October 2012 - ffmpeg-devel - lists.ffmpeg.org

[PATCH 1/2] img2: added support for %t output pattern
by Yuval Adam 05 Dec '16

05 Dec '16

From: Yuval Adam <yuv.adm(a)gmail.com> The image2 muxer now supports timestamps in output filenames. When used in an output patterm '%t' will be replaced with the frames timestamp in hours, minutes and seconds (hh:mm:ss). Signed-off-by: Yuval Adam <yuv.adm(a)gmail.com> --- doc/muxers.texi | 5 +++++ libavformat/avformat.h | 6 ++++-- libavformat/img2.c | 13 ++++++++----- libavformat/utils.c | 32 +++++++++++++++++++++++++------- 4 files changed, 42 insertions(+), 14 deletions(-) diff --git a/doc/muxers.texi b/doc/muxers.texi index 60a72d6..72212fe 100644 --- a/doc/muxers.texi +++ b/doc/muxers.texi @@ -109,6 +109,11 @@ If the pattern contains "%d" or "%0@var{N}d", the first filename of the file list specified will contain the number 1, all the following numbers will be sequential. +The pattern may also contain "%t", this string specifies the timestamp +of the frame in hours, minutes and seconds. Note that if two frames are written at the +same second, they will be overwritten, yielding a maximum output of 1 fps. +The "%t" and "%d" patterns may not be used simultaneously. + The pattern may contain a suffix which is used to automatically determine the format of the image files to write. diff --git a/libavformat/avformat.h b/libavformat/avformat.h index 2ed6b56..a9ab0cc 100644 --- a/libavformat/avformat.h +++ b/libavformat/avformat.h @@ -2067,7 +2067,8 @@ attribute_deprecated int find_info_tag(char *arg, int arg_size, const char *tag1 #endif /** - * Return in 'buf' the path with '%d' replaced by a number. + * Return in 'buf' the path with '%d' replaced by the frame number, + * and '%t' replaced by the frame timestamp. * * Also handles the '%0nd' format where 'n' is the total number * of digits and '%%'. @@ -2076,10 +2077,11 @@ attribute_deprecated int find_info_tag(char *arg, int arg_size, const char *tag1 * @param buf_size destination buffer size * @param path numbered sequence string * @param number frame number + * @param ts frame timestamp in seconds * @return 0 if OK, -1 on format error */ int av_get_frame_filename(char *buf, int buf_size, - const char *path, int number); + const char *path, int number, int ts); /** * Check whether filename actually is a numbered sequence generator. diff --git a/libavformat/img2.c b/libavformat/img2.c index bc35591..8012e5b 100644 --- a/libavformat/img2.c +++ b/libavformat/img2.c @@ -142,7 +142,7 @@ static int find_image_range(int *pfirst_index, int *plast_index, /* find the first image */ for(first_index = 0; first_index < 5; first_index++) { - if (av_get_frame_filename(buf, sizeof(buf), path, first_index) < 0){ + if (av_get_frame_filename(buf, sizeof(buf), path, first_index, 0) < 0){ *pfirst_index = *plast_index = 1; if (avio_check(buf, AVIO_FLAG_READ) > 0) @@ -165,7 +165,7 @@ static int find_image_range(int *pfirst_index, int *plast_index, else range1 = 2 * range; if (av_get_frame_filename(buf, sizeof(buf), path, - last_index + range1) < 0) + last_index + range1, 0) < 0) goto fail; if (avio_check(buf, AVIO_FLAG_READ) <= 0) break; @@ -310,7 +310,7 @@ static int read_packet(AVFormatContext *s1, AVPacket *pkt) if (s->img_number > s->img_last) return AVERROR_EOF; if (av_get_frame_filename(filename, sizeof(filename), - s->path, s->img_number)<0 && s->img_number > 1) + s->path, s->img_number, 0)<0 && s->img_number > 1) return AVERROR(EIO); for(i=0; i<3; i++){ if (avio_open2(&f[i], filename, AVIO_FLAG_READ, @@ -389,12 +389,15 @@ static int write_packet(AVFormatContext *s, AVPacket *pkt) VideoData *img = s->priv_data; AVIOContext *pb[3]; char filename[1024]; - AVCodecContext *codec= s->streams[ pkt->stream_index ]->codec; + AVStream *stream = s->streams[ pkt->stream_index ]; + AVCodecContext *codec = stream->codec; int i; + int ts = av_rescale_q(pkt->pts, stream->time_base, AV_TIME_BASE_Q) / AV_TIME_BASE; + if (!img->is_pipe) { if (av_get_frame_filename(filename, sizeof(filename), - img->path, img->img_number) < 0 && img->img_number>1 && !img->updatefirst) { + img->path, img->img_number, ts) < 0 && img->img_number>1 && !img->updatefirst) { av_log(s, AV_LOG_ERROR, "Could not get frame filename number %d from pattern '%s'\n", img->img_number, img->path); diff --git a/libavformat/utils.c b/libavformat/utils.c index f0df759..6db2db0 100644 --- a/libavformat/utils.c +++ b/libavformat/utils.c @@ -319,7 +319,7 @@ int av_append_packet(AVIOContext *s, AVPacket *pkt, int size) int av_filename_number_test(const char *filename) { char buf[1024]; - return filename && (av_get_frame_filename(buf, sizeof(buf), filename, 1)>=0); + return filename && (av_get_frame_filename(buf, sizeof(buf), filename, 1, 0)>=0); } AVInputFormat *av_probe_input_format3(AVProbeData *pd, int is_opened, int *score_ret) @@ -3774,15 +3774,16 @@ int find_info_tag(char *arg, int arg_size, const char *tag1, const char *info) #endif int av_get_frame_filename(char *buf, int buf_size, - const char *path, int number) + const char *path, int number, int ts) { const char *p; char *q, buf1[20], c; - int nd, len, percentd_found; + int nd, len, percent_found; + int hours, mins, secs; q = buf; p = path; - percentd_found = 0; + percent_found = 0; for(;;) { c = *p++; if (c == '\0') @@ -3800,9 +3801,9 @@ int av_get_frame_filename(char *buf, int buf_size, case '%': goto addchar; case 'd': - if (percentd_found) + if (percent_found) goto fail; - percentd_found = 1; + percent_found = 1; snprintf(buf1, sizeof(buf1), "%0*d", nd, number); len = strlen(buf1); if ((q - buf + len) > buf_size - 1) @@ -3810,6 +3811,23 @@ int av_get_frame_filename(char *buf, int buf_size, memcpy(q, buf1, len); q += len; break; + case 't': + if (percent_found) + goto fail; + percent_found = 1; + secs = ts % 60; + ts /= 60; + mins = ts % 60; + ts /= 60; + hours = ts; + snprintf(buf1, sizeof(buf1), + "%02d:%02d:%02d", hours, mins, secs); + len = strlen(buf1); + if ((q - buf + len) > buf_size - 1) + goto fail; + memcpy(q, buf1, len); + q += len; + break; default: goto fail; } @@ -3819,7 +3837,7 @@ int av_get_frame_filename(char *buf, int buf_size, *q++ = c; } } - if (!percentd_found) + if (!percent_found) goto fail; *q = '\0'; return 0; -- 1.7.2.2

