[ Japanese | English | Back to top ]
Last update: "2005/09/29 04:03:21"

MKSS(1) MKSS(1)

NAME

mkss - compute average spectrum from mic input for SS

SYNOPSIS

mkss [options..] filename

DESCRIPTION

mkss is a tool for estimating noise spectrum for spectral subtraction on Julius. It reads a few seconds of sound data from microphone input, and saves the average spectrum to a file. The output file can be used as a noise spec- trum data in Julius (option "-ssload"). Output format is RAW (no header), 16bit (signed short, big endian), monoral. If the file already exist, it will be overridden.

OPTIONS

-freq threshold Sampling frequency (Hz, default=16000) -len msec length to record in milliseconds (default: 3000). -fsize samplenum frame size in number of samples (default: 400). -fshift samplenum frame shift in number of samples (default: 160).

USAGE

Start mkss and stay quiet for 3 seconds. Then the average noise spectrum will be recorded to the specified file. When you start Julius, specify option "-ssload thatfile", and you can use microphone input with spectral subtrac- tion.

SEE ALSO

julius(1)

COPYRIGHT

Copyright (c) 1991-2005 Kyoto University, Japan Copyright (c) 2000-2005 Nara Institute of Science and Technology, Japan Copyright (c) 2005 Nagoya Institute of Technology, Japan

AUTHORS

LEE Akinobu (Nagoya Institute of Technology, Japan) contact: julius@kuis.kyoto-u.ac.jp

LICENSE

Same as Julius. LOCAL MKSS(1)

$Id: mkss.html.en,v 1.1.1.1 2007/01/10 08:01:57 kudravka_ Exp $