October 2018 - ffmpeg-cvslog - lists.ffmpeg.org

cbs_h265: Add PTL parsing for Main 10 Still Picture profile
by Mark Thompson 31 Oct '18

31 Oct '18

ffmpeg | branch: master | Mark Thompson <sw(a)jkqxz.net> | Sat Oct 27 22:39:47 2018 +0100| [2dee0679e9a88e1fea51ebd3dcd2bbbf6e1777d6] | committer: Mark Thompson cbs_h265: Add PTL parsing for Main 10 Still Picture profile This was added in the 2018 version of the standard. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=2dee0679e9a88e1fe… --- libavcodec/cbs_h265_syntax_template.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/libavcodec/cbs_h265_syntax_template.c b/libavcodec/cbs_h265_syntax_template.c index d4e4f7b1c2..e43f3caf99 100644 --- a/libavcodec/cbs_h265_syntax_template.c +++ b/libavcodec/cbs_h265_syntax_template.c @@ -130,6 +130,11 @@ static int FUNC(profile_tier_level)(CodedBitstreamContext *ctx, RWContext *rw, fixed(24, general_reserved_zero_34bits, 0); fixed(10, general_reserved_zero_34bits, 0); } + } else if (profile_compatible(2)) { + fixed(7, general_reserved_zero_7bits, 0); + flag(general_one_picture_only_constraint_flag); + fixed(24, general_reserved_zero_35bits, 0); + fixed(11, general_reserved_zero_35bits, 0); } else { fixed(24, general_reserved_zero_43bits, 0); fixed(19, general_reserved_zero_43bits, 0);

1 0

cbs_h264: Include SEI type names in trace output
by Mark Thompson 31 Oct '18

31 Oct '18

ffmpeg | branch: master | Mark Thompson <sw(a)jkqxz.net> | Sat Oct 27 22:39:46 2018 +0100| [30fcc10937966cd24b253d4b8267c455f2212b8a] | committer: Mark Thompson cbs_h264: Include SEI type names in trace output > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=30fcc10937966cd24… --- libavcodec/cbs_h264_syntax_template.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/libavcodec/cbs_h264_syntax_template.c b/libavcodec/cbs_h264_syntax_template.c index 9e29132fec..dbf9ff1268 100644 --- a/libavcodec/cbs_h264_syntax_template.c +++ b/libavcodec/cbs_h264_syntax_template.c @@ -513,6 +513,8 @@ static int FUNC(sei_buffering_period)(CodedBitstreamContext *ctx, RWContext *rw, const H264RawSPS *sps; int err, i, length; + HEADER("Buffering Period"); + ue(seq_parameter_set_id, 0, 31); sps = h264->sps[current->seq_parameter_set_id]; @@ -605,6 +607,8 @@ static int FUNC(sei_pic_timing)(CodedBitstreamContext *ctx, RWContext *rw, const H264RawSPS *sps; int err; + HEADER("Picture Timing"); + sps = h264->active_sps; if (!sps) { // If there is exactly one possible SPS but it is not yet active @@ -674,6 +678,8 @@ static int FUNC(sei_pan_scan_rect)(CodedBitstreamContext *ctx, RWContext *rw, { int err, i; + HEADER("Pan-Scan Rectangle"); + ue(pan_scan_rect_id, 0, UINT32_MAX - 1); flag(pan_scan_rect_cancel_flag); @@ -699,6 +705,8 @@ static int FUNC(sei_user_data_registered)(CodedBitstreamContext *ctx, RWContext { int err, i, j; + HEADER("User Data Registered ITU-T T.35"); + u(8, itu_t_t35_country_code, 0x00, 0xff); if (current->itu_t_t35_country_code != 0xff) i = 1; @@ -731,6 +739,8 @@ static int FUNC(sei_user_data_unregistered)(CodedBitstreamContext *ctx, RWContex { int err, i; + HEADER("User Data Unregistered"); + #ifdef READ if (*payload_size < 16) { av_log(ctx->log_ctx, AV_LOG_ERROR, @@ -758,6 +768,8 @@ static int FUNC(sei_recovery_point)(CodedBitstreamContext *ctx, RWContext *rw, { int err; + HEADER("Recovery Point"); + ue(recovery_frame_cnt, 0, 65535); flag(exact_match_flag); flag(broken_link_flag); @@ -771,6 +783,8 @@ static int FUNC(sei_display_orientation)(CodedBitstreamContext *ctx, RWContext * { int err; + HEADER("Display Orientation"); + flag(display_orientation_cancel_flag); if (!current->display_orientation_cancel_flag) { flag(hor_flip); @@ -788,6 +802,8 @@ static int FUNC(sei_mastering_display_colour_volume)(CodedBitstreamContext *ctx, { int err, c; + HEADER("Mastering Display Colour Volume"); + for (c = 0; c < 3; c++) { us(16, display_primaries_x[c], 0, 50000, 1, c); us(16, display_primaries_y[c], 0, 50000, 1, c);

1 0

doc/fate.texi: Mention that samples should be uploaded before pushing dependent commits
by Michael Niedermayer 31 Oct '18

31 Oct '18

ffmpeg | branch: master | Michael Niedermayer <michael(a)niedermayer.cc> | Tue Oct 23 23:31:24 2018 +0200| [793a3e7bc909ee96bb253950bce3d3b0e2e188a2] | committer: Michael Niedermayer doc/fate.texi: Mention that samples should be uploaded before pushing dependent commits Signed-off-by: Michael Niedermayer <michael(a)niedermayer.cc> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=793a3e7bc909ee96b… --- doc/fate.texi | 2 ++ 1 file changed, 2 insertions(+) diff --git a/doc/fate.texi b/doc/fate.texi index a352994230..2be61d639c 100644 --- a/doc/fate.texi +++ b/doc/fate.texi @@ -155,6 +155,8 @@ space on each client, network bandwidth and so on benefit from smaller test case Also keep in mind older checkouts use existing sample files, that means in practice generally do not replace, remove or overwrite files as it likely would break older checkouts or releases. +Also all needed samples for a commit should be uploaded, ideally 24 +hours, before the push. @example #First update your local samples copy:

1 0

[ffmpeg-web] branch master updated. c38da5e web/documentation: add new book about FFmpeg in China
by ffmpeg-git＠ffmpeg.org 31 Oct '18

31 Oct '18

The branch, master has been updated via c38da5ec2cdb67c634c4ad6001ad5a7b23383e78 (commit) from e51f0b6f978d9958a0abfb59da2e674d7e22ea3f (commit) - Log ----------------------------------------------------------------- commit c38da5ec2cdb67c634c4ad6001ad5a7b23383e78 Author: Steven Liu <lq(a)chinaffmpeg.org> AuthorDate: Thu Oct 18 07:40:09 2018 +0800 Commit: Lou Logan <lou(a)lrcd.com> CommitDate: Wed Oct 31 12:50:10 2018 -0800 web/documentation: add new book about FFmpeg in China Signed-off-by: Steven Liu <lq(a)chinaffmpeg.org> diff --git a/src/documentation b/src/documentation index e3bbf4c..a975002 100644 --- a/src/documentation +++ b/src/documentation @@ -131,6 +131,8 @@ <ul> <li><a href="http://ffmpeg.tv">FFmpeg Basics</a> by Frantisek Korbel, describes various FFmpeg features and common tasks.</li> + <li><a href="http://book.chinaffmpeg.com">FFmpeg Beginner's handbook Chinese Version</a> by Steven Liu, + describes FFmpeg common use method in Chinese, from command line to API usage.</li> </ul> </div> </div>  ----------------------------------------------------------------------- Summary of changes: src/documentation | 2 ++ 1 file changed, 2 insertions(+) hooks/post-receive --

