Open jenweilee opened 5 years ago
For calculating the spectrogram/mfcc, I saw the preprocessing of audio uses the function contrib_audio.audio_spectrogram and contrib_audio.mfcc by tensorflow. But where is the parameter of FFT size or what is the default value of it.
FFT used in this example is 1024 size, normally sizes of fft is always 2th power (128,256,512.....)
For calculating the spectrogram/mfcc, I saw the preprocessing of audio uses the function contrib_audio.audio_spectrogram and contrib_audio.mfcc by tensorflow. But where is the parameter of FFT size or what is the default value of it.