internal__normalize_audio

Audio normalization of a tensor in-place. The normalization can be a bool,
a number, or a function that takes the audio tensor as an input. SoX uses
32-bit signed integers internally, thus bool normalizes based on that assumption.

internal

Provides access to datasets, models and processing
facilities for deep learning in audio.

Sigrid Keydana

torchaudio

R Interface to 'pytorch''s 'torchaudio'

Athos Damiani

Daniel Falbel

internal__normalize_audio function

<dl><dt>signal</dt>
<dd>(Tensor): waveform</dd>
<dt>normalization</dt>
<dd>(bool, int or function): Optional normalization.
If boolean <code>TRUE</code>, then output is divided by <code>2^31</code>.
Assuming the input is signed 32-bit audio, this normalizes to <code>[-1, 1]</code>.
If <code>numeric</code>, then output is divided by that number.
If <code>function</code>, then the output is passed as a paramete to the given function,
then the output is divided by the result. (Default: <code>TRUE</code>)</dd></dl>

Arguments

Audio Normalization — internal__normalize_audio

<dl>

<dt>signal</dt>
<dd>(Tensor): waveform</dd>


<dt>normalization</dt>
<dd>(bool, int or function): Optional normalization.
If boolean <code>TRUE</code>, then output is divided by <code>2^31</code>.
Assuming the input is signed 32-bit audio, this normalizes to <code>[-1, 1]</code>.
If <code>numeric</code>, then output is divided by that number.
If <code>function</code>, then the output is passed as a paramete to the given function,
then the output is divided by the result. (Default: <code>TRUE</code>)</dd>

</dl>

internal__normalize_audio: Audio Normalization

Description

Usage

Arguments