July 2013 - 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

[PATCH]Accept startcode 001 when remuxing H264 into mpegts
by Carl Eugen Hoyos 26 Feb '15

26 Feb '15

Hi! Attached patch fixes remuxing of h264 into mpeg-ts for some wtv input files. Please review, Carl Eugen

2 3

[PATCH 1/2] lavfi: remove mp=ilpack filter
by Paul B Mahol 27 Jan '15

27 Jan '15

The filter does not work, asserts here. Signed-off-by: Paul B Mahol <onemda(a)gmail.com> --- doc/filters.texi | 1 - libavfilter/Makefile | 1 - libavfilter/libmpcodecs/vf_ilpack.c | 458 ------------------------------------ libavfilter/vf_mp.c | 2 - 4 files changed, 462 deletions(-) delete mode 100644 libavfilter/libmpcodecs/vf_ilpack.c Boring stuff removed.

6 13

[PATCH] cdxl: fix duration
by Piotr Bandurski 22 Dec '14

22 Dec '14

fixes ticket #1937

4 10

One pass volume normalization (ebur128)
by Jan Ehrhardt 15 Jul '14

15 Jul '14

I am once again proposing a patch for one pass volume normalization base on ebur128, as I see this still did not make it into FFMpeg 2.0. My patch is heaveily based on Clément Bsch's proposal in http://permalink.gmane.org/gmane.comp.video.ffmpeg.devel/159978 We have been using this patch now for more than 4 months and 1800+ videos of approximately 1 hour have been transcoded with it. Part of our FFMpeg commandline reads as -filter_complex \ "[0:v]setpts=PTS-STARTPTS[v0];[0:a]asetpts=PTS-STARTPTS,ebur128=metadata=1,volume=metadata=lavfi.r128.I,ebur128[a0]" \ -map [v0] -map [a0] It uses the already present ebur128 meta injection to adjust the volume on the fly. What would be the objection to move this into the FFMpeg core, so I do not have to patch my FFMpeg every time I compile a new one? I applied the patch below to FFMpeg Release/v.2.0. Jan diff --git a/libavfilter/af_volume.c b/libavfilter/af_volume.c index a2ac1e2..6372bb2 100644 --- a/libavfilter/af_volume.c +++ b/libavfilter/af_volume.c @@ -51,18 +51,24 @@ static const AVOption volume_options[] = { { "fixed", "select 8-bit fixed-point", 0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_FIXED }, INT_MIN, INT_MAX, A|F, "precision" }, { "float", "select 32-bit floating-point", 0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_FLOAT }, INT_MIN, INT_MAX, A|F, "precision" }, { "double", "select 64-bit floating-point", 0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_DOUBLE }, INT_MIN, INT_MAX, A|F, "precision" }, + { "metadata", "set the metadata key for loudness normalization", OFFSET(metadata), AV_OPT_TYPE_STRING, { .str = NULL }, .flags = A|F }, { NULL }, }; AVFILTER_DEFINE_CLASS(volume); +static void set_fixed_volume(VolumeContext *vol, double volume) +{ + vol->volume_i = (int)(volume * 256 + 0.5); + vol->volume = vol->volume_i / 256.0; +} + static av_cold int init(AVFilterContext *ctx) { VolumeContext *vol = ctx->priv; if (vol->precision == PRECISION_FIXED) { - vol->volume_i = (int)(vol->volume * 256 + 0.5); - vol->volume = vol->volume_i / 256.0; + set_fixed_volume(vol, vol->volume); av_log(ctx, AV_LOG_VERBOSE, "volume:(%d/256)(%f)(%1.2fdB) precision:fixed\n", vol->volume_i, vol->volume, 20.0*log(vol->volume)/M_LN10); } else { @@ -171,13 +177,13 @@ static av_cold void volume_init(VolumeContext *vol) switch (av_get_packed_sample_fmt(vol->sample_fmt)) { case AV_SAMPLE_FMT_U8: - if (vol->volume_i < 0x1000000) + if (vol->volume_i < 0x1000000 && !vol->metadata) vol->scale_samples = scale_samples_u8_small; else vol->scale_samples = scale_samples_u8; break; case AV_SAMPLE_FMT_S16: - if (vol->volume_i < 0x10000) + if (vol->volume_i < 0x10000 && !vol->metadata) vol->scale_samples = scale_samples_s16_small; else vol->scale_samples = scale_samples_s16; @@ -216,11 +222,30 @@ static int config_output(AVFilterLink *outlink) static int filter_frame(AVFilterLink *inlink, AVFrame *buf) { - VolumeContext *vol = inlink->dst->priv; - AVFilterLink *outlink = inlink->dst->outputs[0]; + AVFilterContext *ctx = inlink->dst; + VolumeContext *vol = ctx->priv; + AVFilterLink *outlink = ctx->outputs[0]; int nb_samples = buf->nb_samples; AVFrame *out_buf; + if (vol->metadata) { + double loudness, new_volume, timestamp, mx; + AVDictionaryEntry *e; + mx = 20; + timestamp = (float)(1.0 * buf->pts / outlink->sample_rate); + mx = fmin(mx, timestamp); + e = av_dict_get(buf->metadata, vol->metadata, NULL, 0); + if (e) { + loudness = av_strtod(e->value, NULL); + if (loudness > -69) { + new_volume = fmax(-mx,fmin(mx,(-23 - loudness))); + av_log(NULL, AV_LOG_VERBOSE, "loudness=%f => %f => volume=%f\n", + loudness, new_volume, pow(10, new_volume / 20)); + set_fixed_volume(vol, pow(10, new_volume / 20)); + } + } + } + if (vol->volume == 1.0 || vol->volume_i == 256) return ff_filter_frame(outlink, buf); @@ -269,6 +294,12 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *buf) return ff_filter_frame(outlink, out_buf); } +static av_cold void uninit(AVFilterContext *ctx) +{ + VolumeContext *vol = ctx->priv; + av_opt_free(vol); +} + static const AVFilterPad avfilter_af_volume_inputs[] = { { .name = "default", @@ -294,6 +325,7 @@ AVFilter avfilter_af_volume = { .priv_size = sizeof(VolumeContext), .priv_class = &volume_class, .init = init, + .uninit = uninit, .inputs = avfilter_af_volume_inputs, .outputs = avfilter_af_volume_outputs, .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC, diff --git a/libavfilter/af_volume.h b/libavfilter/af_volume.h index bd7932e..4deca9c 100644 --- a/libavfilter/af_volume.h +++ b/libavfilter/af_volume.h @@ -48,6 +48,7 @@ typedef struct VolumeContext { void (*scale_samples)(uint8_t *dst, const uint8_t *src, int nb_samples, int volume); int samples_align; + char *metadata; } VolumeContext; void ff_volume_init_x86(VolumeContext *vol); diff --git a/libavfilter/f_ebur128.c b/libavfilter/f_ebur128.c index 88d37e8..f4ce6d9 100644 --- a/libavfilter/f_ebur128.c +++ b/libavfilter/f_ebur128.c @@ -410,7 +410,7 @@ static av_cold int init(AVFilterContext *ctx) if (ebur128->loglevel != AV_LOG_INFO && ebur128->loglevel != AV_LOG_VERBOSE) { - if (ebur128->do_video || ebur128->metadata) + if (ebur128->do_video) ebur128->loglevel = AV_LOG_VERBOSE; else ebur128->loglevel = AV_LOG_INFO; @@ -689,7 +689,7 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *insamples) SET_META("LRA.high", ebur128->lra_high); } - av_log(ctx, ebur128->loglevel, "t: %-10s " LOG_FMT "\n", + av_log(ctx, ebur128->metadata || !ebur128->do_video ? AV_LOG_VERBOSE : ebur128->loglevel, "t: %-10s " LOG_FMT "\n", av_ts2timestr(pts, &outlink->time_base), loudness_400, loudness_3000, ebur128->integrated_loudness, ebur128->loudness_range); diff --git a/libavfilter/x86/af_volume_init.c b/libavfilter/x86/af_volume_init.c index 81d605f..fab5a03 100644 --- a/libavfilter/x86/af_volume_init.c +++ b/libavfilter/x86/af_volume_init.c @@ -39,7 +39,7 @@ av_cold void ff_volume_init_x86(VolumeContext *vol) enum AVSampleFormat sample_fmt = av_get_packed_sample_fmt(vol->sample_fmt); if (sample_fmt == AV_SAMPLE_FMT_S16) { - if (EXTERNAL_SSE2(mm_flags) && vol->volume_i < 32768) { + if (EXTERNAL_SSE2(mm_flags) && vol->volume_i < 32768 && !vol->metadata) { vol->scale_samples = ff_scale_samples_s16_sse2; vol->samples_align = 8; }

