Mingde
Mingde
I read the paper and codes and have some trouble about the calculation of pitch? The code which I think is to calculate the pitch: in `lpcnet_enc.c line 526-546` ```c...
I want the `D` map to multi `d` like `2d`, `3d` , `4d` and so on..., but it can't.
I extract feature file(.f32) by modify /Tacotron-2/preprocessor.py([link](https://github.com/alokprasad/LPCTron/blob/master/Tacotron-2/datasets/preprocessor.py#L133)) ```python # I change the preprocessor.py line 133: feature_name = 'feature-{}.f32'.format(wavfile) mel_filename = 'mel-{}.npy'.format(wavfile) linear_filename = 'linear-{}.npy'.format(wavfile) mel_spectrogram = mel_spectrogram.T np.save(os.path.join(mel_dir, mel_filename), mel_spectrogram,...
In you code, you use f32 (LPCNet feature ) to mel_target to train tacotron2. Can we use keep the mel spectrum and change f32 as linear_spectrum_target ?