January 2006 - ffmpeg-devel - lists.ffmpeg.org

[BUG] "too many threads" with brd_scale >= 5
by bugfood-ml＠fatooh.org 29 Jan '06

29 Jan '06

Hello, I've been testing vb_stategy=2 and brd_scale, and I ran into a bug. Whenever brd_scale is set to 5 or higher, libavcodec reports: [mpeg4 @ 0xb93650]too many threads ...and mencoder segfaults. Here's the command: mencoder small.vob -nosound -ovc lavc -lavcopts vcodec=mpeg4:vpass=1:vmax_b_frames=2:vb_strategy=2:brd_scale=5 \ -ofps 24000/1001 -o test.avi -v I've attached the source file and a verbose log. ----- Versions: MEncoder dev-CVS-060119-10:39-4.0.3 (C) 2000-2006 MPlayer Team gcc version 4.0.3 20060115 (prerelease) (Debian 4.0.2-7) Linux bugfood 2.6.14 #1 Sun Jan 15 16:45:19 PST 2006 x86_64 GNU/Linux (also tested on) Linux fire 2.6.13.2 #1 Fri Sep 30 10:49:45 PDT 2005 i686 GNU/Linux Is that enough information, or am I getting too lazy with my bug reports? :) -Corey

3 5

[Ffmpeg-devel-old] トキメキ学園３年女子の皆さんへ 02
by info＠ffcom.info 29 Jan '06

29 Jan '06

$B%H%-%a%-3X1`#3G/=w;R$N3'$5$s$XO"Mm$G$9!#(B $B at hF|M=Dj$7$F$*$j$^$7$?!#(B $B?HBN8!::$NF|Dx$,JQ99$K$J$j$^$7$?!#(B --------------$B;}$AJ*(B-------------- $B!{1x$l$F$bNI$$%Q%s%D(B $B!{@8M}MQIJ(B $B!{BNA`I~(B ---------------------------------- $B"(>0!"@8M}Cf$NJ}$O;vA0$K?=$7>e$2$F$/$@$5$$!#(B $B>\$7$$!"F|Dx!&>l=j$K$D$$$F$O0J2<$N#U#R#L$h$j3NG'$7$F$/$@$5$$!#(B http://ffcom.info/wty/ $B%H%-%a%-3X1`!!J]7r2J(B $B;38}$h$j(B $B4V0c$($FFO$$$F$7$^$C$?>l9g$O$*<j?t$G$9$,!"(B $B%3%A%i$^$G$4O"Mm$*4j$$?=$7>e$2$^$9!#(B refusal at ok.kz ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642 _______________________________________________ Ffmpeg-devel mailing list Ffmpeg-devel at lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ffmpeg-devel

1 0

Re: ffmpeg-devel Digest, Vol 10, Issue 105
by kostya.forjunk＠gmail.com 28 Jan '06

28 Jan '06

> From: Diego Biurrun <diego at biurrun.de> > > On Thu, Jan 26, 2006 at 09:55:21PM +0100, Michael Niedermayer wrote: > > > > On Thu, Jan 26, 2006 at 08:40:19PM +0100, Roberto Togni wrote: > > > On Sun, 22 Jan 2006 21:30:02 +0200 > > > Kostya <kostya.shishkov at gmail.com> wrote: > > > > > > > Please test this version. It now correctly decodes most of test data > > > > (except 12-bit images and images with subsampling). I also tried to > > > > remove all tabs and whitespace line ends. > > > > Comments are welcomed. > > > > > > Works for me, but the only sample i have worked also without the patch. > > > > > > Any comments about this? > > > > why not apply it? > > Well, Kostya never bothered to send me username or password, so he does > not have an account yet... > > Kostya, willing to become an FFmpeg dev or not? :) > > Diego I want, just cannot right now. My computer is not very suitable for development (too few MBs free). I hope to move on MacMini in a month (it needs a monitor), then I'll start developing (and solving endianness issues).

2 1

BUG REPORT: Garbled audio when AC3 stream changes channel count
by nr＠eecs.harvard.edu 28 Jan '06

28 Jan '06

FFmpeg becomes confused when transcoding certain audio files. The resulting audio files are OK up to a point, after which everything sounds like chipmunks. I suspect that the problem is an A/52 (AC3) audio stream that changes between 5.1 channels and 2 channels in the middle of the stream. Details follow. I am using the CVS version of FFmpeg compiled today, 18 Dec 2005. * What I am trying to accomplish: transcode US ATSC high-definition television broadcast (MPEG-2 TS) to MPEG-2 file suitable for passing to dvdauthor and burning to DVD. * The problem I encountered: at some point partway through a file, audio suddenly sounds like 'the chipmunks': it is very high-pitched, unintelligble, and no longer synchronized with video. Audio typically then finishes well before video. * The exact command line I was using: myffmpeg -i bud-light.ts -target ntsc-dvd -b 4000 -bufsize 400 -ab 384 \ -maxrate 7000 -mbd 2 -trell zzz -acodec ac3 ffm-bud-light.mpg N.B. 'myffmpeg' is my version compiled from CVS (as opposed to the Debian version in /usr/bin/ffmpeg). * The console output provided by FFmpeg: ffmpeg version CVS, build 3342336, Copyright (c) 2000-2004 Fabrice Bellard configuration: --enable-gpl --enable-a52 built on Dec 18 2005 12:48:26, gcc: 4.0.3 20051111 (prerelease) (Debian 4.0.2-4) Input #0, mpegts, from 'bud-light.ts': Duration: 00:00:25.3, start: 17010.607378, bitrate: 17551 kb/s Stream #0.0[0x31], 59.94 fps(r): Video: mpeg2video, yuv420p, 1280x720, 16800 kb/s Stream #0.1[0x34](eng): Audio: ac3, 48000 Hz, 5:1, 384 kb/s Output #0, dvd, to 'ffm-bud-light.mpg': Stream #0.0, 29.97 fps(c): Video: mpeg2video (hq), yuv420p, 720x480, q=2-31, 4000 kb/s Stream #0.1: Audio: ac3, 48000 Hz, 5:1, 384 kb/s Stream mapping: Stream #0.0 -> #0.0 Stream #0.1 -> #0.1 Press [q] to stop encoding [mpeg2video @ 0x833d388]Warning MVs not available [mpeg2video @ 0x833d388]concealing 3320 DC, 3320 AC, 3320 MV errors frame= 763 q=0.0 Lsize= 12572kB time=18.0 bitrate=5737.0kbits/s video:11304kB audio:842kB global headers:0kB muxing overhead 3.515292% N.B. Although the clip runs 25 seconds, FFmpeg reports only 18 seconds. This report is consistent with the output file, which contains 25 seconds of video and 18 seconds of audio. It is not useful to run ffmpeg_g, because ffmpeg completes execution without errors---only the output is bad. Sample media illustrating the problem may be found at http://http://www.eecs.harvard.edu/~nr/ffmpeg/chipmunks/bud-light.ts This bug report may be found as file README in that directory. I will be happy to upload this file to ftp.mplayerhq.hu, but even this 25-second clip is 53MB (greater than the recommended 10MB). HD files are large :-( I suspect this bug might be a fairly easy one to fix, if only I knew the code base. Probably the audio decoder is assuming that the format of the AC3 stream remains unchanged. Perhaps the problem could be fixed by borrowing a more pessimistic decoder from the 'a52dec' package. The 'a52dec' application plays the AC3 stream correctly. Norman Ramsey