4 22

Re: [FFmpeg-devel] [PATCH] Flush last frame in WMA lossless decoder
by Mashiat Sarker Shakkhar 10 Apr '16

10 Apr '16

On 5/2/2012 11:56 AM, Michael Niedermayer wrote: > On Wed, May 02, 2012 at 03:42:41AM +0600, Mashiat Sarker Shakkhar wrote: >> On 5/2/2012 2:49 AM, Michael Niedermayer wrote: >>> On Sun, Apr 29, 2012 at 10:43:47PM +0200, Jakub Stachowski wrote: >>>> Hello >>>> >>>> Here is cleaned up version of my patches to flush last frames and >>>> skip right number of samples in WMA lossless decoder. This causes >>>> ffmpeg to output the same number of samples as wmal2pcm.exe which I >>>> use as reference. [...] Hi The patch looks OK to the extent of my apprehension. And for the record, I have tested this patch and it fixes luckynight.wma. Patch has cosmetic issues, but that's another matter. Shakkhar P.S. Jakub, why did you use git send-email to send the patch? I tried and failed to review your patch on Gmane.

3 9

[PATCH] Support Ctrl+Break in ffmpeg.exe on Windows as if it was Ctrl+C
by Joe Wreschnig 30 Jun '15

30 Jun '15

On Windows, the Ctrl+Break key combination usually does what Ctrl+C does. It is more common for processes to send other processes Ctrl+Break rather than Ctrl+C, because sending Ctrl+C / SIGINT doesn't work if you started a child in a new process group. This patch makes ffmpeg.exe register what Windows calls a console control event handler, which allows it to intercept Ctrl+Break. It hands it off directly to the usual SIGINT/SIGTERM handler. The same function also processes closing the console window, mapping it to SIGTERM. Obviously, this is only enabled if compiling for a platform where SetConsoleCtrlHandler is available (i.e. modern Windows). (I've also submitted this as a ticket at https://ffmpeg.org/trac/ffmpeg/ticket/1463 but it was recommended I post it here also.)

