mindspore.dataset.audio.PhaseVocoder

class mindspore.dataset.audio.PhaseVocoder(rate, phase_advance)[source]

Given a STFT spectrogram, speed up in time without modifying pitch by a factor of rate.

Parameters

rate (float) – Speed-up factor.
phase_advance (numpy.ndarray) – Expected phase advance in each bin, in shape of (freq, 1).

Raises

TypeError – If rate is not of type float.
ValueError – If rate is not a positive number.
TypeError – If phase_advance is not of type numpy.ndarray .
RuntimeError – If input tensor is not in shape of <…, freq, num_frame, complex=2>.

Supported Platforms:: CPU

Examples

>>> import numpy as np
>>> import mindspore.dataset as ds
>>> import mindspore.dataset.audio as audio
>>>
>>> # Use the transform in dataset pipeline mode
>>> waveform = np.random.random([5, 44, 10, 2])  # 5 samples
>>> numpy_slices_dataset = ds.NumpySlicesDataset(data=waveform, column_names=["audio"])
>>> transforms = [audio.PhaseVocoder(rate=2, phase_advance=np.random.random([44, 1]))]
>>> numpy_slices_dataset = numpy_slices_dataset.map(operations=transforms, input_columns=["audio"])
>>> for item in numpy_slices_dataset.create_dict_iterator(num_epochs=1, output_numpy=True):
...     print(item["audio"].shape, item["audio"].dtype)
...     break
(44, 5, 2) float64
>>>
>>> # Use the transform in eager mode
>>> waveform = np.random.random([44, 10, 2])  # 1 sample
>>> output = audio.PhaseVocoder(rate=2, phase_advance=np.random.random([44, 1]))(waveform)
>>> print(output.shape, output.dtype)
(44, 5, 2) float64

Tutorial Examples:

Illustration of audio transforms