Browse Prior Art Database

Identifying sound source from mixed audio by audio watermark

IP.com Disclosure Number: IPCOM000014491D
Original Publication Date: 2001-Dec-30
Included in the Prior Art Database: 2003-Jun-19
Document File: 1 page(s) / 27K

Publishing Venue

IBM

Abstract

Disclosed is a system for identifying a sound source from mixed sound data. This system embeds a watermark in each sound source before mixing. By detecting the watermark from the mixed sound, this system finds out the sound source at anytime. The embedded watermark has 1 bit information for each sound source. When a certain sound source makes some sound, the bit corresponding to such sound source is set to ON, and otherwise set to OFF. By using this system, for example, we can easily identify who said a certain opinion after a meeting by detecting the watermark within voice data recorded at the meeting.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 100% of the total text.

Page 1 of 1

Identifying sound source from mixed audio by audio watermark

Disclosed is a system for identifying a sound source from mixed sound data. This system embeds a
watermark in each sound source before mixing. By detecting the watermark from the mixed sound,
this system finds out the sound source at anytime. The embedded watermark has 1 bit information
for each sound source. When a certain sound source makes some sound, the bit corresponding to
such sound source is set to ON, and otherwise set to OFF. By using this system, for example, we
can easily identify who said a certain opinion after a meeting by detecting the watermark within
voice data recorded at the meeting.

It was difficult to identify the sound source from mixed sounds if each sound source's frequency
range is near. This system can find out the sound source speedy and accurately.

This system has following applications.
1. Identifying the speaker from a discussion tape
2. Identifying the speaker on a telephone conference
3. Automatically recording who speaks what by using with a voice recognition system

1