Hi all, This should please all those who want to know more about Speex or Vorbis. Monty and I have just finished writing this paper: "Improved Noise Weighting in CELP Coding of Speech - Applying the Vorbis Psychoacoustic Model To Speex" which you can get at: http://people.xiph.org/~jm/papers/aes120_speex_vorbis.pdf It's probably the best description of the Speex encoder to data and it has a lot of details about the Vorbis psycho-acoustic model. For those who actually want to try the improvement described in the paper, just configure Speex with --enable-vorbis-psy . Jean-Marc P.S. This will be presented at the next AES convention in Paris (May).
On Mon, Mar 20, 2006 at 03:55:33PM +1100, Jean-Marc Valin wrote:> This should please all those who want to know more about Speex or > Vorbis. Monty and I have just finished writing this paper: "Improved > Noise Weighting in CELP Coding of Speech - Applying the Vorbis > Psychoacoustic Model To Speex" which you can get at: > http://people.xiph.org/~jm/papers/aes120_speex_vorbis.pdfCool, thanks for posting a link. BTW, for the second footnote in the Introduction, http://xiph.org/vorbis/ would be a better url to give. There's no website at vorbis.org, though we do own the domain. If they'll let you, I'd also reference the vorbis spec itself at http://xiph.org/vorbis/doc/Vorbis_I_spec.html -r
Le dimanche 19 mars 2006 ? 23:30 -0800, Ralph Giles a ?crit :> On Mon, Mar 20, 2006 at 03:55:33PM +1100, Jean-Marc Valin wrote: > > > This should please all those who want to know more about Speex or > > Vorbis. Monty and I have just finished writing this paper: "Improved > > Noise Weighting in CELP Coding of Speech - Applying the Vorbis > > Psychoacoustic Model To Speex" which you can get at: > > http://people.xiph.org/~jm/papers/aes120_speex_vorbis.pdf > > Cool, thanks for posting a link. > > BTW, for the second footnote in the Introduction, http://xiph.org/vorbis/ > would be a better url to give. There's no website at vorbis.org, though > we do own the domain.Oops. Could we add a redirect?> If they'll let you, I'd also reference the vorbis spec itself at > http://xiph.org/vorbis/doc/Vorbis_I_spec.htmlWould have been a good idea. Too late to change unfortunately. Jean-Marc