search

neuro-team-femto / cleese

Combinatorial Expressive Speech Engine
MIT License
42 stars 10 forks source link

`run_cleese` fails with `ValueError` #2

Closed logc closed 5 years ago

logc commented 5 years ago

There is no supplied example.wav in the repository. If one is created, then running run_cleese.py does not succeed.

$ python run_cleese.py
stretch variation 1/10
Traceback (most recent call last):
  File "run_cleese.py", line 15, in <module>
    cleese.process(soundData=inputFile, configFile=configFile)
  File "/Users/luis/Documents/code/cleese/cleese/cleeseProcess.py", line 138, in process
    waveOut = processWithPV(waveIn=waveIn, pars=pars, BPF=BPF, doPitchShift=doPitchShift)
  File "/Users/luis/Documents/code/cleese/cleese/cleeseProcess.py", line 302, in processWithPV
    stftMat = stft(waveIn,win,n_fft,anaHopVec)
  File "/Users/luis/Documents/code/cleese/cleese/cleeseEngine.py", line 23, in stft
    waveForm = np.concatenate((waveForm,np.zeros(numExtraSamples)))
ValueError: all the input arrays must have same number of dimensions
jjburred commented 5 years ago

works here. Did you use a valid wav file? Anyway, will replace 'example.wav' in the code with a reference to one of the provided wav files for the tutorial

logc commented 5 years ago

I recorded an aifc file with Quicktime, then converted it with ffmpeg -i ~/Desktop/example.aifc ./example.wav. Not sure if that results in a valid WAV file ...

jjburred commented 5 years ago

could you send me over your wav file?

jjburred commented 5 years ago

ah yes, CLEESE does not handle stereo files. I'm committing a check and warning for stereo files, in which case only the left channel will be processed. thanks