December 2007 - ffmpeg-devel - lists.ffmpeg.org

Fwd: [PATCH] add rtsp/realmedia (RDT) protocol declaration
by rsbultje＠gmail.com 27 Dec '07

27 Dec '07

Hi, On 11/4/07, Luca Barbato <lu_zero at gentoo.org> wrote: > > Ronald S. Bultje wrote: > > Luca, > > > > can you please review these patches for inclusion? I'd like to get > > Sure, just let me read them, sadly if you don't attach them as text I > need to be prodded to read them =/ I'll move to another mail client once > I find one working as should. > > > integration of my RDT (RTSP / realmedia) patchset into ffmpeg going. I > have > > quite a couple after this one going, touching rtsp/rtp*.c and rmdec.c, > so > > I'll need various maintainers' approval for this stuff. > > > > rtsp-protocol-comment and rtsp-protocol-declaration are obviously fine > the reindent seems to have a stray "`" Whoops, how did that get in there? Here's a fixed version (I probably forgot to refresh the patch after getting it to compile w/o the "`"). Thanks for the review! More coming up (I hope). Ronald

1 1

cuda
by nenashev_as＠mail.ru 27 Dec '07

27 Dec '07

Hi everyone! I'm just a new member in this mailing list thats why my question wouldn't be original. Did anyone try to add CUDA acceleration for dsputil acceleration, DCT and IDCT for example? CUDA is a new Nvidia hardware platform with additional development tools. Thanks for your answers.

2 1

[PATCH] fix clicking sound in xa-adpcm
by astrange＠ithinksw.com 27 Dec '07

27 Dec '07

The XA ADPCM decoder resets the last sample predictors to 0 after every frame. This is wrong, and causes loud clicking constantly in all my samples. Attached patch fixes it. I can't tell if seeking is affected, since the STR demuxer can't seek in ffplay (not sure why) and can't be remuxed (no codec_tag).

4 5

[PATCH] djgpp port
by mik＠niipt.ru 27 Dec '07

27 Dec '07

Known problems: ffmpeg.exe produces broken files, e.g. riff header is absent, huge amount of data is zero-filled etc. (other binaries such as mencoder and ffmpeg2theora work fine); the test suite looks not working because of bash.exe limitations; wma decoding causes sigill when sse is used, gcc version doesn't matter; at least inet_aton() is not a part of standard library, configure checks should use extra settings. Michael

4 11

ffmpeg misdetects frame rate on divx files
by gforte＠wyplay.com 27 Dec '07

27 Dec '07

Hi, I have uploaded a divx 4 stream at ftp://samples.mplayerhq.hu/MPlayer/incoming/mpeg4_framerate of which the frame rate is detected anormally by ffmpeg: [NULL @ 0xb7db1450]looks like this file was encoded with (divx4/(old)xvid/opendivx) -> forcing low_delay flag Input #0, avi, from 'fightclub.avi': Duration: 00:01:40.0, start: 0.000000, bitrate: 757 kb/s Stream #0.0: Video: mpeg4, yuv420p, 640x288 [PAR 0:1 DAR 0:1], 1.00 tb(r) Stream #0.1: Audio: mp3, 48000 Hz, stereo, 32 kb/s playing it with ffplay renders video very slowly. this wrong 1.00 fps comes from libavcodec/h263.c's decode_vol_header() function, where AVCodecContext's time_base field is overwritten with seemingly bogus values when fixed_vop_header == 0. in this case, the stream contains time_increment_resolution == 1. I have another stream where time_increment_resolution == 2, leading to 2.00 fps. I don't know which of the stream or decoder is broken, but attached is a patch that appears to circumvent the problem. as I don't feel familiar enough with the mpeg4p2 standard, I don't know exactly if it is appropriate to do this in such a situation... I'm not even sure if it wouldn't break in other cases (most particularly, maybe it could lead to division by zero if s->avctx->time_base isn't properly initialized ?). maybe it is a divx4 encoder bug, as the only such samples I have are divx4... any comment/information/correction would be appreciated if it is not the proper way to fix this. cheers, -- Gabriel Fort? <gforte at wyplay.com>

2 1

[PATCH] djgpp port
by mik＠niipt.ru 26 Dec '07

26 Dec '07

At Wed, Dec 26 2007 04:38:29 CET +0100, Michael Niedermayer wrote: >> +#ifndef HAVE_ROUND >> +static av_always_inline double round(double x) >> +{ >> + return (x > 0) ? floor(x + 0.5) : ceil(x - 0.5); >> +} >> +#endif /* HAVE_ROUND */ > iam ok with this, though if i were picky i could complain that > your round implementation fails for 0.49999999999999994449 amongth others > but i guess its good enough. Fixing that would make it slower and > uglier with no real advantage ... IIRC the glibc round has the same behavior inside the __DBL_EPSILON__ neighborhood of x.5. In addition you could complain that my fake round uses uncommon rounding mode. Michael

2 2

[RFC] check avcodec_open codec paramert != NULL
by Reimar.Doeffinger＠stud.uni-karlsruhe.de 25 Dec '07

