Audio Toolbox

Design and analyze speech, acoustic, and audio processing systems

Audio Toolbox provides signal processing and analysis tools for audio, speech, and acoustics. It includes algorithms for processing audio signals, estimating acoustic metrics, labeling and augmenting audio data sets, and extracting audio features.

You can run measurements or prototype algorithms in real time by streaming low-latency audio to and from ASIO, CoreAudio, and other sound cards. The toolbox lets you control algorithm parameters via graphical interfaces or MIDI events. You can validate your algorithm by turning it into a VST or Audio Unit plugin to run in external host applications. The toolbox also offers plugin hosting, so you can process MATLAB arrays using external audio plugins.

The toolbox includes pre-trained machine learning and deep learning models that support transfer learning. You can apply the models directly to speech and acoustic signals for high-level tasks such as embedding extraction, sound classification, speaker verification, speech transcription and synthesis, source separation, and background noise reduction.

Audio Toolbox screenshot of speech command recognition using deep learning, showing two versions side-by-side.

Streaming Acquisition and Playback

Read and write audio samples from and to sounds cards (such as USB or Thunderbolt™) using standard audio drivers (such as ASIO, WASAPI, CoreAudio, and ALSA) across Windows^®, Mac^®, and Linux^® operating systems. Process live audio in MATLAB with milliseconds of round-trip latency.

Real-Time Audio in MATLAB

Documentation | Examples

Stacked plot of a recorded waveform including two speakers, followed by the separated tracks of the two respective speakers and the residual signal. The sums of the separated components is showed to match the original signal.

AI for Audio, Speech, and Acoustics

Apply the latest deep learning and machine learning models to audio, speech, and acoustic signals. Create, label, and augment audio data for tuning models using transfer learning. Extract features and compute time-frequency transformations. Develop predictive models with Statistics and Machine Learning Toolbox and Deep Learning Toolbox.

Pretrained AI models

Documentation | Examples

Audio Processing Algorithms

Generate standard waveforms, apply common audio effects, and design audio processing systems with dynamic parameter tuning and live visualization in MATLAB and Simulink.

Parametric Equalizer Design

Documentation | Examples

Simulink model of a two-way loudspeaker, including DSP modules like equalizers and compressors, as well as a circuit-based electro-mechanical model of a loudspeaker drive, complete with visualization scopes and spectrum analyzers.

System Modeling with Simulink

Design system models using libraries of audio processing blocks for Simulink. Tune parameters and visualize system behavior using interactive controls and dynamic plots. Simulate DSP, analog circuits, and deep learning models.

Documentation | Examples

Live Prototyping and Visualization

Automatically create user interfaces for tunable parameters of audio processing algorithms. Test algorithms with the Audio Test Bench app and tune parameters in running programs with auto-generated interactive controls via MIDI.

Documentation | Examples

The Impulse Response Measurer app shows controls to select estimation methods and parameters, details of the measured time-domain impulse response, and the frequency-domain magnitude and phase.

Room Acoustics and Spatial Audio

Measure room impulse responses using maximum-length sequences (MLS) and Exponential Swept Sinusoids (ESS), read and write SOFA files, analyze and interpolate head-related transfer functions (HRTF), and encode and decode ambisonic formats. Run efficient convolutions using partitioned frequency-domain methods.

Documentation | Examples

Psychoacoustics and Loudness Metering

Apply sound pressure level (SPL) meters and loudness meters to recorded or live signals. Analyze signals with octave and fractional-octave filters. Apply standard-compliant A-, C-, or K-weighting filters to raw recordings. Monitor peak and true peak values. Measure acoustic sharpness, roughness, and fluctuation strength.

Documentation | Examples

User Interface of a binary plugin generated with MATLAB and running in REAPER. The plugin shows nine sliders and knobs for the nine tunable parameters of a three-band parametric equalizers, with the same look and feel of a user-designed MATLAB App.

Audio Plugin Creation and Hosting

Generate VST plugins, AU plugins, and standalone executable plugins directly from MATLAB code without requiring manual design of user interfaces. Use external VST and AU plugins as regular MATLAB objects to process MATLAB arrays, changing plugin parameters programmatically, with user interfaces or MIDI controls.

Design your first audio plugin using MATLAB

Documentation | Examples

A RaspberryPi board, representing an example of embedded platform usable to prototype audio system models developed in MATLAB or Simulink, without additional embedded software efforts.

Target Embedded and Real-Time Audio Systems

With MATLAB and Simulink coder products, generate C and C++ source code from signal processing and machine learning algorithms provided as toolbox functions, objects, and blocks. Generate CUDA^® source code from select feature extraction functions. Prototype audio processing designs on Raspberry Pi™, mobile apps for Android^® or iOS devices, Speedgoat audio machines, and ST Discovery boards.

Documentation | Examples