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

Fix for drawtext under FreeBSD
by eduardo＠bragatto.com 09 Jan '08

09 Jan '08

Hi there -- I'm not a developer but I just came across an problem using drawtext vhook under FreeBSD that I managed to fix. Basically drawtext was complaining that I was not passing the parameters -t and -f, although I was. The vhook module has the following lines to parse it's command line parameters: optind = 0; while ((c = getopt(argc, argv, "f:t:T:x:y:s:c:C:bo")) > 0) { On FreeBSD, getopt() will return -1 when the first non-argument is found. So, if optind is set to ZERO, the first string it will check will be the module's filename, which of course is not considered an argument (as it doesn't have "-" on the beginning). So, regardless of the parameters I was passing, it was always failing. See getopt() manpage for FreeBSD[1]. On Linux, getopt() permutes the contents of argv during scan to make sure all non-argument strings are at the end, so having optind set to ZERO doesn't really make any difference as the module's filename would be shifted to the end and the parameters would be properly parsed. See getopt() manpage for Linux[2]. The solution is simply change: optind = 1; That way, both Linux and FreeBSD will skip the vhook's filename and will check only the parameters as we want. By the way, the vhook watermark.c uses this same approach I'm suggesting here (setting optind=1). I have tested the module with this little modification both on Linux and FreeBSD and it now works. Best regards, Eduardo M. Bragatto.

4 11

[PATCH] MLP/TrueHD decoder
by ian.caulfield＠gmail.com 08 Jan '08

08 Jan '08

Hi all, I've been working for a while on reverse engineering the MLP audio stream format, and writing my own decoder. I've finally got to a point where I have something ready for review. I'm going to try to get the documentation I've produced on the format up on the multimediawiki as well. I've attached the code as two patches - the first adds a parser, the second the actual decoder. Current caveats: - The decoder is limited by a #define to a maximum of six channels. I imagine that 7.1 streams will be more-or-less a case of changing the #defines, but I'd like to see a sample stream first to check. - Surround files are output with the channels in codec order (L, R, C, LFE, Ls, Rs for 5.1). There was some discussion on the list earlier on adding this kind of code to libavcodec instead of individual codecs, so I've left out reordering code for now. - Funky MLP stuff, like having some channels at 96KHz and some at 48KHz isn't supported - Blu-Ray TrueHD streams aren't supported. Apparently these consist of an AC3 base layer, plus a TrueHD layer with residual data or somesuch in. Anyone have any samples? Extracting a two-channel downmix from a 5.1 stream is supported using avctx->request_channels Comments welcome :) Ian

7 65

[PATCH][RFC] TechnoTrend PVA Demuxer
by ivop＠euronet.nl 06 Jan '08

06 Jan '08

Hi, In order to make it possible to deprecate MPlayer's libmpdemux in the future, I started working on a few demuxers that lavf is still missing. This is the first one. http://wiki.multimedia.cx/index.php?title=PVA http://samples.mplayerhq.hu/PVA/ All samples play correctly. I did not however implement seeking yet. That's the RFC. How should I proceed? There's no index. PVA files are encapsulated MPEG-ES (for video) and MPEG-PES (for audio) packets. Should I just rely on the caller (MPlayer, ...) feeding it a new byte-position and resync on the AV-syncword at the start of read_packet() ? And if I do so, can I use url_fseek or does that ruin the ability to pipe the input to ffmpeg? --Ivo

11 35

[RFC] RV30 1/3pel MC functions
by kostya.shishkov＠gmail.com 04 Jan '08

04 Jan '08

Here is my attempt to use RV30 1/3pel MC in the same way as standard H.264 ones.

