Python librosa.filters() Examples

The following are 26 code examples of librosa.filters(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module librosa , or try the search function .
Example #1
Source File: audio_utils.py    From WaveGlow with MIT License 5 votes vote down vote up
def _build_mel_basis():
    n_fft = hparams.n_fft
    return librosa.filters.mel(hparams.sample_rate, n_fft, n_mels=hparams.num_mels) 
Example #2
Source File: audio.py    From representation_mixing with BSD 3-Clause "New" or "Revised" License 5 votes vote down vote up
def _build_mel_basis():
    assert hparams.fmax <= hparams.sample_rate // 2
    return librosa.filters.mel(hparams.sample_rate, hparams.fft_size,
                               fmin=hparams.fmin, fmax=hparams.fmax,
                               n_mels=hparams.num_mels) 
Example #3
Source File: audio.py    From representation_mixing with BSD 3-Clause "New" or "Revised" License 5 votes vote down vote up
def _build_mel_basis():
    assert hparams.fmax <= hparams.sample_rate // 2
    return librosa.filters.mel(hparams.sample_rate, hparams.fft_size,
                               fmin=hparams.fmin, fmax=hparams.fmax,
                               n_mels=hparams.num_mels) 
Example #4
Source File: audio.py    From representation_mixing with BSD 3-Clause "New" or "Revised" License 5 votes vote down vote up
def _build_mel_basis():
    assert hparams.fmax <= hparams.sample_rate // 2
    return librosa.filters.mel(hparams.sample_rate, hparams.fft_size,
                               fmin=hparams.fmin, fmax=hparams.fmax,
                               n_mels=hparams.num_mels) 
Example #5
Source File: audio.py    From representation_mixing with BSD 3-Clause "New" or "Revised" License 5 votes vote down vote up
def _build_mel_basis():
    assert hparams.fmax <= hparams.sample_rate // 2
    return librosa.filters.mel(hparams.sample_rate, hparams.fft_size,
                               fmin=hparams.fmin, fmax=hparams.fmax,
                               n_mels=hparams.num_mels) 
Example #6
Source File: audio.py    From style-token_tacotron2 with MIT License 5 votes vote down vote up
def _build_mel_basis(hparams):
	assert hparams.fmax <= hparams.sample_rate // 2
	return librosa.filters.mel(hparams.sample_rate, hparams.n_fft, n_mels=hparams.num_mels,
							   fmin=hparams.fmin, fmax=hparams.fmax) 
Example #7
Source File: audio.py    From Tacotron2-Wavenet-Korean-TTS with MIT License 5 votes vote down vote up
def _build_mel_basis(hparams):
    #assert hparams.fmax <= hparams.sample_rate // 2
    
    #fmin: Set this to 55 if your speaker is male! if female, 95 should help taking off noise. (To test depending on dataset. Pitch info: male~[65, 260], female~[100, 525])
    #fmax: 7600, To be increased/reduced depending on data.
    #return librosa.filters.mel(hparams.sample_rate, hparams.fft_size, n_mels=hparams.num_mels,fmin=hparams.fmin, fmax=hparams.fmax)
    return librosa.filters.mel(hparams.sample_rate, hparams.fft_size, n_mels=hparams.num_mels)  # fmin=0, fmax= sample_rate/2.0 
Example #8
Source File: audio.py    From Tacotron2-PyTorch with MIT License 5 votes vote down vote up
def _build_mel_basis():
	n_fft = (hps.num_freq - 1) * 2
	return librosa.filters.mel(hps.sample_rate, n_fft, n_mels=hps.num_mels) 
Example #9
Source File: audio.py    From gmvae_tacotron with MIT License 5 votes vote down vote up
def _build_mel_basis():
	assert hparams.fmax <= hparams.sample_rate // 2
	return librosa.filters.mel(hparams.sample_rate, hparams.fft_size, n_mels=hparams.num_mels,
							   fmin=hparams.fmin, fmax=hparams.fmax) 
Example #10
Source File: audio.py    From gmvae_tacotron with MIT License 5 votes vote down vote up
def _build_mel_basis():
	assert hparams.fmax <= hparams.sample_rate // 2
	return librosa.filters.mel(hparams.sample_rate, hparams.fft_size, n_mels=hparams.num_mels,
							   fmin=hparams.fmin, fmax=hparams.fmax) 
Example #11
Source File: audio.py    From WaveRNN-Pytorch with MIT License 5 votes vote down vote up
def _build_mel_basis():
    if hparams.fmax is not None:
        assert hparams.fmax <= hparams.sample_rate // 2
    return librosa.filters.mel(hparams.sample_rate, hparams.fft_size,
                               fmin=hparams.fmin, fmax=hparams.fmax,
                               n_mels=hparams.num_mels) 
Example #12
Source File: audio.py    From Griffin_lim with MIT License 5 votes vote down vote up
def _build_mel_basis():
    n_fft = (hparams.num_freq - 1) * 2
    return librosa.filters.mel(hparams.sample_rate, n_fft, n_mels=hparams.num_mels) 
Example #13
Source File: audio.py    From tacotron with MIT License 5 votes vote down vote up
def _build_mel_basis():
  n_fft = (hparams.num_freq - 1) * 2
  return librosa.filters.mel(hparams.sample_rate, n_fft, n_mels=hparams.num_mels) 
Example #14
Source File: audio.py    From tacotron2-mandarin-griffin-lim with MIT License 5 votes vote down vote up
def _build_mel_basis(hparams):
	assert hparams.fmax <= hparams.sample_rate // 2
	return librosa.filters.mel(hparams.sample_rate, hparams.n_fft, n_mels=hparams.num_mels,
							   fmin=hparams.fmin, fmax=hparams.fmax) 
Example #15
Source File: audio.py    From cnn_vocoder with MIT License 5 votes vote down vote up
def _build_mel_basis():
    assert hparams.fmax <= hparams.sample_rate // 2
    return librosa.filters.mel(hparams.sample_rate, n_fft,
                               fmin=hparams.fmin, fmax=hparams.fmax,
                               n_mels=hparams.num_mels) 
Example #16
Source File: audio.py    From Tacotron-Wavenet-Vocoder-Korean with MIT License 5 votes vote down vote up
def _build_mel_basis(hparams):
    #assert hparams.fmax <= hparams.sample_rate // 2
    