11 24

[PATCH] part 4 - portability fixes (ILP32 vs LP64)
by freebsd＠sopwith.solgatos.com 28 Jan '06

28 Jan '06

Diffs for file: libavcodec/h264.c =================================================================== RCS file: RCS/h264.c,v retrieving revision 1.1 diff -u -r1.1 h264.c --- h264.c 2006/01/26 21:27:25 1.1 +++ h264.c 2006/01/26 21:27:43 @@ -5741,7 +5741,7 @@ return 0; } -void inline compute_mb_neighboors(H264Context *h) +static void inline compute_mb_neighboors(H264Context *h) { MpegEncContext * const s = &h->s; const int mb_xy = s->mb_x + s->mb_y*s->mb_stride;

2 1

[PATCH] part 2 - portability fixes (ILP32 vs LP64)
by freebsd＠sopwith.solgatos.com 28 Jan '06

28 Jan '06

>> Split up into small pieces in hopes of getting at least >> some small improvements checked in. > patch ok, can be applied assuming it compiles FINALLY! Here's part 2: Diffs for file: libavcodec/flicvideo.c =================================================================== RCS file: RCS/flicvideo.c,v retrieving revision 1.1 diff -u -r1.1 flicvideo.c --- flicvideo.c 2006/01/26 21:24:20 1.1 +++ flicvideo.c 2006/01/26 21:24:39 @@ -415,7 +415,7 @@ return buf_size; } -int flic_decode_frame_15_16BPP(AVCodecContext *avctx, +static int flic_decode_frame_15_16BPP(AVCodecContext *avctx, void *data, int *data_size, uint8_t *buf, int buf_size) {

2 1

[PATCH] part 1 - portability fixes (ILP32 vs LP64)
by freebsd＠sopwith.solgatos.com 28 Jan '06

28 Jan '06

Split up into small pieces in hopes of getting at least some small improvements checked in. Diffs for file: libavcodec/alac.c =================================================================== RCS file: RCS/alac.c,v retrieving revision 1.1 diff -u -r1.1 alac.c --- alac.c 2006/01/26 21:20:29 1.1 +++ alac.c 2006/01/26 21:21:18 @@ -84,7 +84,7 @@ alac->outputsamples_buffer_b = av_malloc(alac->setinfo_max_samples_per_frame * 4); } -void alac_set_info(ALACContext *alac) +static void alac_set_info(ALACContext *alac) { unsigned char *ptr = alac->avctx->extradata; @@ -125,7 +125,7 @@ return i; } -void bastardized_rice_decompress(ALACContext *alac, +static void bastardized_rice_decompress(ALACContext *alac, int32_t *output_buffer, int output_size, int readsamplesize, /* arg_10 */

2 1

[PATCH] missing newline
by bugfood-ml＠fatooh.org 27 Jan '06

27 Jan '06

This is trivial, but I don't want to step on anyone's toes by applying without asking first. If nobody objects, I'll apply this patch tomorrow. -Corey

2 2

[PATCH] part 3 - portability fixes (ILP32 vs LP64)
by freebsd＠sopwith.solgatos.com 27 Jan '06

27 Jan '06

Diffs for file: libavcodec/h261.c =================================================================== RCS file: RCS/h261.c,v retrieving revision 1.1 diff -u -r1.1 h261.c --- h261.c 2006/01/26 21:26:09 1.1 +++ h261.c 2006/01/26 21:26:25 @@ -762,7 +762,7 @@ * decodes the H261 picture header. * @return <0 if no startcode found */ -int h261_decode_picture_header(H261Context *h){ +static int h261_decode_picture_header(H261Context *h){ MpegEncContext * const s = &h->s; int format, i; uint32_t startcode= 0;

1 0

JPEG-LS decoder
by kostya.shishkov＠gmail.com 27 Jan '06

27 Jan '06

Please test this version. It now correctly decodes most of test data (except 12-bit images and images with subsampling). I also tried to remove all tabs and whitespace line ends. Comments are welcomed.

