utils#

diffsptk.get_alpha(sr: int, mode: str = 'hts', n_freq: int = 10, n_alpha: int = 100) → float[source]#

Compute an appropriate frequency warping factor given the sample rate.

Parameters:

srint >= 1: The sample rate in Hz.
mode[‘hts’, ‘auto’]: ‘hts’ returns a traditional alpha used in HTS. ‘auto’ computes an appropriate alpha in the L2 sense.
n_freqint >= 2: The number of sample points in the frequency domain.
n_alphaint >= 1: The number of sample points to search alpha.

Returns:

Examples

>>> _, sr = diffsptk.read("assets/data.wav")
>>> alpha = diffsptk.get_alpha(sr)
>>> alpha
0.42

diffsptk.read(filename: str, device: device | None = None, dtype: dtype | None = None, **kwargs) → tuple[Tensor, int][source]#

Read a waveform from the given file.

Parameters:

filenamestr: The path of the wav file.
devicetorch.device or None: The device of the returned tensor.
dtypetorch.dtype or None: The data type of the returned tensor.
**kwargsadditional keyword arguments: Additional arguments passed to soundfile.read.

Returns:

Examples

>>> x, sr = diffsptk.read("assets/data.wav")
>>> x
tensor([ 0.0002,  0.0004,  0.0006,  ...,  0.0006, -0.0006, -0.0007])
>>> sr
16000

diffsptk.write(filename: str, x: Tensor, sr: int, **kwargs) → None[source]#

Write the given waveform to a file.

Parameters:

filenamestr: The path of the wav file.
xTensor: The waveform.
srint: The sample rate in Hz.
**kwargsadditional keyword arguments: Additional arguments passed to soundfile.write.

Examples

>>> x, sr = diffsptk.read("assets/data.wav")
>>> diffsptk.write("out.wav", x, sr)