nnsvs.gen.postprocess_waveform

nnsvs.gen.postprocess_waveform(wav, sample_rate, dtype=<class 'numpy.int16'>, peak_norm=False, loudness_norm=False, target_loudness=-20.0)[source]

Perform post-processing for synthesized waveform

Parameters:
  • wav (ndarray) – The input waveform

  • sample_rate (int) – The sampling rate

  • dtype (np.dtype) – The dtype of output waveform. Default is np.int16.

  • peak_norm (bool) – Whether to perform peak normalization

  • loudness_norm (bool) – Whether to perform loudness normalization

  • target_loudness (float) – Target loudness in dB

Returns:

The post-processed waveform

Return type:

ndarray