spektrogrammin

Spektrogrammin is a term sometimes used to refer to the spectrogram, a visual representation of how the spectral content of a signal evolves over time. In practice, a spectrogram is created by calculating the short-time Fourier transform (STFT) of the signal: the signal is divided into overlapping frames, a spectrum is computed for each frame, and the magnitudes are displayed as a 2D image with time on the horizontal axis and frequency on the vertical axis. The intensity or color of each point encodes the signal’s energy at a given frequency and moment.

Parameters such as window length, overlap, and window function (for example Hann or Hamming) determine the balance between time and frequency resolution. Short windows yield good time resolution but poorer frequency resolution, while longer windows have the opposite effect.

Variants include the mel-spectrogram, which maps frequencies onto a perceptually motivated scale, and log-spectrograms, which display magnitude on a logarithmic scale. In many speech and music applications, amplitude may be converted to decibels.

Applications span speech recognition, phonetics and linguistics, music information retrieval, bioacoustics, and seismic analysis. They enable inspection of formant movement, pitch, timbre, and transient events, and are often used as input features for machine learning models.

Limitations include resolution trade-offs, sensitivity to preprocessing choices, and lack of phase information in standard magnitude spectrograms. Reconstructing the original signal requires additional processing, and artifacts can arise from windowing and calibration.

Related topics include the Fourier transform, short-time Fourier transform, and filter banks (such as mel filter banks).