r9y9 / deepvoice3_pytorch

PyTorch implementation of convolutional neural networks-based text-to-speech synthesis models
https://r9y9.github.io/deepvoice3_pytorch/
Other
1.97k stars 485 forks source link

typecast error 'int 16' not supported #162

Closed Ella77 closed 5 years ago

Ella77 commented 5 years ago

in nikl_m.py

mel_spectrogram = audio.melspectrogram(wav).astype(np.float32)

https://r9y9.github.io/nnmnkwii/stable/references/generated/nnmnkwii.preprocessing.preemphasis.html

Traceback (most recent call last): File "/home/soojin/.conda/envs/soojin/lib/python3.6/concurrent/futures/process.py", line 175, in _process_worker r = call_item.fn(*call_item.args, **call_item.kwargs) File "/home/soojin/deepvoice3_pytorch/nikl_m.py", line 76, in _process_utterance k = audio.spectrogram(wav) File "/home/soojin/deepvoice3_pytorch/audio.py", line 42, in spectrogram D = _lws_processor().stft(preemphasis(y)).T File "/home/soojin/deepvoice3_pytorch/audio.py", line 33, in preemphasis return preemphasis(x, hparams.preemphasis) File "/home/soojin/.conda/envs/soojin/lib/python3.6/site-packages/nnmnkwii/preprocessing/generic.py", line 185, in preemphasis return signal.lfilter(b, a, x) File "/home/soojin/.conda/envs/soojin/lib/python3.6/site-packages/scipy/signal/signaltools.py", line 1373, in lfilter raise NotImplementedError("input type '%s' not supported" % dtype) NotImplementedError: input type 'int16' not supported

example here from nnmnkwii.util import example_audio_file

from scipy.io import wavfile fs, x = wavfile.read(example_audio_file()) >>> x = x.astype(np.float64) from nnmnkwii import preprocessing as P y = P.preemphasis(x, coef=0.97) assert x.shape == y.shape

first I faced with audio read no backend error ... I thought that was codec error and had hard time debugging but it was just simple error... That type cast should be done earlier converting numpy to float parameter. wav = audio.load_wav(wav_path) wave = wav.astype(np.float32) spectrogram = audio.spectrogram(wave)

I wonder it can be just version problem

stale[bot] commented 5 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.