9 46

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]Refuse to mux invalid transport streams
by Carl Eugen Hoyos 08 Jun '14

08 Jun '14

Hi! FFmpeg currently allows to mux transport streams that cannot be decoded, attached patch allows to report an error in such a case. Please comment, Carl Eugen

3 8

[PATCH 0/5] Scalar product, take 2
by Christophe Gisquet 06 Jun '14

06 Jun '14

Following the previous comments, I have opted for padding of the buffers in order to use the DSP functions. The troublesome cases for wma lossless are still untested. Patches 3 and 4 still try to address them, 3 being more secure than 4 (which is the shape of the best solution in my opinion). TAK decoder was modified in the same fashion. Christophe Gisquet (5): ra144: use scalarproduct_int16 wma lossless: reuse scalarproduct_and_madd_int16 wma lossless: switching back to C dsp function wma lossless: pad coeff buffer with 0 takdec: pad filter coeff buffer for DSP functions libavcodec/dsputil.c | 4 +-- libavcodec/dsputil.h | 3 +++ libavcodec/ra144.c | 14 ++++------ libavcodec/ra144.h | 6 ++++- libavcodec/ra144dec.c | 3 +++ libavcodec/ra144enc.c | 6 +++-- libavcodec/takdec.c | 33 +++++++----------------- libavcodec/wmalosslessdec.c | 63 ++++++++++++++++++++++++--------------------- 8 files changed, 65 insertions(+), 67 deletions(-) -- 1.8.0.msysgit.0

3 26