5 6

Re: [PATCH] portability fixes (ILP32 vs LP64)
by freebsd＠sopwith.solgatos.com 27 Jan '06

27 Jan '06

Prototypes either moved to .h file or changed to static. Diffs for file: libavcodec/liba52/crc.c =================================================================== RCS file: RCS/crc.c,v retrieving revision 1.1 diff -u -r1.1 crc.c --- crc.c 2006/01/26 21:41:09 1.1 +++ crc.c 2006/01/26 21:43:42 @@ -24,6 +24,7 @@ #include <stdlib.h> #include <stdio.h> #include <inttypes.h> +#include "common.h" static const uint16_t crc_lut[256] = { Diffs for file: libavcodec/alac.c =================================================================== RCS file: RCS/alac.c,v retrieving revision 1.1 diff -u -r1.1 alac.c --- alac.c 2006/01/26 21:20:29 1.1 +++ alac.c 2006/01/26 21:21:18 @@ -84,7 +84,7 @@ alac->outputsamples_buffer_b = av_malloc(alac->setinfo_max_samples_per_frame * 4); } -void alac_set_info(ALACContext *alac) +static void alac_set_info(ALACContext *alac) { unsigned char *ptr = alac->avctx->extradata; @@ -125,7 +125,7 @@ return i; } -void bastardized_rice_decompress(ALACContext *alac, +static void bastardized_rice_decompress(ALACContext *alac, int32_t *output_buffer, int output_size, int readsamplesize, /* arg_10 */ Diffs for file: libavcodec/flicvideo.c =================================================================== RCS file: RCS/flicvideo.c,v retrieving revision 1.1 diff -u -r1.1 flicvideo.c --- flicvideo.c 2006/01/26 21:24:20 1.1 +++ flicvideo.c 2006/01/26 21:24:39 @@ -415,7 +415,7 @@ return buf_size; } -int flic_decode_frame_15_16BPP(AVCodecContext *avctx, +static int flic_decode_frame_15_16BPP(AVCodecContext *avctx, void *data, int *data_size, uint8_t *buf, int buf_size) { Diffs for file: libavcodec/h261.c =================================================================== RCS file: RCS/h261.c,v retrieving revision 1.1 diff -u -r1.1 h261.c --- h261.c 2006/01/26 21:26:09 1.1 +++ h261.c 2006/01/26 21:26:25 @@ -762,7 +762,7 @@ * decodes the H261 picture header. * @return <0 if no startcode found */ -int h261_decode_picture_header(H261Context *h){ +static int h261_decode_picture_header(H261Context *h){ MpegEncContext * const s = &h->s; int format, i; uint32_t startcode= 0; Diffs for file: libavcodec/h264.c =================================================================== RCS file: RCS/h264.c,v retrieving revision 1.1 diff -u -r1.1 h264.c --- h264.c 2006/01/26 21:27:25 1.1 +++ h264.c 2006/01/26 21:27:43 @@ -5741,7 +5741,7 @@ return 0; } -void inline compute_mb_neighboors(H264Context *h) +static void inline compute_mb_neighboors(H264Context *h) { MpegEncContext * const s = &h->s; const int mb_xy = s->mb_x + s->mb_y*s->mb_stride; Diffs for file: libavcodec/mem.c =================================================================== RCS file: RCS/mem.c,v retrieving revision 1.1 diff -u -r1.1 mem.c --- mem.c 2006/01/26 20:57:04 1.1 +++ mem.c 2006/01/26 20:57:57 @@ -46,7 +46,7 @@ { void *ptr; #ifdef MEMALIGN_HACK - int diff; + long diff; #endif /* lets disallow possible ambiguous cases */ @@ -55,7 +55,7 @@ #ifdef MEMALIGN_HACK ptr = malloc(size+16+1); - diff= ((-(int)ptr - 1)&15) + 1; + diff= ((-(long)ptr - 1)&15) + 1; ptr += diff; ((char*)ptr)[-1]= diff; #elif defined (HAVE_MEMALIGN) Diffs for file: libavcodec/qdm2.c =================================================================== RCS file: RCS/qdm2.c,v retrieving revision 1.1 diff -u -r1.1 qdm2.c --- qdm2.c 2006/01/26 20:59:38 1.1 +++ qdm2.c 2006/01/26 21:29:25 @@ -230,7 +230,7 @@ static MPA_INT mpa_window[512] __attribute__((aligned(16))); -static void softclip_table_init() { +static void softclip_table_init(void) { int i; double dfl = SOFTCLIP_THRESHOLD - 32767; float delta = 1.0 / -dfl; @@ -240,7 +240,7 @@ // random generated table -static void rnd_table_init() { +static void rnd_table_init(void) { int i,j; uint32_t ldw,hdw; uint64_t tmp64_1; @@ -276,7 +276,7 @@ } -static void init_noise_samples() { +static void init_noise_samples(void) { int i; int random_seed = 0; float delta = 1.0 / 16384.0; @@ -287,7 +287,7 @@ } -static void qdm2_init_vlc() +static void qdm2_init_vlc(void) { init_vlc (&vlc_tab_level, 8, 24, vlc_tab_level_huffbits, 1, 1, @@ -525,7 +525,7 @@ * @param channels number of channels * @param coding_method q->coding_method[0][0][0] */ - void fix_coding_method_array (int sb, int channels, sb_int8_array coding_method) +static void fix_coding_method_array (int sb, int channels, sb_int8_array coding_method) { int j,k; int ch; @@ -1690,7 +1690,7 @@ * * @param q context */ -void qdm2_init(QDM2Context *q) { +static void qdm2_init(QDM2Context *q) { static int inited = 0; if (inited != 0) @@ -1942,7 +1942,7 @@ } -void qdm2_decode (QDM2Context *q, uint8_t *in, int16_t *out) +static void qdm2_decode (QDM2Context *q, uint8_t *in, int16_t *out) { int ch, i; const int frame_size = (q->frame_size * q->channels); Diffs for file: libavcodec/resample2.c =================================================================== RCS file: RCS/resample2.c,v retrieving revision 1.1 diff -u -r1.1 resample2.c --- resample2.c 2006/01/26 21:02:11 1.1 +++ resample2.c 2006/01/26 21:02:50 @@ -28,22 +28,6 @@ #include "common.h" #include "dsputil.h" -#if 1 -#define FILTER_SHIFT 15 - -#define FELEM int16_t -#define FELEM2 int32_t -#define FELEM_MAX INT16_MAX -#define FELEM_MIN INT16_MIN -#else -#define FILTER_SHIFT 22 - -#define FELEM int32_t -#define FELEM2 int64_t -#define FELEM_MAX INT32_MAX -#define FELEM_MIN INT32_MIN -#endif - typedef struct AVResampleContext{ FELEM *filter_bank; Diffs for file: libavcodec/snow.c =================================================================== RCS file: RCS/snow.c,v retrieving revision 1.1 diff -u -r1.1 snow.c --- snow.c 2006/01/26 21:03:39 1.1 +++ snow.c 2006/01/26 21:32:47 @@ -1576,7 +1576,7 @@ spatial_compose97i_dy(&cs, buffer, width, height, stride); } -void ff_spatial_idwt_buffered_init(dwt_compose_t *cs, slice_buffer * sb, int width, int height, int stride_line, int type, int decomposition_count){ +static void ff_spatial_idwt_buffered_init(dwt_compose_t *cs, slice_buffer * sb, int width, int height, int stride_line, int type, int decomposition_count){ int level; for(level=decomposition_count-1; level>=0; level--){ switch(type){ @@ -1589,7 +1589,7 @@ } } -void ff_spatial_idwt_init(dwt_compose_t *cs, DWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count){ +static void ff_spatial_idwt_init(dwt_compose_t *cs, DWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count){ int level; for(level=decomposition_count-1; level>=0; level--){ switch(type){ @@ -1601,7 +1601,7 @@ } } -void ff_spatial_idwt_slice(dwt_compose_t *cs, DWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count, int y){ +static void ff_spatial_idwt_slice(dwt_compose_t *cs, DWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count, int y){ const int support = type==1 ? 3 : 5; int level; if(type==2) return; @@ -1619,7 +1619,7 @@ } } -void ff_spatial_idwt_buffered_slice(dwt_compose_t *cs, slice_buffer * slice_buf, int width, int height, int stride_line, int type, int decomposition_count, int y){ +static void ff_spatial_idwt_buffered_slice(dwt_compose_t *cs, slice_buffer * slice_buf, int width, int height, int stride_line, int type, int decomposition_count, int y){ const int support = type==1 ? 3 : 5; int level; if(type==2) return; @@ -1637,7 +1637,7 @@ } } -void ff_spatial_idwt(DWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count){ +static void ff_spatial_idwt(DWTELEM *buffer, int width, int height, int stride, int type, int decomposition_count){ if(type==2){ int level; for(level=decomposition_count-1; level>=0; level--) @@ -3776,7 +3776,7 @@ return 0; } -static void init_qexp(){ +static void init_qexp(void){ int i; double v=128; Diffs for file: libavcodec/utils.c =================================================================== RCS file: RCS/utils.c,v retrieving revision 1.1 diff -u -r1.1 utils.c --- utils.c 2006/01/26 21:04:22 1.1 +++ utils.c 2006/01/26 21:04:46 @@ -149,9 +149,9 @@ * Call av_free_static automatically before it's too late */ -static void do_free() __attribute__ ((destructor)); +static void do_free(void) __attribute__ ((destructor)); -static void do_free() +static void do_free(void) { av_free_static(); } Diffs for file: libavcodec/vc9.c =================================================================== RCS file: RCS/vc9.c,v retrieving revision 1.1 diff -u -r1.1 vc9.c --- vc9.c 2006/01/26 21:34:40 1.1 +++ vc9.c 2006/01/26 23:13:08 @@ -929,7 +929,7 @@ * @todo TODO: Take into account stride * @todo TODO: Allow use of external buffers ? */ -int alloc_bitplane(BitPlane *bp, int width, int height) +static int alloc_bitplane(BitPlane *bp, int width, int height) { if (!bp || bp->width<0 || bp->height<0) return -1; bp->data = (uint8_t*)av_malloc(width*height); @@ -942,7 +942,7 @@ /** Free the bitplane's buffer * @param bp Bitplane which buffer is to free */ -void free_bitplane(BitPlane *bp) +static void free_bitplane(BitPlane *bp) { bp->width = bp->stride = bp->height = 0; if (bp->data) av_freep(&bp->data); @@ -1832,7 +1832,7 @@ * @todo TODO: Process the blocks * @todo TODO: Use M$ MPEG-4 cbp prediction */ -int vc9_decode_block(VC9Context *v, DCTELEM block[64], int n, int coded, int mquant) +static int vc9_decode_block(VC9Context *v, DCTELEM block[64], int n, int coded, int mquant) { GetBitContext *gb = &v->s.gb; MpegEncContext *s = &v->s; @@ -1979,7 +1979,7 @@ /** Decode one I-frame MB (in Simple/Main profile) * @todo TODO: Extend to AP */ -int vc9_decode_i_mb(VC9Context *v, DCTELEM block[6][64]) +static int vc9_decode_i_mb(VC9Context *v, DCTELEM block[6][64]) { int i, cbp, val; uint8_t *coded_val; @@ -2013,7 +2013,7 @@ * @todo TODO: Extend to AP * @fixme FIXME: DC value for inter blocks not set */ -int vc9_decode_p_mb(VC9Context *v, DCTELEM block[6][64]) +static int vc9_decode_p_mb(VC9Context *v, DCTELEM block[6][64]) { MpegEncContext *s = &v->s; GetBitContext *gb = &s->gb; @@ -2160,7 +2160,7 @@ * @warning XXX: Used for decoding BI MBs * @fixme FIXME: DC value for inter blocks not set */ -int vc9_decode_b_mb(VC9Context *v, DCTELEM block[6][64]) +static int vc9_decode_b_mb(VC9Context *v, DCTELEM block[6][64]) { MpegEncContext *s = &v->s; GetBitContext *gb = &v->s.gb; Diffs for file: libavcodec/vorbis.c =================================================================== RCS file: RCS/vorbis.c,v retrieving revision 1.1 diff -u -r1.1 vorbis.c --- vorbis.c 2006/01/26 21:40:00 1.1 +++ vorbis.c 2006/01/26 21:40:12 @@ -46,7 +46,7 @@ /** * reads 0-32 bits when using the ALT_BITSTREAM_READER_LE bitstream reader */ -unsigned int get_bits_long_le(GetBitContext *s, int n){ +static unsigned int get_bits_long_le(GetBitContext *s, int n){ if(n<=17) return get_bits(s, n); else{ int ret= get_bits(s, 16); Diffs for file: libavformat/avformat.h =================================================================== RCS file: RCS/avformat.h,v retrieving revision 1.1 diff -u -r1.1 avformat.h --- avformat.h 2006/01/26 22:44:56 1.1 +++ avformat.h 2006/01/26 22:45:46 @@ -713,5 +713,7 @@ } #endif +int gif_write(ByteIOContext *, AVImageInfo *); + #endif /* AVFORMAT_H */ Diffs for file: vhook/drawtext.c =================================================================== RCS file: RCS/drawtext.c,v retrieving revision 1.1 diff -u -r1.1 drawtext.c --- drawtext.c 2006/01/26 22:16:09 1.1 +++ drawtext.c 2006/01/26 22:16:35 @@ -300,7 +300,7 @@ -inline void draw_glyph(AVPicture *picture, FT_Bitmap *bitmap, unsigned int x, unsigned int y, unsigned int width, unsigned int height, unsigned char yuv_fgcolor[3], unsigned char yuv_bgcolor[3], int outline) +static inline void draw_glyph(AVPicture *picture, FT_Bitmap *bitmap, unsigned int x, unsigned int y, unsigned int width, unsigned int height, unsigned char yuv_fgcolor[3], unsigned