5 17

drop entire frame when RTP packets are lost
by Martin Carroll 12 Jun '14

12 Jun '14

[Please forgive me if this issue has already been discussed -- I am new to ffmpeg-devel. I could not, however, find any discussion on this topic on this list in the past year, so here goes.] In the current design of ffplay/ffmpeg, when an RTP packet is lost, the containing -- and incomplete -- encoded frame is still passed along to the decoder. I have implemented a patch that has the read thread keep track of how many packets are missed, and when any are missed, the entire containing frame is dropped (that is, never makes it out of the read thread). I would be happy to send this patch upstream, or discuss it further if anyone likes. (The patch touches six files and about 25 lines of code.) Martin Carroll

4 20

[PATCH] Fix nonsense non-mod16 AMV flipping code.
by Reimar Döffinger 02 Jun '14

02 Jun '14

It is obviously nonsense since it produces wrong results or even crashes (crashes should be encode-only though). Fixes trac issue #1092. Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> --- libavcodec/mjpegdec.c | 3 +-- libavcodec/mjpegenc.c | 2 +- libavcodec/mpegvideo_enc.c | 5 ----- 3 files changed, 2 insertions(+), 8 deletions(-) diff --git a/libavcodec/mjpegdec.c b/libavcodec/mjpegdec.c index a841384..fe54b45 100644 --- a/libavcodec/mjpegdec.c +++ b/libavcodec/mjpegdec.c @@ -972,8 +972,7 @@ static int mjpeg_decode_scan(MJpegDecodeContext *s, int nb_components, int Ah, s->coefs_finished[c] |= 1; if (s->flipped) { // picture should be flipped upside-down for this codec - int offset = (linesize[c] * (s->v_scount[i] * - (8 * s->mb_height - ((s->height / s->v_max) & 7)) - 1)); + int offset = linesize[c] * (s->v_scount[c] * s->height / s->v_max - 1); data[c] += offset; reference_data[c] += offset; linesize[c] *= -1; diff --git a/libavcodec/mjpegenc.c b/libavcodec/mjpegenc.c index 9326d4d..4788c47 100644 --- a/libavcodec/mjpegenc.c +++ b/libavcodec/mjpegenc.c @@ -466,7 +466,7 @@ static int amv_encode_picture(AVCodecContext *avctx, AVPacket *pkt, //picture should be flipped upside-down for(i=0; i < 3; i++) { - pic.data[i] += (pic.linesize[i] * (s->mjpeg_vsample[i] * (8 * s->mb_height -((s->height/V_MAX)&7)) - 1 )); + pic.data[i] += pic.linesize[i] * (s->mjpeg_vsample[i] * s->height / V_MAX - 1); pic.linesize[i] *= -1; } return ff_MPV_encode_picture(avctx, pkt, &pic, got_packet); diff --git a/libavcodec/mpegvideo_enc.c b/libavcodec/mpegvideo_enc.c index 6d89cfe..bb4be28 100644 --- a/libavcodec/mpegvideo_enc.c +++ b/libavcodec/mpegvideo_enc.c @@ -1012,11 +1012,6 @@ static int load_input_picture(MpegEncContext *s, AVFrame *pic_arg) uint8_t *src = pic_arg->data[i]; uint8_t *dst = pic->data[i]; - if(s->codec_id == CODEC_ID_AMV){ - av_assert0(!(s->avctx->flags & CODEC_FLAG_EMU_EDGE)); - h= ((s->height+15)/16*16)>>v_shift; - } - if (!s->avctx->rc_buffer_size) dst += INPLACE_OFFSET; -- 1.7.10

3 13

[PATCH] rtpdec: make the NTP time values unsigned.
by flameeyes＠gmail.com 09 May '14

