search for: lsp_to_curve

It seems to me that in this version of the function using float p=.7071067812f; float q=.7071067812f; will just leads after applying all the products and squaring to p and q having a common factor 0.5 which can be extracted from the formulaes. So I propose to replace: float p=.7071067812f; float q=.7071067812f; float w=vorbis_coslook(wdel*k); float *ftmp=lsp; int

Jack, I got the asm we were working last night committed ont he beta 3 branch. It's in use in both vorbisfile:ov_read() and the lsp lookup (when using float lookups) I'd like to have a patch of the other things you did in the past day or so, even if you don't want to commit yet. Monty --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage:

So, it turns out (and another implementation actually explicitly mentions it) that LSP->LPC computation using the FIR algorithm is very sensitive to noise (iterative algorithm) and really really requires doubles [we're not kidding]. This was complicating things for folks pursuing fixed point implementations, and also was a potential source for bugs if FP optimizations got out of hand. This

...is') go hand in hand with vorbis's postbeta2 branch. I think monty expects to merge postbeta2 in with mainline shortly. Other notes: This is a pretty big change. I've undoubtedly missed something. Please test and let me know if there's anything dumb. Monty's recent work on lsp_to_curve is commited as well, and according to monty provides a 2x increase in decode speed. Not bad Monty :) I'll commit Kim's vorbisfile documentation pretty soon. And the ogg documentation is well on it's way to being done. jack. --- >8 ---- List archives: http://www.xiph.org/archive...