December 2009 - ffmpeg-devel - lists.ffmpeg.org

[bug/patch] MPEG-TS muxer: PCR not in sync with PTS/DTS
by Alan.Stratford＠Goodrich.com 17 Feb '10

17 Feb '10

>On 09/04/2009 07:28 AM, Niobos wrote: >>>> * Changed the calculation of total_bit_rate to be more accurate >> This still had a bug in it (bits vs bytes thing), which is corrected in >> the attached version >> >> >> On 03 Sep 2009, at 12:13, Alexandre Ferrieux wrote: >>> Just done preliminary tests with your patches and a set-top-box: >>> >>> - the VBR works but as bitrate increases, the box gets troubled by >>> bursts (loss of picture updates) >>> - the CBR works beautifully at all bitrates >>> >>> Thanks a lot for this work ! >> >> I run my tests in VLC and on iPhone, both work without a glitch; >> bitrates range from 200kbps to 1.5Mbps. > > >Thanks a lot for the work, I'll review ASAP. > Hi, Many thanks for this latest patch. This seems to help make TS files created by ffmpeg more compatible with VLC. I have been looking at the calculation of total_bit_rate in the latest update to this patch. I think that I understand the calculation but I am confused by the line total_bit_rate += 4 * total_bit_rate / (TS_PACKET_SIZE-4) + /* TS header size */ 1000 * 8 * sdt_size / SDT_RETRANS_TIME + /* SDT size */ 1000 * 8 * pat_pmt_size / PAT_RETRANS_TIME + /* PAT+PMT size */ 1000 * 8 * 8 / PCR_RETRANS_TIME; /* PCR size */ It is the following part that is causing confusion: 4 * total_bit_rate / (TS_PACKET_SIZE-4) + /* TS header size */ I believe that this is trying to adjust the total_bit_rate to allow for the fact that the stream will be sent as TS packets, and, therefore, will require an extra 4 bytes of data for each 184 bytes in the stream. If this is not the case then please ignore the rest of this post, but I would greatly appreciate an explanation of what the line does. If my understanding is correct this part should be: total_bit_rate * TS_PACKET_SIZE / (TS_PACKET_SIZE-4) + /*TS header size*/ i.e. multiply the stream size by 188/184 to account for 4 extra bytes every TS packet. Alan -------------------------------------------------------------------------- Rosemount Aerospace Limited. Registered Office: Stratford Road, Solihull, West Midlands, B90 4LA. Registered in England No. 02849033. --------------------------------------------------------------------------

11 45

[PATCH] enc is not a good name for a decoder context
by stefano.sabatini-lala＠poste.it 16 Feb '10

16 Feb '10

Hi, as in subject. Regards. -- FFmpeg = Friendly and Foolish Multimedia Portable Ecstatic Geek

3 10

[PATCH] Implement -pix_fmts for listing all the supported pixel formats
by stefano.sabatini-lala＠poste.it 16 Feb '10

16 Feb '10

Hi, requires two steps, moving list_fmts to cmdutils.c and implement the function. My plan is to remove the -pix_fmt list feature, which I never liked and which would be at this point just redundant. Same could be done for -sample_fmts. At this point I believe there would be no point into leaving its list_fmts() interface in cmdutils.h, so I my plan to remove it (this is why I didn't docuemnted it). Regards. -- FFmpeg = Fundamental & Faboulous Multimedia Peaceful Elected Game

2 7

cavs.c DECLARE_ALIGNED in ff_cavs_filter
by Reimar.Doeffinger＠gmx.de 15 Feb '10

15 Feb '10

Hello, currently there is this code in cavs.c: DECLARE_ALIGNED_8(uint8_t, bs[8]); The reason for it is that kind of code: > *((uint64_t *)bs) = 0x0202020202020202ULL; > if( *((uint64_t *)bs) ) { I thought of "solving" this via a union, but is this code really this speed critical that we need such a mess? Can't we just use memset for the writes and e.g. AV_RN64 for the read?

4 6

[PATCH] Document url_fopen()
by stefano.sabatini-lala＠poste.it 09 Feb '10

09 Feb '10

Hi, as in subject. -- FFmpeg = Frenzy Fundamentalist Mournful Peaceless Elitist Guru

1 4

[PATCH] Document url_fdopen()
by stefano.sabatini-lala＠poste.it 05 Feb '10

05 Feb '10

As in subject. -- FFmpeg = Fast and Fundamentalist Minimalistic Philosophical Elastic Ghost

3 12

[PATCH] make av_get_bit_rate part of public API
by krueger＠signal7.de 05 Feb '10

05 Feb '10

Hi, this patch adds a prototype for av_get_bit_rate (implemented in libavcodec/utils.c) to avcodec.h. Regards, Robert

3 6

[PATCH] Use pkg-config for x264
by felipe.contreras＠gmail.com 03 Feb '10

03 Feb '10

Signed-off-by: Felipe Contreras <felipe.contreras at gmail.com> --- configure | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/configure b/configure index 844133b..31e289a 100755 --- a/configure +++ b/configure @@ -2377,7 +2377,8 @@ enabled libschroedinger && add_cflags $(pkg-config --cflags schroedinger-1.0) && enabled libspeex && require libspeex speex/speex.h speex_decoder_init -lspeex enabled libtheora && require libtheora theora/theora.h theora_info_init -ltheora -logg enabled libvorbis && require libvorbis vorbis/vorbisenc.h vorbis_info_init -lvorbisenc -lvorbis -logg -enabled libx264 && require libx264 x264.h x264_encoder_encode -lx264 -lm && +enabled libx264 && add_cflags $(pkg-config --cflags x264) && + require libx264 x264.h x264_encoder_encode $(pkg-config --libs x264) && { check_cpp_condition x264.h "X264_BUILD >= 78" || die "ERROR: libx264 version must be >= 0.78."; } enabled libxvid && require libxvid xvid.h xvid_global -lxvidcore -- 1.6.5.3

6 14

[PATCH] GSM-MS decoder and encoder
by mbardiaux＠mediaxim.be 01 Feb '10

01 Feb '10

-- Michel Bardiaux R&D Director T +32 [0] 2 790 29 41 F +32 [0] 2 790 29 02 E mailto:mbardiaux at mediaxim.be Mediaxim NV/SA Vorstlaan 191 Boulevard du Souverain Brussel 1160 Bruxelles http://www.mediaxim.com/

6 56

[PATCH] Implement av_get_pix_fmt_by_name()
by stefano.sabatini-lala＠poste.it 30 Jan '10

30 Jan '10

Hi, as in subject. This is going to replace avcodec_get_pix_fmt(). Regards. -- FFmpeg = Frenzy & Fierce Mind-dumbing Patchable Erotic Guide

3 13