June 2007 - MPlayer-dev-eng - lists.mplayerhq.hu

Re: [MPlayer-dev-eng] [PATCH] Make mp3lib SIMD optimizations work on AMD64, the Finale
by Zuxy Meng 07 Jun '07

07 Jun '07

Hi, 2007/5/20, Zuxy Meng <zuxy.meng(a)gmail.com>: > As discussed with Guillaume on IRC, I'll split my previous big patch > (Rewrite synth_1to1_MMX....) into several small parts for easier > review. Here's the first one, rewriting the generic code in > synth_1to1_MMX from assembly to C, so we don't need to deal with > different ABIs. I've tested it and confirmed it doesn't hurt > performance. > > Note I removed a conditional jump in the remaining assembly too. By > analyzing the code I'm sure it's never taken so don't worry about > that. Strictly speaking it should be in a seperate patch but then this > patch would break mplayer... > > Part 2 will replace 32-bit leal to equivalent add/sub (without the 'l' > suffix) so pointer arithmetic will be 64-bit under amd64. > > Part 3 will remove hardcoded registers. > > Part 4 will kill tabinit_mmx.c. We don't need to compute the table at > runtime; it can be predetermined. > > Part 5 will correct data types, replacing 'long' with 'int' where necessary. > > The last patch will deal with Makefile and macros. The attached patch modifies macros and the Makefile, effectively turning everything on for AMD64. The result is 47% faster decoding on a K8. Please note that due to mp3lib's floating nature, the output won't be bit-exact w/ and w/o the patch. But this shouldn't hurt the sound quality. -- Zuxy Beauty is truth, While truth is beauty. PGP KeyID: E8555ED6

6 14

{PATCH] - improve asf demuxer's handling of streaming data
by John Donaghy 07 Jun '07

07 Jun '07

Hi The attached patch enables mplayer to properly play asf steaming video such as the one found here: http://a1729.l2168647534.c21686.g.lm.akamaistream.net/D/1729/21686/v0001/re… Without the patch there is a loss of precision when the timestamp values are large. During tests I discovered that the resultant timestamps were the same for several frames in a row, then there was a jump with several more being the same (higher) value. This causes a jerky picture. I'm guessing that the timestamps wrap back to zero at some point. Also, I assume that when the timestamp values are relatively low, the stream will play ok without the patch, but I havent confirmed this. John

2 1

Deprecates external vidix support
by Benjamin Zores 07 Jun '07

07 Jun '07

Now that internal vidix support is up to date, we may consider dropping external lib support for easier configure and potential libvo/*vidix* code changes. Attached patch to configure is supposed to do so. Please comment. Ben

3 3

japanese ISDB sample stream
by Marcus Metzler 06 Jun '07

06 Jun '07

Hi, does anyone have a sample TS from japanese digital TV (ISDB). We need to check if it works with a new HDTV decoder hardware that we are writing linux drivers for. Marcus -- /--------------------------------------------------------------------\ | Dr. Marcus O.C. Metzler | | | mocm(a)metzlerbros.de | http://www.metzlerbros.de/ | \--------------------------------------------------------------------/ |>>> Quis custodiet ipsos custodes <<<|

5 6

[PATCH] change demux_mov.c to use stream_skip instead of stream_seek
by elupus 06 Jun '07

06 Jun '07

Hi, This patch changes demux_mov.c to use stream_skip's instead of stream_seeks to move about in stream, this makes it abit more robust on http streams where server doesn't support seeks. Using stream_seek instead could make demuxer crash when cache was low, as that would then force it to attempt a seek. It also makes sure demuxer checks return value of the seek commands and bails out if it fails. demuxer would end up in a broken loop One thing to note could be that this might make it a tiny bit slower on filesystems where seeks are very inexpensive and the mov files is read in some odd non linear faction, but reading 2*STREAM_BUFFER_SIZE on those filesystems probably won't be much slower. /Joakim PS. This patch relies on the fix for stream_skip for it to work

1 1

[LinuxTag] Thanks
by Måns Rullgård 06 Jun '07

06 Jun '07

Another LinuxTag has passed, and as usual the FFmpeg and MPlayer teams were there. I would like to thank all those who put in time and money for our booth. It was especially nice to see so many of our recent recruits joining the effort (and sharing the fun, of course). A special thanks goes to Marc "Shuttleworth" Hoffman for buying us all dinner. I'm making a note so we won't forget to invite him next year too ;-) Thanks also to all visitors to our booths, particularly those who left something in our donations box. Finally, a somewhat off-topic thanks to Jörg Schilling for the fantastic real-life flame war. To bad we didn't make the "asbestos underwear" T-shirts after all. -- Måns Rullgård mans(a)mansr.com

6 7

[PATCH] mp3lib: Align output so that movaps can be used instead of movups
by Zuxy Meng 06 Jun '07

06 Jun '07

Hi, The attached patch aligns output address of dct64 functions to 16-byte boundary, so movaps can be used instead of movups in dct64_sse, resulting in 2.5% faster decode. I tested against mp3 and mp2 and both worked fine; couldn't find an mp1 sample though. -- Zuxy Beauty is truth, While truth is beauty. PGP KeyID: E8555ED6

8 18

[PATCH] nostatus-line option
by Diego Biurrun 05 Jun '07

05 Jun '07

Attila wanted it, I've been meaning to implement it for ages, here it is: A small option that turns off displaying of the status line. Useful if you wish to have full debug output but your terminal is slow or you wish to redirect output to a file without the status line crap. Comments welcome. If somebody has a better name, that is also welcome. Diego

5 11

[PATCH] use avcodec_decode_audio2
by Reimar Döffinger 05 Jun '07

05 Jun '07

Hello, does attached patch look okay? Greetings, Reimar Döffinger

3 7

[PATCH] do not center GUI message boxes since that crashes with nontrusted X connection
by Reimar Doeffinger 05 Jun '07

05 Jun '07

Hello, attached patch does this. Unfortunately I have not found a way to detect if we are a trusted X client or not, which would allow for a more proper solution :-(. For me working with untrusted connections is more relevant than the position of the message boxes, but please speak up if you find this important enough to keep. Greetings, Reimar Doeffinger

1 0