---

Identify Speakers with sndpeek

[ Thanks to Jayne
for this link. ]

“Reliable authentication of an individual through voice-print
analysis is complex and difficult. However, sndpeek and some custom
algorithms can provide a voice-print matching configuration of
considerably reduced complexity while retaining a great deal of
usefulness. This article demonstrates the tools and code you’ll
need to modify sndpeek to record individual voice-print files for a
given speaker. Each of these files are then compared to the
incoming real-time audio stream to provide best-guess matches and
visualizations for the current speaker…”


Complete Story

Related Story:
IBM
DeveloperWorks: Whistle while you work
(Jan 12, 2007)

Get the Free Newsletter!

Subscribe to Developer Insider for top news, trends, & analysis