NOISANAL

NAME

noisanal - estimate quantity of aperiodicity in signal

SYNOPSIS

noisanal (-i item) (-w windowsize) (-e) (-n) file

DESCRIPTION

noisanal estimates the degree of aperiodicity in a signal as a function of time. It uses a type of autocorrelation measure to estimate what proportion of the signal is predictable at each time instant. The output is a track item normalised to overall energy. The measure is probably irrelevant in silent intervals.

Options and their meanings are:

-I Identify program and exit.

-i item Select input item number and type.

-w windowsize Set the analysis window size. Default 30ms.

-e Output energy track as well as aperiodicity track.

-n Normalise energy track to 0..1

INPUT ITEMS

1.xx Speech waveform.

OUTPUT ITEMS

16 Periodicity estimate by autocorrelation.

16 Overall energy track (decibels).

VERSION/AUTHOR

1.0 - Mark Huckvale.

SEE ALSO

fxac(SFS1)
Fri Jul 09 14:54:43 2004