char yuv_bgcolor[3], int outline) { int r, c; int spixel, dpixel[3], in_glyph=0; @@ -363,7 +363,7 @@ } -inline void draw_box(AVPicture *picture, unsigned int x, unsigned int y, unsigned int width, unsigned int height, unsigned char yuv_color[3]) +static inline void draw_box(AVPicture *picture, unsigned int x, unsigned int y, unsigned int width, unsigned int height, unsigned char yuv_color[3]) { int i, j; Diffs for file: vhook/ppm.c =================================================================== RCS file: RCS/ppm.c,v retrieving revision 1.1 diff -u -r1.1 ppm.c --- ppm.c 2006/01/26 22:14:17 1.1 +++ ppm.c 2006/01/26 22:15:26 @@ -40,7 +40,7 @@ /** Create a bidirectional pipe for the given command. */ -rwpipe *rwpipe_open( int argc, char *argv[] ) +static rwpipe *rwpipe_open( int argc, char *argv[] ) { rwpipe *this = av_mallocz( sizeof( rwpipe ) ); @@ -94,7 +94,7 @@ /** Read data from the pipe. */ -FILE *rwpipe_reader( rwpipe *this ) +static FILE *rwpipe_reader( rwpipe *this ) { if ( this != NULL ) return this->reader; @@ -105,7 +105,7 @@ /** Write data to the pipe. */ -FILE *rwpipe_writer( rwpipe *this ) +static FILE *rwpipe_writer( rwpipe *this ) { if ( this != NULL ) return this->writer; @@ -116,7 +116,7 @@ /* Read a number from the pipe - assumes PNM style headers. */ -int rwpipe_read_number( rwpipe *rw ) +static int rwpipe_read_number( rwpipe *rw ) { int value = 0; int c = 0; @@ -147,7 +147,7 @@ /** Read a PPM P6 header. */ -int rwpipe_read_ppm_header( rwpipe *rw, int *width, int *height ) +static int rwpipe_read_ppm_header( rwpipe *rw, int *width, int *height ) { char line[ 3 ]; FILE *in = rwpipe_reader( rw ); @@ -167,7 +167,7 @@ /** Close the pipe and process. */ -void rwpipe_close( rwpipe *this ) +static void rwpipe_close( rwpipe *this ) { if ( this != NULL ) { Diffs for file: libavutil/common.h =================================================================== RCS file: RCS/common.h,v retrieving revision 1.1 diff -u -r1.1 common.h --- common.h 2006/01/26 20:50:28 1.1 +++ common.h 2006/01/26 22:51:23 @@ -610,4 +610,52 @@ #endif /* HAVE_AV_CONFIG_H */ +/* moved from libavcodec/resample2.c + * needed here for av_build_filter() prototype below + */ +#if 1 +#define FILTER_SHIFT 15 + +#define FELEM int16_t +#define FELEM2 int32_t +#define FELEM_MAX INT16_MAX +#define FELEM_MIN INT16_MIN +#else +#define FILTER_SHIFT 22 + +#define FELEM int32_t +#define FELEM2 int64_t +#define FELEM_MAX INT32_MAX +#define FELEM_MIN INT32_MIN +#endif + + + /* prototypes */ +double av_int2dbl(int64_t); +float av_int2flt(int32_t); +int64_t av_dbl2int(double); +int32_t av_flt2int(float); +double bessel(double); +void av_build_filter(FELEM *, double, int, int, int, int); +void ff_spatial_dwt(int *, int, int, int, int, int); +void parse_arg_file(const char *); +uint16_t a52_crc16_block(uint8_t *, uint32_t); +void ff_h264_idct_add_mmx2(uint8_t *, int16_t *, int); +void ff_idct_xvid_mmx(short *); +void ff_idct_xvid_mmx2(short *); +void ff_vp3_dsp_init_mmx(void); +void ff_vp3_idct_mmx(int16_t *); +void ff_vp3_idct_sse2(int16_t *); +uint16_t a52_crc16_block(uint8_t *, uint32_t); +void* a52_resample_init(uint32_t, int, int); +void deinterlace_16(int32_t *, int32_t *, int16_t *, int, int, uint8_t, uint8_t); +uint16_t getbe16(const uint8_t *); +int get_nibble(const uint8_t *, int); +void set_service_cb(void *, int); +int ParseColor(char *, unsigned char [3]); +void parse_matrix_coeffs(uint16_t *, const char *); +void opt_inter_matrix(const char *); +void opt_intra_matrix(const char *); +void get_audio_frame(int16_t *, int, int); + #endif /* COMMON_H */ Diffs for file: ./output_example.c =================================================================== RCS file: RCS/output_example.c,v retrieving revision 1.1 diff -u -r1.1 output_example.c --- output_example.c 2006/01/26 22:11:23 1.1 +++ output_example.c 2006/01/26 22:13:15 @@ -51,7 +51,7 @@ /* * add an audio output stream */ -AVStream *add_audio_stream(AVFormatContext *oc, int codec_id) +static AVStream *add_audio_stream(AVFormatContext *oc, int codec_id) { AVCodecContext *c; AVStream *st; @@ -73,7 +73,7 @@ return st; } -void open_audio(AVFormatContext *oc, AVStream *st) +static void open_audio(AVFormatContext *oc, AVStream *st) { AVCodecContext *c; AVCodec *codec; @@ -139,7 +139,7 @@ } } -void write_audio_frame(AVFormatContext *oc, AVStream *st) +static void write_audio_frame(AVFormatContext *oc, AVStream *st) { AVCodecContext *c; AVPacket pkt; @@ -163,7 +163,7 @@ } } -void close_audio(AVFormatContext *oc, AVStream *st) +static void close_audio(AVFormatContext *oc, AVStream *st) { avcodec_close(st->codec); @@ -179,7 +179,7 @@ int frame_count, video_outbuf_size; /* add a video output stream */ -AVStream *add_video_stream(AVFormatContext *oc, int codec_id) +static AVStream *add_video_stream(AVFormatContext *oc, int codec_id) { AVCodecContext *c; AVStream *st; @@ -224,7 +224,7 @@ return st; } -AVFrame *alloc_picture(int pix_fmt, int width, int height) +static AVFrame *alloc_picture(int pix_fmt, int width, int height) { AVFrame *picture; uint8_t *picture_buf; @@ -244,7 +244,7 @@ return picture; } -void open_video(AVFormatContext *oc, AVStream *st) +static void open_video(AVFormatContext *oc, AVStream *st) { AVCodec *codec; AVCodecContext *c; @@ -293,7 +293,7 @@ } /* prepare a dummy image */ -void fill_yuv_image(AVFrame *pict, int frame_index, int width, int height) +static void fill_yuv_image(AVFrame *pict, int frame_index, int width, int height) { int x, y, i; @@ -315,7 +315,7 @@ } } -void write_video_frame(AVFormatContext *oc, AVStream *st) +static void write_video_frame(AVFormatContext *oc, AVStream *st) { int out_size, ret; AVCodecContext *c; @@ -380,7 +380,7 @@ frame_count++; } -void close_video(AVFormatContext *oc, AVStream *st) +static void close_video(AVFormatContext *oc, AVStream *st) { avcodec_close(st->codec); av_free(picture->data[0]);

