Did some surfing today and found some very interesting pages:
visit http://www.eas.asu.edu/~speech/ndtc/ and download
"A Review of Algorithms for Perceptual Coding of Digital Audio Signals."
I haven't read this yet, but it looks good & much more details
than the Pan article.
But even more interesting is:
http://www-ccrma.stanford.edu/~bosse/
(Any of you know of Bo Lincoln, [EMAIL PROTECTED]?)
There is a detailed report (& source code). The codec has
temporal masking and some kind of tonal model, but not joint stereo.
He also mentions a wavelet based codec where "the authors claim
perceptually nealy lossless coding at about 80 kbs at a coding
complexity which requires only about 50% of the processing
power of a pentium-75 for either decoding or encoding"
ICASSP 97, pp 327-330. Hard to believe.
Mark
