On Tue, Feb 28, 2012 at 05:31:27AM -0800, Mashiat Sarker Shakkhar wrote: > > > ----- Original Message ----- > > From: Christophe Gisquet <[email protected]> > > To: libav development <[email protected]> > > Cc: > > Sent: Tuesday, February 28, 2012 7:09 PM > > Subject: Re: [libav-devel] [PATCH] Windows Media Audio Lossless decoder > > > > 2012/2/28 Mashiat Sarker Shakkhar <[email protected]>: > >> https://github.com/shahriman/libav/tree/wmall > > > > I was in the process of profiling it. A very quick analysis (I'm > > sparing you the details) show that > > first code blocks of lms_update and lms_predict account for more than > > 60% of the decoding time, while the rest is scattered without clear > > candidates. > > > The code is not really as optimized as it could be, because optimizing > it wasn't really the first priority and I am not terribly experienced > at that. > > > > > Most operations seems to involve int16_t, so dspizing should not be > > very difficult (famous last words). > > Intermediate operations require 32-bit integer according to the spec, > (that is, Pred is int32_t). Ref-code uses 64-bit for some reason that > I don't understand.
it's hard to understand reference decoder indeed > I don't have experience with SIMD, but will try to write some if that's > necessary. Priority for now is to implement 24-bit. Yes, you don't have to worry about SIMD. _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