1 0

avfilter: add (a)graphmonitor filter(s)
by Paul B Mahol 31 Oct '18

31 Oct '18

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Sat Oct 27 19:17:08 2018 +0200| [323c2cfd384a06ad1e5233306156eb931185193f] | committer: Paul B Mahol avfilter: add (a)graphmonitor filter(s) > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=323c2cfd384a06ad1… --- Changelog | 1 + doc/filters.texi | 57 ++++++ libavfilter/Makefile | 2 + libavfilter/allfilters.c | 2 + libavfilter/f_graphmonitor.c | 425 +++++++++++++++++++++++++++++++++++++++++++ libavfilter/version.h | 2 +- 6 files changed, 488 insertions(+), 1 deletion(-) diff --git a/Changelog b/Changelog index de0383047e..5ce1233b10 100644 --- a/Changelog +++ b/Changelog @@ -41,6 +41,7 @@ version <next>: - decoding S12M timecode in h264 - xstack filter - pcm vidc decoder and encoder +- (a)graphmonitor filter version 4.0: diff --git a/doc/filters.texi b/doc/filters.texi index 9b84b1145b..200af0d2d1 100644 --- a/doc/filters.texi +++ b/doc/filters.texi @@ -10300,6 +10300,63 @@ gradfun=radius=8 @end itemize +@section graphmonitor, agraphmonitor +Show various filtergraph stats. + +With this filter one can debug complete filtergraph. +Especially issues with links filling with queued frames. + +The filter accepts the following options: + +@table @option +@item size, s +Set video output size. Default is @var{hd720}. + +@item opacity, o +Set video opacity. Default is @var{0.9}. Allowed range is from @var{0} to @var{1}. + +@item mode, m +Set output mode, can be @var{fulll} or @var{compact}. +In @var{compact} mode only filters with some queued frames have displayed stats. + +@item flags, f +Set flags which enable which stats are shown in video. + +Available values for flags are: +@table @samp +@item queue +Display number of queued frames in each link. + +@item frame_count_in +Display number of frames taken from filter. + +@item frame_count_out +Display number of frames given out from filter. + +@item pts +Display current filtered frame pts. + +@item time +Display current filtered frame time. + +@item timebase +Display time base for filter link. + +@item format +Display used format for filter link. + +@item size +Display video size or number of audio channels in case of audio used by filter link. + +@item rate +Display video frame rate or sample rate in case of audio used by filter link. +@end table + +@item rate, r +Set upper limit for video rate of output stream, Default value is @var{25}. +This guarantee that output video frame rate will not be higher than this value. +@end table + @section greyedge A color constancy variation filter which estimates scene illumination via grey edge algorithm and corrects the scene colors accordingly. diff --git a/libavfilter/Makefile b/libavfilter/Makefile index a98c64b7ce..c35cd8f422 100644 --- a/libavfilter/Makefile +++ b/libavfilter/Makefile @@ -240,6 +240,7 @@ OBJS-$(CONFIG_FSPP_FILTER) += vf_fspp.o OBJS-$(CONFIG_GBLUR_FILTER) += vf_gblur.o OBJS-$(CONFIG_GEQ_FILTER) += vf_geq.o OBJS-$(CONFIG_GRADFUN_FILTER) += vf_gradfun.o +OBJS-$(CONFIG_GRAPHMONITOR_FILTER) += f_graphmonitor.o OBJS-$(CONFIG_GREYEDGE_FILTER) += vf_colorconstancy.o OBJS-$(CONFIG_HALDCLUT_FILTER) += vf_lut3d.o framesync.o OBJS-$(CONFIG_HFLIP_FILTER) += vf_hflip.o @@ -437,6 +438,7 @@ OBJS-$(CONFIG_NULLSINK_FILTER) += vsink_nullsink.o # multimedia filters OBJS-$(CONFIG_ABITSCOPE_FILTER) += avf_abitscope.o OBJS-$(CONFIG_ADRAWGRAPH_FILTER) += f_drawgraph.o +OBJS-$(CONFIG_AGRAPHMONITOR_FILTER) += f_graphmonitor.o OBJS-$(CONFIG_AHISTOGRAM_FILTER) += avf_ahistogram.o OBJS-$(CONFIG_APHASEMETER_FILTER) += avf_aphasemeter.o OBJS-$(CONFIG_AVECTORSCOPE_FILTER) += avf_avectorscope.o diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c index b2cb58fc38..d5a211bda5 100644 --- a/libavfilter/allfilters.c +++ b/libavfilter/allfilters.c @@ -227,6 +227,7 @@ extern AVFilter ff_vf_fspp; extern AVFilter ff_vf_gblur; extern AVFilter ff_vf_geq; extern AVFilter ff_vf_gradfun; +extern AVFilter ff_vf_graphmonitor; extern AVFilter ff_vf_greyedge; extern AVFilter ff_vf_haldclut; extern AVFilter ff_vf_hflip; @@ -418,6 +419,7 @@ extern AVFilter ff_vsink_nullsink; /* multimedia filters */ extern AVFilter ff_avf_abitscope; extern AVFilter ff_avf_adrawgraph; +extern AVFilter ff_avf_agraphmonitor; extern AVFilter ff_avf_ahistogram; extern AVFilter ff_avf_aphasemeter; extern AVFilter ff_avf_avectorscope; diff --git a/libavfilter/f_graphmonitor.c b/libavfilter/f_graphmonitor.c new file mode 100644 index 0000000000..7052c84d9b --- /dev/null +++ b/libavfilter/f_graphmonitor.c @@ -0,0 +1,425 @@ +/* + * Copyright (c) 2018 Paul B Mahol + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "float.h" + +#include "libavutil/pixdesc.h" +#include "libavutil/eval.h" +#include "libavutil/intreadwrite.h" +#include "libavutil/opt.h" +#include "libavutil/timestamp.h" +#include "libavutil/xga_font_data.h" +#include "avfilter.h" +#include "filters.h" +#include "formats.h" +#include "internal.h" +#include "video.h" + +typedef struct GraphMonitorContext { + const AVClass *class; + + int w, h; + float opacity; + int mode; + int flags; + AVRational frame_rate; + + int64_t pts; + uint8_t white[4]; + uint8_t yellow[4]; + uint8_t red[4]; + uint8_t green[4]; + uint8_t bg[4]; +} GraphMonitorContext; + +enum { + MODE_QUEUE = 1 << 0, + MODE_FCIN = 1 << 1, + MODE_FCOUT = 1 << 2, + MODE_PTS = 1 << 3, + MODE_TIME = 1 << 4, + MODE_TB = 1 << 5, + MODE_FMT = 1 << 6, + MODE_SIZE = 1 << 7, + MODE_RATE = 1 << 8, +}; + +#define OFFSET(x) offsetof(GraphMonitorContext, x) +#define VF AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM + +static const AVOption graphmonitor_options[] = { + { "size", "set monitor size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, 0, 0, VF }, + { "s", "set monitor size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, 0, 0, VF }, + { "opacity", "set video opacity", OFFSET(opacity), AV_OPT_TYPE_FLOAT, {.dbl=.9}, 0, 1, VF }, + { "o", "set video opacity", OFFSET(opacity), AV_OPT_TYPE_FLOAT, {.dbl=.9}, 0, 1, VF }, + { "mode", "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, VF, "mode" }, + { "m", "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, VF, "mode" }, + { "full", NULL, 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, VF, "mode" }, + { "compact", NULL, 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, VF, "mode" }, + { "flags", "set flags", OFFSET(flags), AV_OPT_TYPE_FLAGS, {.i64=MODE_QUEUE}, 0, INT_MAX, VF, "flags" }, + { "f", "set flags", OFFSET(flags), AV_OPT_TYPE_FLAGS, {.i64=MODE_QUEUE}, 0, INT_MAX, VF, "flags" }, + { "queue", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_QUEUE}, 0, 0, VF, "flags" }, + { "frame_count_in", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_FCOUT}, 0, 0, VF, "flags" }, + { "frame_count_out", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_FCIN}, 0, 0, VF, "flags" }, + { "pts", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_PTS}, 0, 0, VF, "flags" }, + { "time", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_TIME}, 0, 0, VF, "flags" }, + { "timebase", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_TB}, 0, 0, VF, "flags" }, + { "format", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_FMT}, 0, 0, VF, "flags" }, + { "size", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_SIZE}, 0, 0, VF, "flags" }, + { "rate", NULL, 0, AV_OPT_TYPE_CONST, {.i64=MODE_RATE}, 0, 0, VF, "flags" }, + { "rate", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str = "25"}, 0, INT_MAX, VF }, + { "r", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str = "25"}, 0, INT_MAX, VF }, + { NULL } +}; + +static int query_formats(AVFilterContext *ctx) +{ + AVFilterLink *outlink = ctx->outputs[0]; + static const enum AVPixelFormat pix_fmts[] = { + AV_PIX_FMT_RGBA, + AV_PIX_FMT_NONE + }; + int ret; + + AVFilterFormats *fmts_list = ff_make_format_list(pix_fmts); + if ((ret = ff_formats_ref(fmts_list, &outlink->in_formats)) < 0) + return ret; + + return 0; +} + +static void clear_image(GraphMonitorContext *s, AVFrame *out, AVFilterLink *outlink) +{ + int bg = AV_RN32(s->bg); + + for (int i = 0; i < out->height; i++) + for (int j = 0; j < out->width; j++) + AV_WN32(out->data[0] + i * out->linesize[0] + j * 4, bg); +} + +static void drawtext(AVFrame *pic, int x, int y, const char *txt, uint8_t *color) +{ + const uint8_t *font; + int font_height; + int i; + + font = avpriv_cga_font, font_height = 8; + + if (y + 8 >= pic->height || + x + strlen(txt) * 8 >= pic->width) + return; + + for (i = 0; txt[i]; i++) { + int char_y, mask; + + uint8_t *p = pic->data[0] + y*pic->linesize[0] + (x + i*8)*4; + for (char_y = 0; char_y < font_height; char_y++) { + for (mask = 0x80; mask; mask >>= 1) { + if (font[txt[i] * font_height + char_y] & mask) { + p[0] = color[0]; + p[1] = color[1]; + p[2] = color[2]; + } + p += 4; + } + p += pic->linesize[0] - 8 * 4; + } + } +} + +static int filter_have_queued(AVFilterContext *filter) +{ + for (int j = 0; j < filter->nb_inputs; j++) { + AVFilterLink *l = filter->inputs[j]; + size_t frames = ff_inlink_queued_frames(l); + + if (frames) + return 1; + } + + for (int j = 0; j < filter->nb_outputs; j++) { + AVFilterLink *l = filter->outputs[j]; + size_t frames = ff_inlink_queued_frames(l); + + if (frames) + return 1; + } + + return 0; +} + +static void draw_items(AVFilterContext *ctx, AVFrame *out, + int xpos, int ypos, + AVFilterLink *l, + size_t frames) +{ + GraphMonitorContext *s = ctx->priv; + char buffer[1024] = { 0 }; + + if (s->flags & MODE_FMT) { + if (l->type == AVMEDIA_TYPE_VIDEO) { + snprintf(buffer, sizeof(buffer)-1, " | format: %s", + av_get_pix_fmt_name(l->format)); + } else if (l->type == AVMEDIA_TYPE_AUDIO) { + snprintf(buffer, sizeof(buffer)-1, " | format: %s", + av_get_sample_fmt_name(l->format)); + } + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + } + if (s->flags & MODE_SIZE) { + if (l->type == AVMEDIA_TYPE_VIDEO) { + snprintf(buffer, sizeof(buffer)-1, " | size: %dx%d", l->w, l->h); + } else if (l->type == AVMEDIA_TYPE_AUDIO) { + snprintf(buffer, sizeof(buffer)-1, " | channels: %d", l->channels); + } + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + } + if (s->flags & MODE_RATE) { + if (l->type == AVMEDIA_TYPE_VIDEO) { + snprintf(buffer, sizeof(buffer)-1, " | fps: %d/%d", l->frame_rate.num, l->frame_rate.den); + } else if (l->type == AVMEDIA_TYPE_AUDIO) { + snprintf(buffer, sizeof(buffer)-1, " | samplerate: %d", l->sample_rate); + } + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + } + if (s->flags & MODE_TB) { + snprintf(buffer, sizeof(buffer)-1, " | tb: %d/%d", l->time_base.num, l->time_base.den); + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + } + if (s->flags & MODE_QUEUE) { + snprintf(buffer, sizeof(buffer)-1, " | queue: "); + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + snprintf(buffer, sizeof(buffer)-1, "%"PRId64, frames); + drawtext(out, xpos, ypos, buffer, frames > 0 ? frames >= 10 ? frames >= 50 ? s->red : s->yellow : s->green : s->white); + xpos += strlen(buffer) * 8; + } + if (s->flags & MODE_FCIN) { + snprintf(buffer, sizeof(buffer)-1, " | in: %"PRId64, l->frame_count_in); + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + } + if (s->flags & MODE_FCOUT) { + snprintf(buffer, sizeof(buffer)-1, " | out: %"PRId64, l->frame_count_out); + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + } + if (s->flags & MODE_PTS) { + snprintf(buffer, sizeof(buffer)-1, " | pts: %s", av_ts2str(l->current_pts_us)); + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + } + if (s->flags & MODE_TIME) { + snprintf(buffer, sizeof(buffer)-1, " | time: %s", av_ts2timestr(l->current_pts_us, &AV_TIME_BASE_Q)); + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + } +} + +static int create_frame(AVFilterContext *ctx, int64_t pts) +{ + GraphMonitorContext *s = ctx->priv; + AVFilterLink *outlink = ctx->outputs[0]; + AVFrame *out; + int xpos, ypos = 0; + + out = ff_get_video_buffer(outlink, outlink->w, outlink->h); + if (!out) + return AVERROR(ENOMEM); + + clear_image(s, out, outlink); + + for (int i = 0; i < ctx->graph->nb_filters; i++) { + AVFilterContext *filter = ctx->graph->filters[i]; + char buffer[1024] = { 0 }; + + if (s->mode && !filter_have_queued(filter)) + continue; + + xpos = 0; + drawtext(out, xpos, ypos, filter->name, s->white); + xpos += strlen(filter->name) * 8 + 10; + drawtext(out, xpos, ypos, filter->filter->name, s->white); + ypos += 10; + for (int j = 0; j < filter->nb_inputs; j++) { + AVFilterLink *l = filter->inputs[j]; + size_t frames = ff_inlink_queued_frames(l); + + if (s->mode && !frames) + continue; + + xpos = 10; + snprintf(buffer, sizeof(buffer)-1, "in%d: ", j); + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + drawtext(out, xpos, ypos, l->src->name, s->white); + xpos += strlen(l->src->name) * 8 + 10; + draw_items(ctx, out, xpos, ypos, l, frames); + ypos += 10; + } + + ypos += 2; + for (int j = 0; j < filter->nb_outputs; j++) { + AVFilterLink *l = filter->outputs[j]; + size_t frames = ff_inlink_queued_frames(l); + + if (s->mode && !frames) + continue; + + xpos = 10; + snprintf(buffer, sizeof(buffer)-1, "out%d: ", j); + drawtext(out, xpos, ypos, buffer, s->white); + xpos += strlen(buffer) * 8; + drawtext(out, xpos, ypos, l->dst->name, s->white); + xpos += strlen(l->dst->name) * 8 + 10; + draw_items(ctx, out, xpos, ypos, l, frames); + ypos += 10; + } + ypos += 5; + } + + out->pts = pts; + s->pts = pts; + return ff_filter_frame(outlink, out); +} + +static int activate(AVFilterContext *ctx) +{ + GraphMonitorContext *s = ctx->priv; + AVFilterLink *inlink = ctx->inputs[0]; + AVFilterLink *outlink = ctx->outputs[0]; + int64_t pts = AV_NOPTS_VALUE; + + FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink); + + if (ff_inlink_queued_frames(inlink)) { + AVFrame *frame = NULL; + int ret; + + ret = ff_inlink_consume_frame(inlink, &frame); + if (ret < 0) + return ret; + if (ret > 0) { + pts = frame->pts; + av_frame_free(&frame); + } + } + + if (pts != AV_NOPTS_VALUE) { + pts = av_rescale_q(pts, inlink->time_base, outlink->time_base); + if (s->pts < pts && ff_outlink_frame_wanted(outlink)) + return create_frame(ctx, pts); + } + + FF_FILTER_FORWARD_STATUS(inlink, outlink); + FF_FILTER_FORWARD_WANTED(outlink, inlink); + + return FFERROR_NOT_READY; +} + +static int config_output(AVFilterLink *outlink) +{ + GraphMonitorContext *s = outlink->src->priv; + + s->bg[3] = 255 * s->opacity; + s->white[0] = s->white[1] = s->white[2] = 255; + s->yellow[0] = s->yellow[1] = 255; + s->red[0] = 255; + s->green[1] = 255; + outlink->w = s->w; + outlink->h = s->h; + outlink->sample_aspect_ratio = (AVRational){1,1}; + outlink->frame_rate = s->frame_rate; + outlink->time_base = av_inv_q(s->frame_rate); + + return 0; +} + +#if CONFIG_GRAPHMONITOR_FILTER + +AVFILTER_DEFINE_CLASS(graphmonitor); + +static const AVFilterPad graphmonitor_inputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_VIDEO, + }, + { NULL } +}; + +static const AVFilterPad graphmonitor_outputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_VIDEO, + .config_props = config_output, + }, + { NULL } +}; + +AVFilter ff_vf_graphmonitor = { + .name = "graphmonitor", + .description = NULL_IF_CONFIG_SMALL("Show various filtergraph stats."), + .priv_size = sizeof(GraphMonitorContext), + .priv_class = &graphmonitor_class, + .query_formats = query_formats, + .activate = activate, + .inputs = graphmonitor_inputs, + .outputs = graphmonitor_outputs, +}; + +#endif // CONFIG_GRAPHMONITOR_FILTER + +#if CONFIG_AGRAPHMONITOR_FILTER + +#define agraphmonitor_options graphmonitor_options +AVFILTER_DEFINE_CLASS(agraphmonitor); + +static const AVFilterPad agraphmonitor_inputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_AUDIO, + }, + { NULL } +}; + +static const AVFilterPad agraphmonitor_outputs[] = { + { + .name = "default", + .type = AVMEDIA_TYPE_VIDEO, + .config_props = config_output, + }, + { NULL } +}; + +AVFilter ff_avf_agraphmonitor = { + .name = "agraphmonitor", + .description = NULL_IF_CONFIG_SMALL("Show various filtergraph stats."), + .priv_size = sizeof(GraphMonitorContext), + .priv_class = &agraphmonitor_class, + .query_formats = query_formats, + .activate = activate, + .inputs = agraphmonitor_inputs, + .outputs = agraphmonitor_outputs, +}; +#endif // CONFIG_AGRAPHMONITOR_FILTER diff --git a/libavfilter/version.h b/libavfilter/version.h index 77e1a77b50..e2572d623e 100644 --- a/libavfilter/version.h +++ b/libavfilter/version.h @@ -30,7 +30,7 @@ #include "libavutil/version.h" #define LIBAVFILTER_VERSION_MAJOR 7 -#define LIBAVFILTER_VERSION_MINOR 38 +#define LIBAVFILTER_VERSION_MINOR 39 #define LIBAVFILTER_VERSION_MICRO 100 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \

1 0

avcodec/vdpau: Enable HEVC support for working Nvidia driver versions
by ManojGuptaBonda 31 Oct '18

31 Oct '18

ffmpeg | branch: master | ManojGuptaBonda <mbonda(a)nvidia.com> | Mon Oct 29 13:57:54 2018 +0530| [4a6d5f3cadaabefe6c3548e575bb7e713997762f] | committer: Philip Langdale avcodec/vdpau: Enable HEVC support for working Nvidia driver versions The driver bugs that caused decoded HEVC content to have an incorrect memory layout have been fully fixed in the 410.xx driver release so we can start exposing support. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=4a6d5f3cadaabefe6… --- libavcodec/vdpau.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/libavcodec/vdpau.c b/libavcodec/vdpau.c index 1b2ec989cd..59454e8c4b 100644 --- a/libavcodec/vdpau.c +++ b/libavcodec/vdpau.c @@ -208,8 +208,12 @@ int ff_vdpau_common_init(AVCodecContext *avctx, VdpDecoderProfile profile, return vdpau_error(status); if (avctx->codec_id == AV_CODEC_ID_HEVC && strncmp(info_string, "NVIDIA ", 7) == 0 && !(avctx->hwaccel_flags & AV_HWACCEL_FLAG_ALLOW_PROFILE_MISMATCH)) { - av_log(avctx, AV_LOG_VERBOSE, "HEVC with NVIDIA VDPAU drivers is buggy, skipping.\n"); - return AVERROR(ENOTSUP); + int driver_version; + sscanf(info_string, "NVIDIA VDPAU Driver Shared Library %d", &driver_version); + if (driver_version < 410) { + av_log(avctx, AV_LOG_VERBOSE, "HEVC with NVIDIA VDPAU drivers is buggy, skipping.\n"); + return AVERROR(ENOTSUP); + } } status = vdctx->get_proc_address(vdctx->device,

