Audio processing
Foreword: Sorry about the python=2.7
constraint but yaafe
does not support Python 3 at the time of writing this.
Ping me if this is no longer true!
$ conda create --name pyannote python=2.7 anaconda
$ source activate pyannote
$ conda install -c yaafe yaafe=0.65
$ pip install "theano==0.8.2"
$ pip install "keras==1.1.0"
$ pip install pyannote.audio
What did you just install?
keras
(and itstheano
backend) is used for all things deep.yaafe
is used for MFCC feature extraction. You might also want to checkoutlibrosa
(easy to install, but much slower) thoughpyannote.audio
does not support it yet (pull requests are welcome, though!)pyannote.audio
is this library.
Not yet available.