3 11

[PATCH] PC Paintbrush PCX decoder
by ivop＠euronet.nl 04 Jan '08

04 Jan '08

Hi, Here's a decoder for PC Paintbrush PCX files. http://wiki.multimedia.cx/index.php?title=PCX Samples are uploaded to: http://samples.mplayerhq.hu/MPlayer/incoming/pcx Merry christmas, --Ivo

6 20

16:9 DV AVI not detected
by j＠v2v.cc 03 Jan '08

03 Jan '08

16:9 ntsc avi dv files exported with Adobe Premiere, the aspect ratio is not properly detected. it is always set to 1:1. what would be the right place to fix this? the only place i could find that sets sample_aspect_ratio is dv_extract_video_info in libavformat/dv.c dont see this used for avi files... j

4 10

ffmpeg SVN crashes when the input file doesn't exist
by ismail＠pardus.org.tr 02 Jan '08

02 Jan '08

Hi all, SVN revision 11298, [/packages/ffmpeg]> ./ffmpeg_g -i /dev/null FFmpeg version SVN-r11298, Copyright (c) 2000-2007 Fabrice Bellard, et al. configuration: --enable-gpl --enable-libx264 --enable-libvorbis --enable-libfaad --enable-libfaac --disable-ffserver --enable-liba52 --disable-static --enable-shared --enable-pp --enable-libxvid --enable-libmp3lame --disable-debug --prefix=/usr --enable-pthreads --enable-libamr-nb --enable-libamr-wb --cpu=i686 --enable-libtheora --enable-libgsm --extra-cflags=-fno-strict-aliasing --enable-x11grab libavutil version: 49.6.0 libavcodec version: 51.49.0 libavformat version: 52.3.0 built on Dec 22 2007 01:16:51, gcc: 4.3.0 20071221 [trunk revision 131119] zsh: segmentation fault ./ffmpeg_g -i /dev/null Valgrind reports invalid reads : ==5603== Invalid write of size 4 ==5603== at 0x8052579: opt_input_file (ffmpeg.c:2666) ==5603== Address 0x5294e7c is 0 bytes after a block of size 3,900 alloc'd ==5603== at 0x4024A24: memalign (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==5603== by 0x38002FAF: (within /usr/lib/valgrind/x86-linux/memcheck) ==5603== ==5603== Invalid write of size 4 ==5603== at 0x80525B1: opt_input_file (ffmpeg.c:2667) ==5603== Address 0x5294e80 is 4 bytes after a block of size 3,900 alloc'd ==5603== at 0x4024A24: memalign (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==5603== by 0x38002FAF: (within /usr/lib/valgrind/x86-linux/memcheck) ==5603== ==5603== Invalid write of size 4 ==5603== at 0x80525E9: opt_input_file (ffmpeg.c:2668) ==5603== Address 0x5294e84 is 8 bytes after a block of size 3,900 alloc'd ==5603== at 0x4024A24: memalign (in /usr/lib/valgrind/x86-linux/vgpreload_memcheck.so) ==5603== by 0x38002FAF: (within /usr/lib/valgrind/x86-linux/memcheck) Regards, ismail -- Never learn by your mistakes, if you do you may never dare to try again.

2 2

[PATCH] Roundup issue #301
by christophe.gisquet＠free.fr 02 Jan '08

02 Jan '08

Hi, This addresses roundup issue #301 : https://roundup.mplayerhq.hu/roundup/ffmpeg/issue301 Summary of changes: - Use of MANGLE when loading into mmx registers some constants - Conversion to non-static for those constants and thus renaming with ff_prefix - Removal of last parameter of MSPEL_FILTER13_CORE (was constant) - Use of "+r" instead of stricter but unnecessary "+g" - Use of REG_c and direct loading of some of the above (not tested No speed regression is noticeable with the new code, but the performance vastly varies depending on the compiler. For a 720p test sequence: 2.95 ~ 4.1s 3.3.6 ~ 3.8s (3.2 in 64b) 4.2.1 ~ 3.4s (3.1 in 64b) Considering those numbers, I think I'll pass next time. (And yeah! for hard-coded limitations) Best regards, -- Christophe GISQUET

3 3

[PATCH] rmdec: remove double-nested if
by rsbultje＠gmail.com 02 Jan '08

02 Jan '08

Hi, Patch for $subj attached. It removes a double-nested if that looks kind of stupid. It does nothing functional. I had it in my tree for some reason, so it must've occurred to me at some point that it looks stupid. Ronald

2 2

[PATCH] socklen_t definition
by daveryeo＠telus.net 01 Jan '08

01 Jan '08

With more talk about libossupport I thought it might be a good time to resurrect KO Myung-Hun's socklen_t patch. Still typedefs it to int as that is what OS/2 and Beos use. Not sure about the windows bit that KO added. Dave

5 9