>>>>> "M" == Mark Taylor <[EMAIL PROTECTED]> writes:
M> I applied the Huffman patch, it will be in 3.51.
M> The MDCT + FFT patch is more difficult. They are noticably
M> faster for me too, and I would like to get them into lame, but:
M> The before and after patch results are not identical. Maybe
M> just rouding errors, but it also looks like there are some
M> errors syncing the psycho-acoustics and the MDCT encoding? At
M> least the frame analyzer shows them out of sync by about 1
M> granule.
Oh, it is my patch merging mistake.
It's not "about 1 granule" but "exactly 1 granule"
M> I tried to put in just the MDCT patches, but also couldn't
M> reproduce the original output. Takehiro: is the delay of your
M> MDCT different than the original? The current MDCT routine has
M> a delay of 240 (filterbanks) + 288 (MDCT's).
my MDCT routine is removed buffering and the delay decrease 480.
so the delay is 240 + 288 - 480.
BTW, I have also MDCT+huffman patch. The patched code produces "little
bit different" MP3 file because of rounding error, but I hope no sync bug
in it.
I will sent this patch if you want it, after I come home.
---
[EMAIL PROTECTED] // may be source be with you!
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
