Fourier Transform / STFT

The Fourier Transform is a mathematical operation that decomposes a signal into the individual frequencies it contains. Applied to audio, the short-time variant (STFT) slices the signal into short, overlapping frames and transforms each one, so frequency content can be tracked over time rather than averaged across the whole clip. The STFT's time-by-frequency output is exactly what a spectrogram displays, and it is the standard front end for most audio AI and source-separation models.