def feat_extraction(in_wav_dir, file_name_token, out_feats_dir):
# Display:
print("\nAnalysing file: " + file_name_token + '.wav............................')
# File setup:
wav_file = os.path.join(in_wav_dir, file_name_token + '.wav')
mp.analysis_compressed(wav_file, out_dir=out_feats_dir)
return
def feat_extraction(in_wav_dir, file_name_token, out_feats_dir):
# Display:
print("\nAnalysing file: " + file_name_token +
'.wav............................')
# File setup:
wav_file = os.path.join(in_wav_dir, file_name_token + '.wav')
mp.analysis_compressed(wav_file, out_dir=out_feats_dir)
return
def feat_extraction(wav_file, out_feats_dir):
# Parsing path:
file_name_token = os.path.basename(os.path.splitext(wav_file)[0])
# Display:
print("Analysing file: " + file_name_token + '.wav' +
'................................')
# Files setup:
est_file = os.path.join(out_feats_dir, file_name_token + '.est')
# Epochs detection:
la.reaper(wav_file, est_file)
# Feature extraction:
m_mag_mel_log, m_real_mel, m_imag_mel, v_lf0, v_shift, fs, fft_len = mp.analysis_compressed(
wav_file)
if fs != fs_expected:
print(
"The wavefile's sample rate (%dHz) does not match the expected sample rate (%dHz)."
% (fs, fs_expected))
sys.exit(1)
# Zeros for unvoiced segments in phase features:
v_voi = (np.exp(v_lf0) > 5.0).astype(int) # 5.0: tolerance (just in case)
m_real_mel_zeros = m_real_mel * v_voi[:, None]
m_imag_mel_zeros = m_imag_mel * v_voi[:, None]
# Saving features:
lu.write_binfile(m_mag_mel_log,
out_feats_dir + '/' + file_name_token + '.mag')
lu.write_binfile(m_real_mel_zeros,
out_feats_dir + '/' + file_name_token + '.real')
lu.write_binfile(m_imag_mel_zeros,
out_feats_dir + '/' + file_name_token + '.imag')
lu.write_binfile(v_lf0, out_feats_dir + '/' + file_name_token + '.lf0')
# Saving auxiliary feature shift (hop length). It is useful for posterior modifications of labels in Merlin.
lu.write_binfile(v_shift, out_feats_dir + '/' + file_name_token + '.shift')
return
b_plots = True # True if you want to plot the extracted parameters.
# INPUT:==============================================================================
wav_file_orig = 'data_48k/wavs_nat/hvd_593.wav' # Original natural wave file. You can choose anyone provided in the /wavs_nat directory.
b_const_rate = False
mag_dim = 100
phase_dim = 45
# PROCESS:============================================================================
lu.mkdir(out_dir)
# ANALYSIS:
print("Analysing.....................................................")
m_mag_mel_log, m_real_mel, m_imag_mel, v_lf0_smth, v_shift, fs, fft_len = mp.analysis_compressed(
wav_file_orig,
mag_dim=mag_dim,
phase_dim=phase_dim,
b_const_rate=b_const_rate)
# MODIFICATIONS:
# You can modify the parameters here if wanted.
# SYNTHESIS:
print("Synthesising.................................................")
v_syn_sig = mp.synthesis_from_compressed(m_mag_mel_log,
m_real_mel,
m_imag_mel,
v_lf0_smth,
fs,
b_const_rate=b_const_rate,
b_out_hpf=False)
return
if __name__ == '__main__':
# INPUT:==============================================================================
wav_file_orig = 'data_48k/wavs_nat/hvd_593.wav' # Original natural wave file. You can choose anyone provided in the /wavs_nat directory.
out_dir = 'data_48k/wavs_syn' # Where the synthesised waveform will be stored.
b_plots = True # True if you want to plot the extracted parameters.
# PROCESS:============================================================================
lu.mkdir(out_dir)
# ANALYSIS:
print("Analysing.....................................................")
m_mag_mel_log, m_real_mel, m_imag_mel, v_lf0, v_shift, fs, fft_len = mp.analysis_compressed(
wav_file_orig)
# MODIFICATIONS:
# You can modify the parameters here if wanted.
# SYNTHESIS:
print("Synthesising.................................................")
v_syn_sig = mp.synthesis_from_compressed(m_mag_mel_log, m_real_mel,
m_imag_mel, v_lf0, fs, fft_len)
# SAVE WAV FILE:
print("Saving wav file..............................................")
wav_file_syn = out_dir + '/' + lu.get_filename(
wav_file_orig) + '_copy_syn_low_dim.wav'
la.write_audio_file(wav_file_syn, v_syn_sig, fs)