25 Dec '07

Hello, I personally would prefer if the avcodec_open would check if codec is NULL and return an error then. The main reason is laziness: You can then just do avcodec_open(vf->priv->avctx, avcodec_find_encoder(CODEC_ID_PNG)) instead of having an extra variable for the avcodec_find_encoder result and checking it against NULL. What do you think? I admit it encourages sloppy error checking but I just hate doing more checks that absolutely necessary, it clutters the code too much for my tastes. Greetings, Reimar D?ffinger

2 1

Audio Video Merge not in Sync - possible bug?
by spam＠bentheten.com 23 Dec '07

23 Dec '07

Hi I'm trying to convert my avchd files to something readable, but I'm always stuck in having two files or Audio not in sync. Example: 00009.mts ------------- Splitting the file xporthdmv -h 00009.mts 1 1 1 convert video file to yuv ldecod -i bits0001.mpv -o 00009.mts.yuv rename audio file to ac3 mv bits0001.mpa 00009.mts.ac3 merging the two files in an avi file ffmpeg -r pal -s 1440x1080 -i 00009.mts.yuv -i 00009.mts.ac3 -acodec copy -vcodec libx264 -sameq -aspect 16:9 -b 15000k -deinterlace 00009.mts.avi ------------- now when I start "mplayer 00009.mts.avi" then the audio is out of sync BUT when I start "mplayer 00009.mts.avi -audiofile 00009.mts.ac3" the audio is perfectly in sync. I tried converting only the audio, only the video, both, none .. but no matter what I do, if I merge them with ffmpeg or mencoder(svn) which should be more or less the same, the audio moves out of sync(runs slower than the video). The files: original: http://www.bentheten.com/cx7/00009.mts conv. a/v: http://www.bentheten.com/cx7/00009.mts.avi audio only: http://www.bentheten.com/cx7/00009.mts.ac3 I hope this belongs to the developer list. Thanks Ben

1 0

New project using ffmpeg
by diego＠biurrun.de 23 Dec '07

23 Dec '07

On Wed, Dec 12, 2007 at 01:34:25PM +0000, Cezary Falba wrote: > > On 18 Sep 2007, at 15:01, Diego Biurrun wrote: > >> On Tue, Sep 18, 2007 at 02:49:56PM +0100, Cezary Falba wrote: >>> >>> On 18 Sep 2007, at 14:43, Diego Biurrun wrote: >>> >>>> On Tue, Sep 18, 2007 at 12:05:14PM +0100, Cezary Falba wrote: >>>>> >>>>> On 18 Sep 2007, at 12:03, Diego Biurrun wrote: >>>>> >>>>>> On Fri, Sep 14, 2007 at 05:27:28PM +0100, Cezary Falba wrote: >>>>>>> Hi, could you please note that we've started new project using >>>>>>> ffmpeg? It's homepage is s2.wap.filmnight.com/trailers/software.html >>>>>>> but it will soon change to saffrondigital.com. >>>>>>> It's a free download application that prepares video source from our >>>>>>> customers to be used by our system internally. >>>>>> >>>>>> Let us know when you have found a permanent URL for it. >>>>>> >>>>>> Also, where is the source? Where is the credit? >>>>> >>>>> The source code is available on request to everyone who needs it. >>>>> Credits are displayed during instalation process... >>>> >>>> I just downloaded and installed the Mac OS X version of Saffron Master >>>> Encoder. No credits were displayed anywhere. Nowhere is the word >>>> FFmpeg or x264 mentioned. There is no copy of the GPL. You do not >>>> provide source code nor describe where to get it. >>>> >>>> In other words, you are violating the license in every way imaginable. >>>> >>>> Why? >>> >>> Please forgive me that - it's my first software release. I'll add copy of >>> GPL as well as proper information in Readme File that is going to be >>> displayed during installation. I will let you know when it's done (surely >>> by the end of this week). >> >> Let us know when you have done this. Also, if you haven't already, you >> should really read the GPL. > > I'm sorry for a long delay with solving those issues. Could you please > revise that software's site and give me your comments? First off, you mention FFmpeg (misspelled on your site), AviSynth and Perl, but you don't link to them. This is the internet, place a link! > I'm just not sure if there is enough information about the way we use > ffmpeg. The information you provide is OK I would say, but please do not misspell FFmpeg (notice the two capital F) and use ffmpeg.org as the homepage link. > I'd like to be 100% sure everything is clear and ok with ffmpeg team before > we include that page on company's website and allow other people to use it. I just installed the Mac version of Saffron Encoder again but you made no changes at all! Everything I criticized above is still valid and none of the things you promise above have been implemented. What's it gonna be? Diego

3 7

how to insert the 'vsync' into this example?
by redsuyu＠163.com 23 Dec '07

23 Dec '07

Hi, 1 how to insert the 'vsync' into this example? 2 how the parameter 'vsync' works? 1 how to insert the 'vsync' into this example? ./ffmpeg -f x11grab -s cif -i :0.0 -r 12 -vcodec mpeg4 -f mpegts -an udp:192.168.119.1:1234 2 how the parameter 'vsync' works? i see this functions parse_options(argc, argv, options, opt_output_file); but how does it know the vsync? i can't find the trace.. thank you ! `-vsync parameter' Video sync method. Video will be stretched/squeezed to match the timestamps, it is done by duplicating and dropping frames. With -map you can select from which stream the timestamps should be taken. You can leave either video or audio unchanged and sync the remaining stream(s) to the unchanged one. 2007-12-23 redsuyu

1 0