09 May '14

As per RFC3550, section 4, the NTP time is provided as 64-bit unsigned integer, so follow the same logic here. --- libavformat/rtpdec.h | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/libavformat/rtpdec.h b/libavformat/rtpdec.h index 86af2b9..87fd419 100644 --- a/libavformat/rtpdec.h +++ b/libavformat/rtpdec.h @@ -148,8 +148,8 @@ struct RTPDemuxContext { RTPStatistics statistics; ///< Statistics for this stream (used by RTCP receiver reports) /* rtcp sender statistics receive */ - int64_t last_rtcp_ntp_time; // TODO: move into statistics - int64_t first_rtcp_ntp_time; // TODO: move into statistics + uint64_t last_rtcp_ntp_time; // TODO: move into statistics + uint64_t first_rtcp_ntp_time; // TODO: move into statistics uint32_t last_rtcp_timestamp; // TODO: move into statistics /* rtcp sender statistics */ -- 1.7.1.1

4 3

[PATCH 0/5] Windows Resource File support
by jamal 26 Nov '13

26 Nov '13

This patchset adds support for Windows Resource files in Windows shared libraries. I thought about doing it for the executables as well, but I'd rather wait until i can get an icon for them. The first four commits add some missing version macros and header dependencies. With this the macros across all libraries will be consistent. This was done because they are needed by the rc files. I implemented this in all libraries except for libavresample. It's not enabled by default and the ffmpeg project isn't maintaining it. Regards jamal (5): avcodec: Add missing include to version.h postproc: Create version.h header swresample: Create version.h header Add missing version macros to libraries Add Windows Resource File support for shared libraries Makefile | 3 +- common.mak | 7 +++++- configure | 15 +++++++++-- libavcodec/Makefile | 3 ++ libavcodec/avcodecres.rc | 49 ++++++++++++++++++++++++++++++++++++++++ libavcodec/version.h | 2 + libavdevice/Makefile | 3 ++ libavdevice/avdeviceres.rc | 49 ++++++++++++++++++++++++++++++++++++++++ libavdevice/version.h | 2 + libavfilter/Makefile | 3 ++ libavfilter/avfilterres.rc | 49 ++++++++++++++++++++++++++++++++++++++++ libavfilter/version.h | 2 + libavformat/Makefile | 3 ++ libavformat/avformatres.rc | 49 ++++++++++++++++++++++++++++++++++++++++ libavutil/Makefile | 3 ++ libavutil/avutilres.rc | 49 ++++++++++++++++++++++++++++++++++++++++ libpostproc/Makefile | 6 ++++- libpostproc/postprocess.h | 18 +-------------- libpostproc/postprocessres.rc | 49 ++++++++++++++++++++++++++++++++++++++++ libpostproc/version.h | 47 ++++++++++++++++++++++++++++++++++++++ library.mak | 2 +- libswresample/Makefile | 6 ++++- libswresample/swresample.h | 8 +----- libswresample/swresampleres.rc | 49 ++++++++++++++++++++++++++++++++++++++++ libswresample/version.h | 45 ++++++++++++++++++++++++++++++++++++ libswscale/Makefile | 3 ++ libswscale/swscaleres.rc | 49 ++++++++++++++++++++++++++++++++++++++++ 27 files changed, 541 insertions(+), 32 deletions(-) create mode 100644 libavcodec/avcodecres.rc create mode 100644 libavdevice/avdeviceres.rc create mode 100644 libavfilter/avfilterres.rc create mode 100644 libavformat/avformatres.rc create mode 100644 libavutil/avutilres.rc create mode 100644 libpostproc/postprocessres.rc create mode 100644 libpostproc/version.h create mode 100644 libswresample/swresampleres.rc create mode 100644 libswresample/version.h create mode 100644 libswscale/swscaleres.rc -- 1.7.8.6

6 21

[PATCH] lavu: add av_get_color_string() and use it in -colors option
by stefano.sabatini-lala＠poste.it 27 Oct '13