2 1

uncompressed AVI output problem
by DKorchev＠aol.com 27 Jan '06

27 Jan '06

Hi, I was trying to use ffmpeg to convert video into uncompressed AVI ("-vcodec rawvideo" option) format and I found that output uncompressed AVI file cannot be used by Microsoft utility UncompAVIToWMV. I was able to fix this problem in ffmpeg\libavformat\avienc.c file by changing depth to 12 instead of 24 and frame buffer size to 1.5 * (width*height) instead of 3, see the changes in red below. I also think that these parameters are more logical for the video format I420 (yuv420). Any suggestions? Thanks, Dmitriy Korchev /* BITMAPINFOHEADER header */ void put_bmp_header(ByteIOContext *pb, AVCodecContext *enc, const CodecTag *tags, int for_asf) { put_le32(pb, 40 + enc->extradata_size); /* size */ put_le32(pb, enc->width); put_le32(pb, enc->height); put_le16(pb, 1); /* planes */ put_le16(pb, enc->bits_per_sample ? enc->bits_per_sample : 12); /* depth */ /* compression type */ put_le32(pb, for_asf ? codec_get_asf_tag(tags, enc->codec_id) : enc->codec_tag); put_le32(pb, enc->width * enc->height * 3/2); put_le32(pb, 0); put_le32(pb, 0); put_le32(pb, 0); put_le32(pb, 0); put_buffer(pb, enc->extradata, enc->extradata_size); if (enc->extradata_size & 1) put_byte(pb, 0); }

2 1

video4linux2 input format?
by lucabe72＠email.it 27 Jan '06

27 Jan '06

Hi all, someone managed to scare me telling me that video4linux support will be removed from future linux kernels... So, I think libavformat rally needs video4linux2 support. Is anyone already working on it? If not, I'll start implementing video4linux2 support today (I am asking just to avoid duplicate work). As usual, pointers to useful information are apreciated :) (I am currently reading http://v4l2spec.bytesex.org/spec ) Luca -- _____________________________________________________________________________ Copy this in your signature, if you think it is important: N O W A R ! ! ! -- Email.it, the professional e-mail, gratis per te: http://www.email.it/f Sponsor: L'auto dei tuoi sogni? Findomestic ti offre finanziamenti chiari e semplici senza anticipi! clicca qui Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=3937&d=25-1

5 14

compile vorbis error in ffmpeg using MinGW
by Baohua.WANG＠alcatel-sbell.com.cn 27 Jan '06