    #fmin: Set this to 55 if your speaker is male! if female, 95 should help taking off noise. (To test depending on dataset. Pitch info: male~[65, 260], female~[100, 525])
    #fmax: 7600, To be increased/reduced depending on data.
    #return librosa.filters.mel(hparams.sample_rate, hparams.fft_size, n_mels=hparams.num_mels,fmin=hparams.fmin, fmax=hparams.fmax)
    return librosa.filters.mel(hparams.sample_rate, hparams.fft_size, n_mels=hparams.num_mels)  # fmin=0, fmax= sample_rate/2.0 
Example #17
Source File: audio.py    From ZeroSpeech-TTS-without-T with MIT License 5 votes vote down vote up
def _build_mel_basis():
	n_fft = (config.num_freq - 1) * 2
	return librosa.filters.mel(config.sample_rate, n_fft, n_mels=config.num_mels) 
Example #18
Source File: audio.py    From Tacotron-2 with MIT License 5 votes vote down vote up
def _build_mel_basis(hparams):
	assert hparams.fmax <= hparams.sample_rate // 2
	return librosa.filters.mel(hparams.sample_rate, hparams.n_fft, n_mels=hparams.num_mels,
							   fmin=hparams.fmin, fmax=hparams.fmax) 
Example #19
Source File: audio.py    From arabic-tacotron-tts with MIT License 5 votes vote down vote up
def _build_mel_basis():
  n_fft = (hparams.num_freq - 1) * 2
  return librosa.filters.mel(hparams.sample_rate, n_fft, n_mels=hparams.num_mels) 
Example #20
Source File: audio.py    From libfaceid with MIT License 5 votes vote down vote up
def _build_mel_basis():
  n_fft = (hparams.num_freq - 1) * 2
  return librosa.filters.mel(hparams.sample_rate, n_fft, n_mels=hparams.num_mels) 
Example #21
Source File: audio.py    From vae_tacotron2 with MIT License 5 votes vote down vote up
def _build_mel_basis():
	assert hparams.fmax <= hparams.sample_rate // 2
	return librosa.filters.mel(hparams.sample_rate, hparams.fft_size, n_mels=hparams.num_mels,
							   fmin=hparams.fmin, fmax=hparams.fmax) 
Example #22
Source File: audio.py    From vae_tacotron2 with MIT License 5 votes vote down vote up
def _build_mel_basis():
	assert hparams.fmax <= hparams.sample_rate // 2
	return librosa.filters.mel(hparams.sample_rate, hparams.fft_size, n_mels=hparams.num_mels,
							   fmin=hparams.fmin, fmax=hparams.fmax) 
Example #23
Source File: audio.py    From vae_tacotron with MIT License 5 votes vote down vote up
def _build_mel_basis():
  n_fft = (hparams.num_freq - 1) * 2
  return librosa.filters.mel(hparams.sample_rate, n_fft, n_mels=hparams.num_mels) 
Example #24
Source File: utils.py    From Tacotron-pytorch with MIT License 5 votes vote down vote up
def __init__(self, sample_rate, num_mels, num_freq, frame_length_ms, frame_shift_ms, preemphasis,
            min_level_db, ref_level_db, griffin_lim_iters, power):
        self.sr = sample_rate
        self.n_mels = num_mels
        self.n_fft = (num_freq - 1) * 2
        self.hop_length = int(frame_shift_ms / 1000 * sample_rate)
        self.win_length = int(frame_length_ms / 1000 * sample_rate)
        self.preemph = preemphasis
        self.min_level_db = min_level_db
        self.ref_level_db = ref_level_db
        self.GL_iter = griffin_lim_iters
        self.mel_basis = librosa.filters.mel(self.sr, self.n_fft, n_mels=self.n_mels)
        self.power = power 
Example #25
Source File: speech_utils.py    From OpenSeq2Seq with Apache License 2.0 4 votes vote down vote up
def get_mel(
    log_mag_spec,
    fs=22050,
    n_fft=1024,
    n_mels=80,
    power=2.,
    feature_normalize=False,
    mean=0,
    std=1,
    mel_basis=None,
    data_min=1e-5,
    htk=True,
    norm=None
):
  """
  Method to get mel spectrograms from magnitude spectrograms

