Skip to content

TimeStretch

Added in v0.2.0

Change the speed or duration of the signal without changing the pitch. This transform lets you choose between method="signalsmith_stretch" and method="librosa_phase_vocoder". If you need other time stretching methods, consider the following alternatives:

Input-output example

In this example we speed up a sound by 25% (i.e. a rate of 1.25), using the signalsmith_stretch method:

Input-output waveforms and spectrograms

Input sound Transformed sound

Usage example

from audiomentations import TimeStretch

transform = TimeStretch(
    min_rate=0.8,
    max_rate=1.25,
    leave_length_unchanged=True,
    p=1.0
)

augmented_sound = transform(my_waveform_ndarray, sample_rate=16000)

TimeStretch API

min_rate: float • range: [0.1, 10.0]
Default: 0.8. Minimum time-stretch rate. Values less than 1.0 slow down the audio (reduce the playback speed).
max_rate: float • range: [0.1, 10.0]
Default: 1.25. Maximum time-stretch rate. Values greater than 1.0 speed up the audio (increase the playback speed).
leave_length_unchanged: bool
Default: True. If True, the output audio will have the same duration as the input audio. If False, the duration of the output audio will be altered by the time-stretch rate.
method: str • choices: "librosa_phase_vocoder", "signalsmith_stretch"

Default: "signalsmith_stretch".

  • "signalsmith_stretch": Use signalsmith-stretch. It is 50-100% faster than librosa_phase_vocoder, and provides significantly higher perceived audio quality.
  • "librosa_phase_vocoder": Use librosa.effects.time_stretch. Pro: Supports any number of channels. Con: phase vocoding can significantly degrade the audio quality by "smearing" transient sounds, altering the timbre of harmonic sounds, and distorting pitch modulations. This may result in a loss of sharpness, clarity, or naturalness in the transformed audio.
p: float • range: [0.0, 1.0]
Default: 0.5. The probability of applying this transform.

Source code

audiomentations/augmentations/time_stretch.py