spectran - transform speech waveforms to spectral data by fourier transform


spectran [(-b bandwidth)|(-w window) (-s stepsize|-o overlap)] (-n numpixel) (-t offset) (-c) (-p) (-i item) file


spectran performs a fourier transform on selected windows of a speech signal to produce a set of spectral coefficients suitable for spectrographic examination. The windows may be of fixed length with a selected overlap or selected stepsize, or may be positioned by means of a TX item (in which case the unvoiced periods are analysed according to the fixed length parameters). A Hamming window is used for all analyses. Pre-emphasis may be de-selected. This program does not use the array processor.

Options and their meanings are:

-I Identify program and version number.

-i item Select input item number.

-b bandwidth Select analysis bandwidth in Hz. (default 300Hz). This can be used as an alternative to specifying window size directly.

-w window Select analysis window size in milliseconds. Default is calculated from analysis bandwidth.

-s stepsize Select analysis window step size in milliseconds. Default is calculated from analysis bandwidth.

-o overlap Select analysis window overlap in milliseconds. Default is calculated from analysis bandwidth.

-n numpixel Specify number of energies returned in spectrum. Default 128. Must be power of 2 between 32 and 16384.

-t offset Select excitation synchronous analysis with supplied offset in micro-seconds between tx markers and start point of analysis window. Negative is earlier, positive later, default is 0.

-c Use cepstral filtering to smooth spectrum (cut off at 2ms).

-p Do not pre-emphasize speech data.


1.xx Any speech waveform.

3.xx (Optional) Excitation markers.


11 Spectral coefficients.


1.5 - Mark Huckvale (from flfft and others by Lynn Whitaker & David Pearce)
Fri Jul 09 14:54:36 2004