Spectrogram (Short-time FT log magnitude)

Computes the Short-time Fourier transform and displays as grayscale image or 3-D surface

This program is a modification and extension to Paul Kienzle's spectrogram program for Octave. It is for Matlab 6r12.1. The short-time fourier transform of a signal is computed by FFTs of overlapping windowed signal segments. You can then exit with the complete STFT, as a matrix, or you can go on to plot first a grayscale pseudocolor spectrogram with interpolated shading, and then a color 3-D surface spectrogram viewed from an optimum angle.
I made this because the specgram function in the Matlab toolbox is a joke, it doesn't even use a Fast Fourier Transform, electing to compute the DFT directly in its algorithm. Second, neither that algorithm nor Paul's original program used the pseudo plot, which allows interpolated shading of the colors, which makes the spectrogram look acceptable. The other way (i.e.\ Matlab SP toolbox, Paul's program) of displaying a 2-D plot with a third dimension as color is through an indexed image, but this doesn't allow interpolated shading, and so the spectrograms are totally useless for speech analysis.