MPEG7EASE 1.0 - MPEG-7 Extract Audio Spectrum Envelope
By Tak-Shing Chan <chan12@sourceforge.net>


0. Important

If you use this program, you might be liable to MPEG-7 licensing
fees.  Caveat emptor.


1. Introduction

This package consists of two programs: EASE and WASP.


1.1. EASE (Extract Audio Spectrum Envelope)

This program extracts audio spectrum envelopes from monaural
sound files.  The extraction algorithm conforms to the
International Standard ISO/IEC 15938-4:2002 (also known as MPEG-7
Audio).

When invoked with the -i switch, the resulting output would no
longer be standard-compliant.  Instead, it would be in an
intermediate format, where the data consists of an int which
denotes the column size n, followed by a mxn matrix in double
precision, with m frames and n bins in row major order
(n * i + j).

Synopsis: ease [-i] [-h hopsize] [-w winsize] [-l loEdge]
               [-u hiEdge] [-o octaveResolution] infile outfile

Example: ease -i -l 545.25 -u 6727 -o 1/16 test.wav test.eas

Note that hopsize should be specified using mediaDurationType.


1.2. WASP (Write Audio Spectrum Projection)

This program reads a binary input file, extracts its audio
spectrum projection, and writes the extraction to the output file
as tab-delimited text.  Optionally, its audio spectrum basis can
be extracted to a separate text file.

The binary input file shall contain an audio spectrum envelope in
the intermediate EASE format.

The extraction algorithm conforms to the MPEG-7 standard.

Synposis: wasp [-n k] [-t icatype] [-b basisfile] infile outfile

Valid values for icatype:
     0    No ICA
     1    Symmetric fastICA (tanh)
     2    Jade
     3    Extended infomax

Example: wasp -n 15 -t 1 -b test.bas test.eas test.was


2. Compilation

If you wish to create autoconf files from scratch, try:

    % aclocal -I .
    % autoheader
    % automake --add-missing --copy
    % autoconf

Otherwise just follow the instructions in the INSTALL file.


3. References

ISO/IEC 15938-4:2002, Information Technology -- Multimedia
     Content Description Interface -- Part 4: Audio.
