Estimation of the Signal Short-Time Stationarity Duration

Estimation of the short-time stationarity duration (in terms of frame lengths) of a given nonstationary signal using a novel method.

The present code is a Matlab function that implements a novel technique named “statiogram” concerning the estimation of the short-time stationarity frame duration of a given nonstationary signal. The procedure answers what should be the length of the frames into which a nonstationary signal is divided so that the signal be considered as short-time wide-sense stationary in the boundaries of the (most of the) frames, with some predefined tolerance. This is of particular importance when one is going to perform some kind of time-localized analysis (e.g., STFT or signal feature extraction), in the fields of speech and sound processing; vibration and seismic analysis; neuroscience, etc.

A few examples are given to clarify the usage of the function. For convenience, the input and output arguments are given at the beginning of the function.

The code is based on the theory described in:

[1] H. Zhivomirov. An Empirical Method for Signal Stationarity Estimation. The Scientific Bulletin, Series C: Electrical Engineering and Computer Science, ISSN: 2286-3540, Vol. 86, Issue 2, pp. 243-256, 2024. (https://www.scientificbulletin.upb.ro/rev_docs_arhiva/rezb81_132280).

[2] --- A research article dedicated to this estimation technique is about to come by the end of March 2025. ---