1 0

avcodec/cbs_vp9: store profile in the private context
by James Almer 31 Oct '18

31 Oct '18

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Tue Oct 30 15:41:21 2018 -0300| [7aaf092f8ce8708bc745e0c27746d3312a2421d5] | committer: James Almer avcodec/cbs_vp9: store profile in the private context Derived from profile_low_bit and profile_high_bit. Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=7aaf092f8ce8708bc… --- libavcodec/cbs_vp9.h | 2 ++ libavcodec/cbs_vp9_syntax_template.c | 13 ++++++------- 2 files changed, 8 insertions(+), 7 deletions(-) diff --git a/libavcodec/cbs_vp9.h b/libavcodec/cbs_vp9.h index dac8171c4a..4c9b2f880d 100644 --- a/libavcodec/cbs_vp9.h +++ b/libavcodec/cbs_vp9.h @@ -190,6 +190,8 @@ typedef struct VP9ReferenceFrameState { } VP9ReferenceFrameState; typedef struct CodedBitstreamVP9Context { + int profile; + // Frame dimensions in 8x8 mode info blocks. uint16_t mi_cols; uint16_t mi_rows; diff --git a/libavcodec/cbs_vp9_syntax_template.c b/libavcodec/cbs_vp9_syntax_template.c index bbf104c8be..898cede329 100644 --- a/libavcodec/cbs_vp9_syntax_template.c +++ b/libavcodec/cbs_vp9_syntax_template.c @@ -278,15 +278,14 @@ static int FUNC(uncompressed_header)(CodedBitstreamContext *ctx, RWContext *rw, VP9RawFrameHeader *current) { CodedBitstreamVP9Context *vp9 = ctx->priv_data; - int profile, i; - int err; + int err, i; f(2, frame_marker); f(1, profile_low_bit); f(1, profile_high_bit); - profile = (current->profile_high_bit << 1) + current->profile_low_bit; - if (profile == 3) + vp9->profile = (current->profile_high_bit << 1) + current->profile_low_bit; + if (vp9->profile == 3) fixed(1, reserved_zero, 0); f(1, show_existing_frame); @@ -304,7 +303,7 @@ static int FUNC(uncompressed_header)(CodedBitstreamContext *ctx, RWContext *rw, if (current->frame_type == VP9_KEY_FRAME) { CHECK(FUNC(frame_sync_code)(ctx, rw, current)); - CHECK(FUNC(color_config)(ctx, rw, current, profile)); + CHECK(FUNC(color_config)(ctx, rw, current, vp9->profile)); CHECK(FUNC(frame_size)(ctx, rw, current)); CHECK(FUNC(render_size)(ctx, rw, current)); @@ -324,8 +323,8 @@ static int FUNC(uncompressed_header)(CodedBitstreamContext *ctx, RWContext *rw, if (current->intra_only == 1) { CHECK(FUNC(frame_sync_code)(ctx, rw, current)); - if (profile > 0) { - CHECK(FUNC(color_config)(ctx, rw, current, profile)); + if (vp9->profile > 0) { + CHECK(FUNC(color_config)(ctx, rw, current, vp9->profile)); } else { infer(color_space, 1); infer(subsampling_x, 1);