  Args:
    log_mag_spec (np.array): log of the magnitude spec
    fs (int): sampling frequency in Hz
    n_fft (int): size of fft window in samples
    n_mels (int): number of mel features
    power (float): power of the mag spectrogram
    feature_normalize (bool): whether the mag spec was normalized
    mean (float): normalization param of mag spec
    std (float): normalization param of mag spec
    mel_basis (np.array): optional pre-computed mel basis to save computational
      time if passed. If not passed, it will call librosa to construct one
    data_min (float): min clip value prior to taking the log.
    htk (bool): whther to compute the mel spec with the htk or slaney algorithm
    norm: Should be None for htk, and 1 for slaney

  Returns:
    np.array: mel_spec with shape [time, n_mels]
  """
  if mel_basis is None:
    mel_basis = librosa.filters.mel(
        fs,
        n_fft,
        n_mels=n_mels,
        htk=htk,
        norm=norm
    )
  log_mag_spec = log_mag_spec * power
  mag_spec = np.exp(log_mag_spec)
  mel_spec = np.dot(mag_spec, mel_basis.T)
  mel_spec = np.log(np.clip(mel_spec, a_min=data_min, a_max=None))
  if feature_normalize:
    mel_spec = normalize(mel_spec, mean, std)
  return mel_spec 
Example #26
Source File: speech_utils.py    From OpenSeq2Seq with Apache License 2.0 4 votes vote down vote up
def inverse_mel(
    log_mel_spec,
    fs=22050,
    n_fft=1024,
    n_mels=80,
    power=2.,
    feature_normalize=False,
    mean=0,
    std=1,
    mel_basis=None,
    htk=True,
    norm=None
):
  """
  Reconstructs magnitude spectrogram from a mel spectrogram by multiplying it
  with the transposed mel basis.

  Args:
    log_mel_spec (np.array): log of the mel spec
    fs (int): sampling frequency in Hz
    n_fft (int): size of fft window in samples
    n_mels (int): number of mel features
    power (float): power of the mag spectrogram that was used to generate the
      mel spec
    feature_normalize (bool): whether the mel spec was normalized
    mean (float): normalization param of mel spec
    std (float): normalization param of mel spec
    mel_basis (np.array): optional pre-computed mel basis to save computational
      time if passed. If not passed, it will call librosa to construct one
    htk (bool): whther to compute the mel spec with the htk or slaney algorithm
    norm: Should be None for htk, and 1 for slaney

  Returns:
    np.array: mag_spec with shape [time, n_fft/2 + 1]
  """
  if mel_basis is None:
    mel_basis = librosa.filters.mel(
        fs,
        n_fft,
        n_mels=n_mels,
        htk=htk,
        norm=norm
    )
  if feature_normalize:
    log_mel_spec = denormalize(log_mel_spec, mean, std)
  mel_spec = np.exp(log_mel_spec)
  mag_spec = np.dot(mel_spec, mel_basis)
  mag_spec = np.power(mag_spec, 1. / power)
  return mag_spec