27 Oct '13

--- cmdutils.c | 13 +++++++++++++ cmdutils.h | 6 ++++++ cmdutils_common_opts.h | 1 + doc/fftools-common-opts.texi | 3 +++ libavutil/parseutils.c | 18 ++++++++++++++++++ libavutil/parseutils.h | 14 ++++++++++++++ 6 files changed, 55 insertions(+), 0 deletions(-) diff --git a/cmdutils.c b/cmdutils.c index 514ebad..549c673 100644 --- a/cmdutils.c +++ b/cmdutils.c @@ -713,6 +713,19 @@ void show_pix_fmts(void) } } +void show_colors(void) +{ + char buf[256]; + int i; + + /* print header */ + fprintf(stderr, av_get_color_string(buf, sizeof(buf), -1)); + + for (i = 0; av_get_color_string(buf, sizeof(buf), i); i++) { + printf("%s", buf); + } +} + int read_yesno(void) { int c = getchar(); diff --git a/cmdutils.h b/cmdutils.h index c3d8a42..191ee36 100644 --- a/cmdutils.h +++ b/cmdutils.h @@ -195,6 +195,12 @@ void show_formats(void); void show_codecs(void); /** + * Print a listing containing all the color names supported by the + * program. + */ +void show_colors(void); + +/** * Print a listing containing all the filters supported by the * program. */ diff --git a/cmdutils_common_opts.h b/cmdutils_common_opts.h index da30997..1e3799a 100644 --- a/cmdutils_common_opts.h +++ b/cmdutils_common_opts.h @@ -6,6 +6,7 @@ { "version", OPT_EXIT, {(void*)show_version}, "show version" }, { "formats" , OPT_EXIT, {(void*)show_formats }, "show available formats" }, { "codecs" , OPT_EXIT, {(void*)show_codecs }, "show available codecs" }, + { "colors" , OPT_EXIT, {(void*)show_colors }, "show recognized color names" }, { "bsfs" , OPT_EXIT, {(void*)show_bsfs }, "show available bit stream filters" }, { "protocols", OPT_EXIT, {(void*)show_protocols}, "show available protocols" }, { "filters", OPT_EXIT, {(void*)show_filters }, "show available filters" }, diff --git a/doc/fftools-common-opts.texi b/doc/fftools-common-opts.texi index d72ca5c..5b25971 100644 --- a/doc/fftools-common-opts.texi +++ b/doc/fftools-common-opts.texi @@ -59,6 +59,9 @@ Codec can handle input truncated at random locations instead of only at frame bo @item -bsfs Show available bitstream filters. + at item -colors +Show recognized color names. + @item -protocols Show available protocols. diff --git a/libavutil/parseutils.c b/libavutil/parseutils.c index d67d31b..e701620 100644 --- a/libavutil/parseutils.c +++ b/libavutil/parseutils.c @@ -373,6 +373,24 @@ int av_parse_color(uint8_t *rgba_color, const char *color_string, int slen, return 0; } +char *av_get_color_string(char *buf, int buf_size, int color_idx) +{ + /* print header */ + if (color_idx < 0) { + snprintf(buf, buf_size, "%-32s #RRGGBB\n", "name"); + } else { + ColorEntry *color; + if (color_idx >= FF_ARRAY_ELEMS(color_table)) + return NULL; + color = &color_table[color_idx]; + snprintf(buf, buf_size, "%-32s #%02X%02X%02X\n", + color->name, + color->rgb_color[0], color->rgb_color[1], color->rgb_color[2]); + } + + return buf; +} + /* get a positive number between n_min and n_max, for a maximum length of len_max. Return -1 if error. */ static int date_get_num(const char **pp, diff --git a/libavutil/parseutils.h b/libavutil/parseutils.h index c3986af..3e3a34b 100644 --- a/libavutil/parseutils.h +++ b/libavutil/parseutils.h @@ -73,6 +73,20 @@ int av_parse_color(uint8_t *rgba_color, const char *color_string, int slen, void *log_ctx); /** + * Print in buf the string corresponding to the color with + * number color_idx, or an header if color_idx is negative. + * + * @param buf the buffer where to write the string + * @param buf_size the size of buf + * @param color_idx the number of the pixel format to print the + * corresponding info string, or a negative value to print the + * corresponding header. + * @return the printed string or NULL if color_idx doesn't correspond to any + * color + */ +char *av_get_color_string(char *buf, int buf_size, int color_idx); + +/** * Parses timestr and returns in *time a corresponding number of * microseconds. * -- 1.7.2.3

