Wolfram Language

Audio Effects via STFT Transformations

AudioSpectralMap is a function that allows for modifications of signals in the time-frequency representation of the short-time Fourier transform (STFT). It is easy to manipulate the STFT to reduce the background noise in a time-and-frequency-dependent way.

Start with a noisy audio signal.

Define a function that decreases the amplitude of quiet components.

Apply the function to the STFT using AudioSpectralMap.

AudioSpectralTransformation is another useful function for manipulating signals from a time-frequency representation. It constructs a new signal by computing a backward transformation between the input and output signals' STFTs.

One can create "special effects"such as fisheyein the frequency domain by defining such a mapping. Start with an amusing signal.

show complete Wolfram Language input

Define a transformation that creates a fisheye effect.

Apply the function to the input signal using AudioSpectralTransformation.

Related Examples

de es fr ja ko pt-br zh