27 Jan '06

Dear All: I use the ffmpeg-0.4.9-pre1 version, add "--enable-vorbis" option to compile the ffmpeg. I had download the libvorbis-1.1.2 and libogg-1.1.3 and build , install successfully. But when I compile the ffmpeg. it shows the following error: ogg.o(.text+0x27):ogg.c: undefined reference to `ogg_stream_init' ogg.o(.text+0xc6):ogg.c: undefined reference to `ogg_stream_packetin' ogg.o(.text+0x1a9):ogg.c: undefined reference to `ogg_stream_packetin' ogg.o(.text+0x1f4):ogg.c: undefined reference to `ogg_stream_pageout' ogg.o(.text+0x255):ogg.c: undefined reference to `ogg_stream_flush' ogg.o(.text+0x2c3):ogg.c: undefined reference to `ogg_stream_flush' ogg.o(.text+0x2cf):ogg.c: undefined reference to `ogg_stream_clear' ogg.o(.text+0x2f9):ogg.c: undefined reference to `ogg_sync_init' ogg.o(.text+0x30a):ogg.c: undefined reference to `ogg_sync_buffer' ogg.o(.text+0x349):ogg.c: undefined reference to `ogg_sync_wrote' ogg.o(.text+0x355):ogg.c: undefined reference to `ogg_sync_pageout' ogg.o(.text+0x35d):ogg.c: undefined reference to `ogg_page_serialno' ogg.o(.text+0x369):ogg.c: undefined reference to `ogg_stream_init' ogg.o(.text+0x377):ogg.c: undefined reference to `ogg_stream_pagein' ogg.o(.text+0x3f9):ogg.c: undefined reference to `ogg_stream_packetout' ogg.o(.text+0x40f):ogg.c: undefined reference to `ogg_sync_buffer' ogg.o(.text+0x43c):ogg.c: undefined reference to `ogg_sync_wrote' ogg.o(.text+0x448):ogg.c: undefined reference to `ogg_sync_pageout' ogg.o(.text+0x457):ogg.c: undefined reference to `ogg_stream_pagein' ogg.o(.text+0x509):ogg.c: undefined reference to `ogg_stream_packetout' ogg.o(.text+0x51f):ogg.c: undefined reference to `ogg_sync_buffer' ogg.o(.text+0x54c):ogg.c: undefined reference to `ogg_sync_wrote' ogg.o(.text+0x558):ogg.c: undefined reference to `ogg_sync_pageout' ogg.o(.text+0x56a):ogg.c: undefined reference to `ogg_stream_pagein' ogg.o(.text+0x62c):ogg.c: undefined reference to `ogg_stream_clear' ogg.o(.text+0x634):ogg.c: undefined reference to `ogg_sync_clear' collect2: ld returned 1 exit status make[1]: *** [avformat.dll] Error 1 make[1]: Leaving directory `/home/Administrator/ffmpeg-0.4.9-pre1/libavformat' It seem that it cannot find the libogg. But I had put it in /MinGW/lib and /MinGW/MSYS/local/lib, How could I do? thanks !

2 1

problem decoding blank/black mpg. Gives 2 vertical strips when frame is saved
by pranavdesai＠gmail.com 27 Jan '06

27 Jan '06

Hello All, Hope this is the right mailing list for this questions. Let me know if it isnt. I am doing some testing on blank frame detection using lavc to detect commercial. I am using some code from mythtv commflag. So I thought I will start by testing it on a blank mpg itself. I created a blank/black mpg using the output_example.c provided with ffmpeg. I made all the Y values to 0 and UV values 128 in fill_yuv_image(). I am not sure if this the right way to do it, but it does generate a blank/black mpg (I can attach it if needed). I then use the avcodec_example.cpp provided here ( http://www.inb.uni-luebeck.de/~boehme/libavcodec_update.html) This link is on the ffmpeg doc pages. This program saves the first 5 frames of an mpg video. Below is a part of the code. When the frame is converted to RGB, it seems to correctly give a blank image. But if YUV is used , there are 2 vertical white stripes on the image. I have attached a jpg version of the frame saved using the program below. I am not sure what am I doing wrong. I would really appreciate if someone could help me out with this. Thanks for taking the time to read this and let me know if you need the above mentioned code. I can send it as a separate mail. Thanks -Pranav ------------------------------------------------------------------- while(av_read_frame(pFormatCtx, &packet)>=0) { // Is this a packet from the video stream? if(packet.stream_index==videoStream) { // Decode video frame avcodec_decode_video(pCodecCtx, pFrame, &frameFinished, packet.data, packet.size); // Did we get a video frame? if(frameFinished) { // Convert the image from its native format to RGB img_convert((AVPicture *)pFrameRGB, PIX_FMT_RGB24, (AVPicture*)pFrame, pCodecCtx->pix_fmt, pCodecCtx->width, pCodecCtx->height); // Save the frame to disk if(++i<=5) { //SaveFrame(pFrameRGB, pCodecCtx->width, pCodecCtx->height, i); SaveFrame(pFrame, pCodecCtx->width, pCodecCtx->height, i); } } } // Free the packet that was allocated by av_read_frame av_free_packet(&packet); } -------------------------------------------------------------------

1 0

Re: JPEG-LS decoder
by kostya.forjunk＠gmail.com 27 Jan '06

27 Jan '06

On Thu, Jan 26, 2006 at 11:17:11PM +0100, ffmpeg-devel-request at mplayerhq.hu wrote: > From: Michael Niedermayer <michaelni at gmx.at> > > Hi > > On Thu, Jan 26, 2006 at 08:40:19PM +0100, Roberto Togni wrote: > > On Sun, 22 Jan 2006 21:30:02 +0200 > > Kostya <kostya.shishkov at gmail.com> wrote: > > > > > Please test this version. It now correctly decodes most of test data > > > (except 12-bit images and images with subsampling). I also tried to > > > remove all tabs and whitespace line ends. > > > Comments are welcomed. > > > > Works for me, but the only sample i have worked also without the patch. > > > > Any comments about this? > > why not apply it? > > [...] > > -- > Michael > That was just a preview, I also plan to implement JPEG-LS encoder and split/rework/optimize some functions. BTW, I placed test images from T.87 standard at http://codecs.multimedia.cx/wp-content/ref_images.tar.bz2

1 0

DKorchev's AOL Address Card
by dkorchev＠aol.com 27 Jan '06

27 Jan '06

A non-text attachment was scrubbed... Name: dkorchev.vcf Type: text/x-vcard Size: 100 bytes Desc: not available URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20060126/6d53a…>

1 0

Makefile.bak removed on distclean
by diego＠biurrun.de 26 Jan '06

26 Jan '06

I've been wondering about "Makefile.bak" that is removed during distclean in some (but not all) Makefiles that MPlayer/FFmpeg uses. No Makefile.bak appears anywhere in the code, nor is it autogenerated. What am I missing, where does this rule come from? Cruft? Diego