1 0

avcodec/cbs_vp9: keep track of reference frames
by James Almer 31 Oct '18

31 Oct '18

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Tue Oct 30 13:44:55 2018 -0300| [ec440339b2cd2b1e9dd306fe69cfcff5d75dcd28] | committer: James Almer avcodec/cbs_vp9: keep track of reference frames Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=ec440339b2cd2b1e9… --- libavcodec/cbs_vp9.h | 16 ++++++++++ libavcodec/cbs_vp9_syntax_template.c | 60 +++++++++++++++++++++++++++++++++--- 2 files changed, 72 insertions(+), 4 deletions(-) diff --git a/libavcodec/cbs_vp9.h b/libavcodec/cbs_vp9.h index b9cb422fc6..dac8171c4a 100644 --- a/libavcodec/cbs_vp9.h +++ b/libavcodec/cbs_vp9.h @@ -181,6 +181,13 @@ typedef struct VP9RawSuperframe { VP9RawSuperframeIndex index; } VP9RawSuperframe; +typedef struct VP9ReferenceFrameState { + int frame_width; // RefFrameWidth + int frame_height; // RefFrameHeight + int subsampling_x; // RefSubsamplingX + int subsampling_y; // RefSubsamplingY + int bit_depth; // RefBitDepth +} VP9ReferenceFrameState; typedef struct CodedBitstreamVP9Context { // Frame dimensions in 8x8 mode info blocks. @@ -190,6 +197,15 @@ typedef struct CodedBitstreamVP9Context { uint16_t sb64_cols; uint16_t sb64_rows; + int frame_width; + int frame_height; + + uint8_t subsampling_x; + uint8_t subsampling_y; + int bit_depth; + + VP9ReferenceFrameState ref[VP9_NUM_REF_FRAMES]; + // Write buffer. uint8_t *write_buffer; size_t write_buffer_size; diff --git a/libavcodec/cbs_vp9_syntax_template.c b/libavcodec/cbs_vp9_syntax_template.c index 5055aa38e1..bbf104c8be 100644 --- a/libavcodec/cbs_vp9_syntax_template.c +++ b/libavcodec/cbs_vp9_syntax_template.c @@ -43,10 +43,14 @@ static int FUNC(frame_sync_code)(CodedBitstreamContext *ctx, RWContext *rw, static int FUNC(color_config)(CodedBitstreamContext *ctx, RWContext *rw, VP9RawFrameHeader *current, int profile) { + CodedBitstreamVP9Context *vp9 = ctx->priv_data; int err; - if (profile >= 2) + if (profile >= 2) { f(1, ten_or_twelve_bit); + vp9->bit_depth = current->ten_or_twelve_bit ? 12 : 10; + } else + vp9->bit_depth = 8; f(3, color_space); @@ -69,6 +73,9 @@ static int FUNC(color_config)(CodedBitstreamContext *ctx, RWContext *rw, } } + vp9->subsampling_x = current->subsampling_x; + vp9->subsampling_y = current->subsampling_y; + return 0; } @@ -81,8 +88,11 @@ static int FUNC(frame_size)(CodedBitstreamContext *ctx, RWContext *rw, f(16, frame_width_minus_1); f(16, frame_height_minus_1); - vp9->mi_cols = (current->frame_width_minus_1 + 8) >> 3; - vp9->mi_rows = (current->frame_height_minus_1 + 8) >> 3; + vp9->frame_width = current->frame_width_minus_1 + 1; + vp9->frame_height = current->frame_height_minus_1 + 1; + + vp9->mi_cols = (vp9->frame_width + 7) >> 3; + vp9->mi_rows = (vp9->frame_height + 7) >> 3; vp9->sb64_cols = (vp9->mi_cols + 7) >> 3; vp9->sb64_rows = (vp9->mi_rows + 7) >> 3; @@ -107,15 +117,33 @@ static int FUNC(render_size)(CodedBitstreamContext *ctx, RWContext *rw, static int FUNC(frame_size_with_refs)(CodedBitstreamContext *ctx, RWContext *rw, VP9RawFrameHeader *current) { + CodedBitstreamVP9Context *vp9 = ctx->priv_data; int err, i; for (i = 0; i < VP9_REFS_PER_FRAME; i++) { fs(1, found_ref[i], 1, i); - if (current->found_ref[i]) + if (current->found_ref[i]) { + VP9ReferenceFrameState *ref = + &vp9->ref[current->ref_frame_idx[i]]; + + vp9->frame_width = ref->frame_width; + vp9->frame_height = ref->frame_height; + + vp9->subsampling_x = ref->subsampling_x; + vp9->subsampling_y = ref->subsampling_y; + vp9->bit_depth = ref->bit_depth; + break; + } } if (i >= VP9_REFS_PER_FRAME) CHECK(FUNC(frame_size)(ctx, rw, current)); + else { + vp9->mi_cols = (vp9->frame_width + 7) >> 3; + vp9->mi_rows = (vp9->frame_height + 7) >> 3; + vp9->sb64_cols = (vp9->mi_cols + 7) >> 3; + vp9->sb64_rows = (vp9->mi_rows + 7) >> 3; + } CHECK(FUNC(render_size)(ctx, rw, current)); return 0; @@ -249,6 +277,7 @@ static int FUNC(tile_info)(CodedBitstreamContext *ctx, RWContext *rw, static int FUNC(uncompressed_header)(CodedBitstreamContext *ctx, RWContext *rw, VP9RawFrameHeader *current) { + CodedBitstreamVP9Context *vp9 = ctx->priv_data; int profile, i; int err; @@ -301,6 +330,10 @@ static int FUNC(uncompressed_header)(CodedBitstreamContext *ctx, RWContext *rw, infer(color_space, 1); infer(subsampling_x, 1); infer(subsampling_y, 1); + vp9->bit_depth = 8; + + vp9->subsampling_x = current->subsampling_x; + vp9->subsampling_y = current->subsampling_y; } f(8, refresh_frame_flags); @@ -339,6 +372,25 @@ static int FUNC(uncompressed_header)(CodedBitstreamContext *ctx, RWContext *rw, f(16, header_size_in_bytes); + for (i = 0; i < VP9_NUM_REF_FRAMES; i++) { + if (current->refresh_frame_flags & (1 << i)) { + vp9->ref[i] = (VP9ReferenceFrameState) { + .frame_width = vp9->frame_width, + .frame_height = vp9->frame_height, + .subsampling_x = vp9->subsampling_x, + .subsampling_y = vp9->subsampling_y, + .bit_depth = vp9->bit_depth, + }; + } + } + + av_log(ctx->log_ctx, AV_LOG_DEBUG, "Frame: size %dx%d " + "subsample %dx%d bit_depth %d tiles %dx%d.\n", + vp9->frame_width, vp9->frame_height, + vp9->subsampling_x, vp9->subsampling_y, + vp9->bit_depth, 1 << current->tile_cols_log2, + 1 << current->tile_rows_log2); + return 0; }

1 0

cbs_vp9: Ensure that reserved zero bits are actually zero
by Mark Thompson 30 Oct '18

