November 2008 - ffmpeg-devel - lists.ffmpeg.org

[PATCH] Ignore file size field in BMP header
by daniel＠drv.nu 16 Feb '09

16 Feb '09

This patch removes the reading and checking of the file size field in the BMP decoder (known as bfSize in the MS docs). This field is documented (see MSDN page for BITMAPFILEHEADER) as containing the size of the file, but other sources document it as containing the size of just the file header or the size of all the headers. This is seen in some files in the wild; for example: http://samples.mplayerhq.hu/bmp-files/rmoney.bmp has 14 in this field (size of file header) http://samples.mplayerhq.hu/bmp-files/test4os2v2.bmp has 78 (size of all headers) This field isn't necessary to decode the bitmap correctly, so the patch removes the read and checks (the second check is changed to use the buffer size instead of the file size from the header). The patch fixes loading of rmoney.bmp; test4os2v2.bmp is OS/2 V2 (different header size), so does not yet decode. -- Daniel Verkamp

2 3

[PATCH] support FCP IMX/D-10 30/40/50 mov muxing
by baptiste.coudurier＠smartjog.com 12 Feb '09

12 Feb '09

Hi, This patch will correctly mux IMX/D-10 essences into a FCP compatible mov. This patch is not really meant to be included in svn (if people are against it), but since too many people are contacting me to get it, Im publishing it here. Feedback very welcome especially when muxing NTSC material. -- Baptiste COUDURIER GnuPG Key Id: 0x5C1ABAAA SMARTJOG S.A. http://www.smartjog.com Key fingerprint 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA Phone: +33 1 49966312

4 3

[PATCH] Add versioning information to dlls
by jkolb＠wsi.com 12 Feb '09

12 Feb '09

This patch adds support for generating .rc files and then linking them with the dlls created by ffmpeg. This information can be used to identify the product version by the user and is also available to installer programs for ease of updating etc. Jeremy

10 102

[PATCH] H.264: decode picture timing SEI message and get field order
by h.yamagata＠nifty.com 10 Feb '09

10 Feb '09

Hi, I wrote a patch which implements decoding of picture timing SEI. [Current problems] libavcodec derives field order from POCs. cur->top_field_first = cur->field_poc[0] < cur->field_poc[1]; It assumes BFF if two POSs are equal. This is not always correct. And it derives AVFrame::interlaced_frame from used decoding process. cur->interlaced_frame = FIELD_OR_MBAFF_PICTURE; Sometimes this is not correct. Interlaced/progressive mode of post-decoding processes such as deinterlacing and color space conversion is not always same as the mode used during decoding. H.264 for example allows to encode an interlaced frame using a progressive algorithm. Sample: premiere-paff.ts (http://x264.nl/h.264.samples/premiere-paff.ts) [Solutions] Picture timing SEI message contains pic_struct which defines interlace/progressive mode of post decoding processes and field order. Most streams have the SEI for every frame or field. My patch implements decoding of the SEI and use of pic_struct. Now the field order is always correct. Interlaced picture is judged interlaced. Field 1 repeat flags are also set. [Limitation] As usual, there are exceptions: badly encoded streams. Some progressive streams from television have pic_struct 3, which indicates interlacing. Fortunately such streams are rare. This is not my fault, other decoder show the same behavior. Please review. ----------------------------- for svn log ----------------------- H.264: Implement decoding of picture timing SEI message Now correct values are set to interlaced_frame, top_field_first and repeat_pict in AVFrame structure. patch by ffdshow tryouts -------------------------------------------------------------- Best regards, Haruhiko Yamagata

3 9

[Patch] Topfield 58XXPVR and 68XXPVR file support
by doctorjames＠gmail.com 10 Feb '09

10 Feb '09

Topfield make a variety of personal video recorders that allow for the transfer of recorded programmes via an onboard USB port. The ".rec" files they produce are MPEG2 transport streams with a 3760 byte header containing archive information for the PVR. The attached patch checks for the presence of "TFrc" in the first four bytes of a file and identifies it as an mpeg transport stream if present. It subsequently skips this header when the services are probed. With the patch in place these files are then fully supported by ffmpeg. Cheers, James

3 8

[PATCH] mpegts.c: fix minor memory leak
by makovick＠gmail.com 06 Feb '09

06 Feb '09

Hi, there is a small memory leak in the MPEG-TS stream detection code: When auto_guess is used in the MPEG-TS demuxer, PESContext structures which do not really belong to PES streams are not freed at the end, because they do not get referenced by the AVStreams. The attached patch fixes the leak by freeing all PESContexts without a corresponding stream. -- Jindrich Makovicka

3 4

[PATCH] change the order of params for av_init_random()
by justinruggles＠bellsouth.net 05 Feb '09

05 Feb '09

$subject includes lavu minor version bump. -Justin

6 26

[PATCH] Proper parsing of DTS-HD MA streams - Patch to TS Parser
by robert.mcnamara＠gmail.com 03 Feb '09

03 Feb '09

Updating this subject to correctly reflect the contents. I am attaching a patch to handle the first of two hurdles for solving the problem of proper parsing of DTS HD MA audio. This is a rework of a previous patch discussed in June. I'll try to briefly summarize the issues at that time. At the time, there was no means of handling TrueHD, and there was some question as to whether the E-AC3 and AC3 stream types should be handles seperately. As both TrueHD/MLP and E-AC3 are handled properly by FFMpeg now, it seems appropriate to bring this back to the table. The stream type definitions are are outlined here: http://www.freepatentsonline.com/y2007/0116302.html I have added a case for the TrueHD streams to be handled by CODEC_ID_MLP. As we know that E-AC3 is now handled by CODEC_ID_AC3, I have added that to the EAC3 stream handling to that case. Finally, the issues with properly decoding DTS-HD are parsing ones for dca_parser. It is useful in working towards that goal to be able to see the streams properly in ffmpeg, so I have added the DTS-HD definitions to the DTS parser case. I have tested all stream types with real world examples from my own collection and from samples provided by madshi. Between that and the patent application delineating the stream types, I am hoping that this will be sufficient to get these changes committed in furtherance of working DTS-HD MA playback. I am by no means a programmer so I beg everyone's pardon for any stupid mistakes, and beleive me when I say you couldn't make me any more embarrassed about my shortcoming in that department than I already am. :) To anyone celebrating the holiday, a happy thanksgiving. Thanks, Robert

5 10

[PATCH] DV timecode
by baptiste.coudurier＠smartjog.com 02 Feb '09

02 Feb '09

Hi 3 patches: - add timecode.h to avoid code duplication for drop frame timecode adjustment - use new lavc timecode api to set timecode in DV. Regression tests change because now default behaviour is non drop frame timecode, you can still activate it by using -flags2 +drop_frame_timecode - timecode helper function for ffmpeg to set timecode. Maybe moving brktimegm to lavu and using it is better ? -- Baptiste COUDURIER GnuPG Key Id: 0x5C1ABAAA SMARTJOG S.A. http://www.smartjog.com Key fingerprint 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA Phone: +33 1 49966312

5 14

[Patch] Enable gpl if enabled nonfree
by cehoyos＠rainbow.studorg.tuwien.ac.at 19 Jan '09

19 Jan '09

Hi! IMO, it does not make much sense to be required to enable nonfree and gpl at the same time if one wants to use a nonfree and a gpl'ed library. Patch attached. Carl Eugen

9 17