2 2

[PATCH] set pix_fmt for video grabbing devices
by lucabe72＠email.it 26 Jan '06

26 Jan '06

Hello, when doing some experiments for testing my new v4l2 code, I noticed that ffmpeg forgets to set AVFormatParameters::pix_fmt when capturing from a v4l card. In other words, if you type ffmpeg -pix_fmt yuv422 -s 352x288 test.avi the "-pix_fmt yuv422" option will have no effect. The attached patch fixes the problem. Luca -- _____________________________________________________________________________ Copy this in your signature, if you think it is important: N O W A R ! ! ! -- Email.it, the professional e-mail, gratis per te: http://www.email.it/f Sponsor: Problemi di Liquidit?? Con Logos Finanziaria 30.000 ? in 24 ore a dipendenti e lavoratori autonomi con rimborsi fino a 120 mesi clicca qui * Clicca qui: http://adv.email.it/cgi-bin/foclick.cgi?mid=2907&d=26-1

3 2

[PATCH] Add support for Samsung Miniket camcorders files
by fabien.cosse＠wanadoo.fr 26 Jan '06

26 Jan '06

Hello, This little patch adds support for Samsung Miniket camcorders "strange" file format wich is in fact DIVX, but with fourcc "SEDG" regards.. Fabien Cosse

3 3

[PATCH] Fix leak of AVPacket data in ffserver
by bmayland＠leoninedev.com 26 Jan '06

26 Jan '06

I was noticing that ffserver memory usage steadily grows over time. There are a couple reasons for this, but here is the big one. When in the HTTPSTATE_SEND_DATA state, we're basically in a av_read_frame()/av_write_frame() loop. If the av_write_frame doesn't need a flush (which is the case when writing low bitrate [96kbit] asf streams), nothing is written to the dynbuff and the code loops back to read more frames to fill the outgoing buffer. In this case however, the data attached to the AVPacket is not freed with a call to av_free_packet. This patch fixes that in the common case as well as in the case a dynbuff could not be allocated. NOTE: This patch must be applied *after* my ffserver_warning_clean2.diff patch.

4 8

[PATCH] ffserver warning cleanup
by bmayland＠leoninedev.com 26 Jan '06

26 Jan '06

This patch cleans up all the compile warnings about signed/unsigned buffer pointers in ffserver.c. gcc version 4.0.2 20050808 (prerelease) (Ubuntu 4.0.1-4ubuntu9) I also added a feed_child_argv_free() method to free the memory allocated for the feed child_argv parameters. Since ffserver never really cleanly exits it isn't used except when the child process is exiting quickly. I can attach a signal handler so the http_server() function exits if requested, but I didn't want to do too much in this patch. Just a little clean up.

3 2

XOpendisplay
by mpegavc＠gmail.com 26 Jan '06

26 Jan '06

Hi, how can I solve this problem in the compilation process: ffplay.o(.text+0x4af4): In function `main': /root/ffmpeg/ffmpeg/ffplay.c:2455: undefined reference to `XOpenDisplay' ffplay.o(.text+0x4b2c):/root/ffmpeg/ffmpeg/ffplay.c:2459: undefined reference to`XCloseDisplay' collect2: ld returned 1 exit status make: *** [ffplay_g] Error 1 Thanks in advance, Peter

2 1

possible memory leak?
by nagyz＠nefty.hu 26 Jan '06

26 Jan '06

Hello! I've been using ffmpeg to play various streams, and I found out that my apps leak memory. the stream parameters: Input #0, sdp, from '/opt/mmsys/videos/2038_capture.sdp': Duration: N/A, bitrate: N/A Stream #0.0: Video: mpeg4, yuv420p, 320x240, 25.00 fps Stream #0.1: Audio: 0x0000 So I've cooked up a little sample application (basically using Martin Bohme's example), and ran it thru valgrind. the app basically does this: while(av_read_frame(pFormatCtx, &packet)>=0) { av_free_packet(&packet); if( SDL_GetTicks() - kezdet > 5000 ) { printf( "time's up!\n" ); break; } } so it quits after 5 secs, and it should free all packets. yet, valgrind sais: ==24793== LEAK SUMMARY: ==24793== definitely lost: 126,350 bytes in 208 blocks. ==24793== indirectly lost: 1,121 bytes in 1 blocks. ==24793== possibly lost: 2,293 bytes in 2 blocks. ==24793== still reachable: 18 bytes in 1 blocks. ==24793== suppressed: 0 bytes in 0 blocks. and the relevant info: ==24793== 84,329 (83,208 direct, 1,121 indirect) bytes in 72 blocks are definitely lost in loss record 9 of 9 ==24793== at 0x401CA6B: memalign (vg_replace_malloc.c:332) ==24793== by 0x40E374E: av_malloc (in /usr/local/lib/libavcodec-CVS.so) ==24793== by 0x403F3D5: av_new_packet (in /usr/local/lib/libavformat-CVS.so) ==24793== by 0x408EE37: rtp_parse_packet (in /usr/local/lib/libavformat-CVS.so) ==24793== by 0x408CF49: (within /usr/local/lib/libavformat-CVS.so) ==24793== by 0x404034A: (within /usr/local/lib/libavformat-CVS.so) ==24793== by 0x804893D: main (avcodec_sample.0.4.9.cpp:44) so basically it sais I didnt free the result of av_read_frame(). but I did! (line 44 is while(av_read_frame(..))) any ideas? I dont understand this, as I always free it. the source can be found at: http://nefty.hu/~nagyz/ffmpeg.c I'd really appreciate any help. Thanks, Zoltan NAGY, Software Engineer

3 4

ffserver Fix for segfault running RTSP streams, remove dummy_frame
by bmayland＠leoninedev.com 26 Jan '06

26 Jan '06

I've found the problem that causes segfaults in rtsp setup as well as required the use of a "dummy frame": /* I'm pretty sure that this is not correct... * However, without it, we crash */ However, I've still got two patches that are uncommited and it is becoming troublesome to supply patches from my branch. Is there any reason they are no good? ffserver_warning_clean.diff ffserver_packet_leak4.diff

1 0