30 Oct '18

ffmpeg | branch: master | Mark Thompson <sw(a)jkqxz.net> | Sat Oct 27 20:41:00 2018 +0100| [edcdf3512376b64d6add61fb5c21b418ebbba1e3] | committer: Mark Thompson cbs_vp9: Ensure that reserved zero bits are actually zero > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=edcdf3512376b64d6… --- libavcodec/cbs_vp9.c | 13 +++++++++++++ libavcodec/cbs_vp9.h | 2 -- libavcodec/cbs_vp9_syntax_template.c | 6 +++--- 3 files changed, 16 insertions(+), 5 deletions(-) diff --git a/libavcodec/cbs_vp9.c b/libavcodec/cbs_vp9.c index 7498be4b73..c03ce986c0 100644 --- a/libavcodec/cbs_vp9.c +++ b/libavcodec/cbs_vp9.c @@ -314,6 +314,12 @@ static int cbs_vp9_write_le(CodedBitstreamContext *ctx, PutBitContext *pbc, current->name = prob; \ } while (0) +#define fixed(width, name, value) do { \ + av_unused uint32_t fixed_value = value; \ + CHECK(ff_cbs_read_unsigned(ctx, rw, width, #name, \ + 0, &fixed_value, value, value)); \ + } while (0) + #define infer(name, value) do { \ current->name = value; \ } while (0) @@ -331,6 +337,7 @@ static int cbs_vp9_write_le(CodedBitstreamContext *ctx, PutBitContext *pbc, #undef fle #undef delta_q #undef prob +#undef fixed #undef infer #undef byte_alignment @@ -370,6 +377,11 @@ static int cbs_vp9_write_le(CodedBitstreamContext *ctx, PutBitContext *pbc, xf(8, name.prob, current->name, subs, __VA_ARGS__); \ } while (0) +#define fixed(width, name, value) do { \ + CHECK(ff_cbs_write_unsigned(ctx, rw, width, #name, \ + 0, value, value, value)); \ + } while (0) + #define infer(name, value) do { \ if (current->name != (value)) { \ av_log(ctx->log_ctx, AV_LOG_WARNING, "Warning: " \ @@ -392,6 +404,7 @@ static int cbs_vp9_write_le(CodedBitstreamContext *ctx, PutBitContext *pbc, #undef fle #undef delta_q #undef prob +#undef fixed #undef infer #undef byte_alignment diff --git a/libavcodec/cbs_vp9.h b/libavcodec/cbs_vp9.h index 5b99c90c2e..b9cb422fc6 100644 --- a/libavcodec/cbs_vp9.h +++ b/libavcodec/cbs_vp9.h @@ -84,7 +84,6 @@ typedef struct VP9RawFrameHeader { uint8_t frame_marker; uint8_t profile_low_bit; uint8_t profile_high_bit; - uint8_t profile_reserved_zero; uint8_t show_existing_frame; uint8_t frame_to_show_map_idx; @@ -99,7 +98,6 @@ typedef struct VP9RawFrameHeader { uint8_t color_range; uint8_t subsampling_x; uint8_t subsampling_y; - uint8_t color_config_reserved_zero; uint8_t refresh_frame_flags; diff --git a/libavcodec/cbs_vp9_syntax_template.c b/libavcodec/cbs_vp9_syntax_template.c index b4a7f65e85..5055aa38e1 100644 --- a/libavcodec/cbs_vp9_syntax_template.c +++ b/libavcodec/cbs_vp9_syntax_template.c @@ -55,7 +55,7 @@ static int FUNC(color_config)(CodedBitstreamContext *ctx, RWContext *rw, if (profile == 1 || profile == 3) { f(1, subsampling_x); f(1, subsampling_y); - f(1, color_config_reserved_zero); + fixed(1, reserved_zero, 0); } else { infer(subsampling_x, 1); infer(subsampling_y, 1); @@ -65,7 +65,7 @@ static int FUNC(color_config)(CodedBitstreamContext *ctx, RWContext *rw, if (profile == 1 || profile == 3) { infer(subsampling_x, 0); infer(subsampling_y, 0); - f(1, color_config_reserved_zero); + fixed(1, reserved_zero, 0); } } @@ -258,7 +258,7 @@ static int FUNC(uncompressed_header)(CodedBitstreamContext *ctx, RWContext *rw, f(1, profile_high_bit); profile = (current->profile_high_bit << 1) + current->profile_low_bit; if (profile == 3) - f(1, profile_reserved_zero); + fixed(1, reserved_zero, 0); f(1, show_existing_frame); if (current->show_existing_frame) {

1 0

doc/hls: fix grammar for HLS options
by Werner Robitza 30 Oct '18

30 Oct '18

ffmpeg | branch: master | Werner Robitza <werner.robitza(a)gmail.com> | Tue Oct 30 20:22:55 2018 +0100| [ad5ca1fb72fc1e0d7cdeec85b279608f8195c96e] | committer: Lou Logan doc/hls: fix grammar for HLS options This fixes the grammar of two HLS option descriptions and makes them less ambiguous. Signed-off-by: Werner Robitza <werner.robitza(a)gmail.com> Signed-off-by: Lou Logan <lou(a)lrcd.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=ad5ca1fb72fc1e0d7… --- doc/muxers.texi | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/doc/muxers.texi b/doc/muxers.texi index f18543e83d..62f4091e31 100644 --- a/doc/muxers.texi +++ b/doc/muxers.texi @@ -764,17 +764,17 @@ Possible values: @table @samp @item mpegts -If this flag is set, the hls segment files will format to mpegts. -the mpegts files is used in all hls versions. +Output segment files in MPEG-2 Transport Stream format. This is +compatible with all HLS versions. @item fmp4 -If this flag is set, the hls segment files will format to fragment mp4 looks like dash. -the fmp4 files is used in hls after version 7. +Output segment files in fragmented MP4 format, similar to MPEG-DASH. +fmp4 files may be used in HLS version 7 and above. @end table @item hls_fmp4_init_filename @var{filename} -set filename to the fragment files header file, default filename is @file{init.mp4}. +Set filename to the fragment files header file, default filename is @file{init.mp4}. When @code{var_stream_map} is set with two or more variant streams, the @var{filename} pattern must contain the string "%v", this string specifies

1 0