5 18

[PATCH] ffprobe: add -ss option
by Stefano Sabatini 19 Sep '13

19 Sep '13

This is useful to test seeking on an input file. TODO: add Changelog entry --- doc/ffprobe.texi | 4 ++++ ffprobe.c | 26 ++++++++++++++++++++++++++ 2 files changed, 30 insertions(+), 0 deletions(-) diff --git a/doc/ffprobe.texi b/doc/ffprobe.texi index 6295eb3..c2adcc4 100644 --- a/doc/ffprobe.texi +++ b/doc/ffprobe.texi @@ -162,6 +162,10 @@ corresponding stream section. Count the number of packets per stream and report it in the corresponding stream section. +@item -ss @var{seek_point} +Seek input to position specified by @var{seek_point}. +@var{seek_point} must be a duration specification. + @item -show_private_data, -private Show private data, that is data depending on the format of the particular shown element. diff --git a/ffprobe.c b/ffprobe.c index 1cf6c42..27f8638 100644 --- a/ffprobe.c +++ b/ffprobe.c @@ -70,6 +70,7 @@ static int show_private_data = 1; static char *print_format; static char *stream_specifier; +static int64_t seek_point = -INT64_MAX; /* section structure definition */ @@ -1863,6 +1864,24 @@ static int open_input_file(AVFormatContext **fmt_ctx_ptr, const char *filename) av_dump_format(fmt_ctx, 0, filename, 0); + if (seek_point > -INT64_MAX) { + // add the stream start time, should it exist + if (fmt_ctx->start_time != AV_NOPTS_VALUE) { + if (seek_point > INT64_MAX - fmt_ctx->start_time) { + av_log(NULL, AV_LOG_ERROR, + "Seek point value %"PRId64" overflow with start time value:%"PRId64"\n", + seek_point, fmt_ctx->start_time); + return AVERROR(EINVAL); + } + seek_point += fmt_ctx->start_time; + } + if ((err = av_seek_frame(fmt_ctx, -1, seek_point, AVSEEK_FLAG_BACKWARD)) < 0) { + av_log(NULL, AV_LOG_ERROR, "Could not seek to position %"PRId64"\n", + seek_point); + return err; + } + } + /* bind a decoder to each input stream */ for (i = 0; i < fmt_ctx->nb_streams; i++) { AVStream *stream = fmt_ctx->streams[i]; @@ -2056,6 +2075,12 @@ void show_help_default(const char *opt, const char *arg) show_help_children(avformat_get_class(), AV_OPT_FLAG_DECODING_PARAM); } +static int opt_seek_point(void *optctx, const char *opt, const char *arg) +{ + seek_point = parse_time_or_die(opt, arg, 1); + return 0; +} + static int opt_pretty(void *optctx, const char *opt, const char *arg) { show_value_unit = 1; @@ -2103,6 +2128,7 @@ static const OptionDef real_options[] = { { "show_private_data", OPT_BOOL, {(void*)&show_private_data}, "show private data" }, { "private", OPT_BOOL, {(void*)&show_private_data}, "same as show_private_data" }, { "bitexact", OPT_BOOL, {&do_bitexact}, "force bitexact output" }, + { "ss", HAS_ARG, {.func_arg = opt_seek_point}, "seek input to the specified seek point", "seek_point" }, { "default", HAS_ARG | OPT_AUDIO | OPT_VIDEO | OPT_EXPERT, {.func_arg = opt_default}, "generic catch all option", "" }, { "i", HAS_ARG, {.func_arg = opt_input_file_i}, "read specified file", "input_file"}, { NULL, }, -- 1.7.5.4

3 12

[RFC] Dynamic filtergraph reconfiguration
by Stefano Sabatini 16 Sep '13

16 Sep '13

Hi, maybe it's time to finally tackle this, since it seems to be a recurring issue. Sometimes the input of the filterchain changes configuration parameters. For audio it means it could change channel layout/sample rate/sample format/timebase, for video it may change frame rate, size, pixel format, sample aspect ratio and timebase. Basically most filters assume these values to be invariant during the whole life of the filtergraph, others (e.g. the scale filter) monitors *some* of the input values and reconfigure themselves accordingly, sometimes calling configure on the following filters (which may or may not be correct). We need to make reconfiguration consistently applied at the framework level, and properly design it, in order to avoid brittle code and glitches. Problems to tackle are: - not all filters may support reconfigurations, for example some filters rely on some specific values defined during configuration, so in general we need a mechanism which allows to reconfigure only the filters which can do that - some filters may support only *one* parameter change (e.g. timebase may easily change). - since we don't mark codecs/containers which support variable configuration parameters (do we?), we may need to "normalize" the output before the terminal sink (depending on the application). Related questions: where should we check the changes (possibly in the framework code)? Where should we add the "buffer" filters which normalize input towards a "non-reconfigurable" filter (possibly we should auto-insert it only when required). Filter supporting reconfiguration should be marked as such (or alternatively we may mark filters *not* supporting it). Also I wonder if we should also support filters where the *number* of inputs or outputs change dinamically (e.g. a movie source which adds a new stream). I'll try to come with a more detailed design and proof of concept once we'll clear some of the design issues. -- FFmpeg = Frightening and Faithful Murdering Pacific Explosive Guru

4 3