c5
|
updated dataset demos
|
5 years ago |
|
Pierre Alexandre Tremblay
|
31d8150946
|
more verbose KDTree KR example and AE example
|
6 years ago |
|
Pierre Alexandre Tremblay
|
21daae30ed
|
now with finalActivation
|
6 years ago |
|
Pierre Alexandre Tremblay
|
9fef109e33
|
tweaks to examples (slowing them down at times)
|
6 years ago |
|
Pierre Alexandre Tremblay
|
efab1d08b2
|
MLPregressor: corrected the helpfiles and toy example for the new parameter in the list
|
6 years ago |
|
Pierre Alexandre Tremblay
|
20278da9b3
|
added merge in the help and class def of dataset
|
6 years ago |
|
Pierre Alexandre Tremblay
|
d6b5ebd686
|
sorted problem of query in synthchaos example
|
6 years ago |
|
Pierre Alexandre Tremblay
|
28c2124c05
|
better mlp-synth-control example
|
6 years ago |
|
Pierre Alexandre Tremblay
|
e44645293d
|
new x-or MLP regressor example with sexy visualisation
|
6 years ago |
|
Pierre Alexandre Tremblay
|
79f58097dc
|
mlp upgrade
|
6 years ago |
|
Pierre Alexandre Tremblay
|
00891b97b0
|
updated chaossynth with more parameter to play with the MLP
|
6 years ago |
|
Pierre Alexandre Tremblay
|
3c036f32d3
|
amended chaos-synth-mlp example
|
6 years ago |
Owen Green
|
603d309f0d
|
Update dataest examples to interface changes
|
6 years ago |
|
Pierre Alexandre Tremblay
|
92270ed000
|
now with a smaller network (2 6 10) and gui to change the training params
|
6 years ago |