Main Content

enhanceSpeech

Enhance speech signal

Since R2024a

collapse all in page

    Syntax

    audioOut = enhanceSpeech(audioIn,fs)
    enhanceSpeech(audioIn,fs)

    Description

    example

    audioOut = enhanceSpeech(audioIn,fs) enhances the speech in the audio signal by reducing non-speech sounds.

    example

    enhanceSpeech(audioIn,fs) with no output arguments displays a plot of the original and enhanced speech.

    This function requires both Audio Toolbox™ and Deep Learning Toolbox™.

    Examples

    collapse all

    This example uses:

    Open Live Script

    Try calling enhanceSpeech in the command line. If the required model files are not installed, then the function throws an error and provides a link to download them. Click the link, and unzip the file to a location on the MATLAB path.

    Alternatively, execute the following commands to download and unzip the enhanceSpeech model files to your temporary directory.

    downloadFolder = fullfile(tempdir,"enhanceSpeechDownload");
loc = websave(downloadFolder,"https://ssd.mathworks.com/supportfiles/audio/enhanceSpeech.zip");
modelsLocation = tempdir;
unzip(loc,modelsLocation)
addpath(fullfile(modelsLocation,"enhanceSpeech"))

    This example uses:

    Open Live Script

    Read in an audio file containing speech and noise. Listen to the signal.

    [noisySpeech,fs] = audioread("NoisySpeech-16-mono-3secs.ogg");
sound(noisySpeech,fs)

    Use enhanceSpeech to reduce the non-speech sounds in the signal. Listen to the enhanced signal.

    enhancedSpeech = enhanceSpeech(noisySpeech,fs);
sound(enhancedSpeech,fs)

    Call enhanceSpeech with no output arguments to plot both the noisy signal and the enhanced signal.

    enhanceSpeech(noisySpeech,fs);

    Figure contains 4 axes objects. Axes object 1 with title Time Analysis, ylabel Audio Input contains an object of type line. Axes object 2 with title Spectral Analysis, ylabel Frequency (kHz) contains an object of type image. Axes object 3 with xlabel Time (s), ylabel Audio Output contains an object of type line. Axes object 4 with xlabel Time (s), ylabel Frequency (kHz) contains an object of type image.

    This example uses:

    Open Live Script

    Read in an audio file containing speech and noise. Also read in an audio file containing the original clean speech to use as a reference signal.

    [noisySpeech,fs] = audioread("NoisySpeech-16-mono-3secs.ogg");
reference = audioread("CleanSpeech-16-mono-3secs.ogg");

    Calculate the STOI metric for the noisy speech signal using stoi.

    noisySpeechSTOI = stoi(noisySpeech,reference,fs)
    noisySpeechSTOI = 0.8370

    Use enhanceSpeech to enhance the speech signal. Evaluate the enhanced signal using the STOI metric and see the improvement compared to the STOI of the noisy signal.

    enhancedSpeech = enhanceSpeech(noisySpeech,fs);
enhancedSpeechSTOI = stoi(enhancedSpeech,reference,fs)
    enhancedSpeechSTOI = single
    0.8808

    This example uses:

    Open Live Script

    Read in an audio file containing speech and noise. Also read in an audio file containing the original clean speech to use as a reference signal.

    [noisySpeech,fs] = audioread("NoisySpeech-16-mono-3secs.ogg");
reference = audioread("CleanSpeech-16-mono-3secs.ogg");

    Calculate the ViSQOL metric for the noisy speech signal using visqol.

    noisySpeechMOS = visqol(noisySpeech,reference,fs,Mode="speech")
    noisySpeechMOS = 2.9550

    Use enhanceSpeech to enhance the speech signal. Evaluate the enhanced signal using the ViSQOL metric and see the improvement compared to the noisy signal.

    enhancedSpeech = enhanceSpeech(noisySpeech,fs);
enhancedSpeechMOS = visqol(enhancedSpeech,reference,fs,Mode="speech")
    enhancedSpeechMOS = single
    3.2205

    Input Arguments

    collapse all

    Audio input containing the speech signal to enhance, specified as a column vector (single channel).

    Data Types: single | double

    Sample rate in Hz, specified as a positive scalar.

    Data Types: single | double

    Output Arguments

    collapse all

    Audio output containing the enhanced speech signal, returned as a column vector with the same size and sample rate as the input signal.

    Data Types: single

    Algorithms

    The enhanceSpeech function uses a pretrained MetricGAN-OKD [1] neural network to enhance speech signals.

    References

    [1] Shin, Wooseok, Byung Hoon Lee, Jin Sob Kim, Hyun Joon Park, and Sung Won Han. "MetricGAN-OKD: multi-metric optimization of MetricGAN via online knowledge distillation for speech enhancement." In International Conference on Machine Learning, pp. 31521-31538. PMLR, 2023.

    Extended Capabilities

    C/C++ Code Generation
    Generate C and C++ code using MATLAB® Coder™.

    GPU Arrays
    Accelerate code by running on a graphics processing unit (GPU) using Parallel Computing Toolbox™.

    Version History

    Introduced in R2024a

    See Also

    | | |