On Sun, Aug 17, 2003 at 10:53:03PM -0600, Andrew Jorgensen wrote: > I definitely like the idea of a speech-optimized codec from the makers > of Ogg, but I'm a little confused that there'd be a need. The main > difference between voice and music is the sampling rate required.
I'm not so sure about this. The sample domain for voice is much more
restricted than that of general-purpose sound. This makes it easier
to apply compression techniques that are specific to the frequency
ranges and sound constructs that a larynx, tongue, and mouth produce.
Compression techniques that might destroy the sound of a violin or a
guitar may work perfectly fine with a ``th'' in ``the''. If I were
developing an audio compression algorithm, I would be delighted to
know that I would get to ignore the possibility of an electric guitar
cord being played in the background of my audio sample. The narrower
you can restrict the domain of information that you compress, the more
liberal you can get with your approaches.
Mike
--
------------------------------------------- | ---------------------
Michael Halcrow | [EMAIL PROTECTED]
Developer, IBM Linux Technology Center |
|
The wonderful thing about a dancing bear is |
not how well he dances, but that he dances |
at all. |
------------------------------------------- | ---------------------
GnuPG Keyprint: 05B5 08A8 713A 64C1 D35D 2371 2D3C FDDA 3EB6 601D
pgp00000.pgp
Description: PGP signature
____________________ BYU Unix Users Group http://uug.byu.edu/ ___________________________________________________________________ List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list
