transcode_step()在转码过程中对pts、dts、duration的处理

简介

对pts、dts、duration的处理主要集中在两大函数里面

1、process_input()读入数据并处理，放到滤镜里面

2、reap_filters()从滤镜读出数据，处理后写入文件

process_input()中对时间的处理

获取数据包

ret = get_input_packet(ifile, &pkt);

通过这条语句实际里面调用

return av_read_frame(f->ctx, pkt);

将读取到的帧放到pkt中。

处理获取的pkt中的时间

//如果pkt.dts或pkt.pts有值，而不是AV_NOPTS_VALUE时，做处理。
/*
ifile->ts_offset在open_input_file函数中有处理，涉及到参数"itsoffset"(o->input_ts_offset)、
"copyts"(copy_ts)、"start_at_zero"(start_at_zero )、"ss"(ic->start_time)
语句如下。
f->ts_offset  = o->input_ts_offset - (copy_ts ? (start_at_zero && ic->start_time != AV_NOPTS_VALUE ? ic->start_time : 0) : timestamp);
如果没有这些设置。f->ts_offset=0
*/
if (pkt.dts != AV_NOPTS_VALUE)
pkt.dts += av_rescale_q(ifile->ts_offset, AV_TIME_BASE_Q, ist->st->time_base);
if (pkt.pts != AV_NOPTS_VALUE)
pkt.pts += av_rescale_q(ifile->ts_offset, AV_TIME_BASE_Q, ist->st->time_base);

//ist->ts_scale在add_input_streams函数中设置
/*
ist->ts_scale = 1.0;//默认是1
MATCH_PER_STREAM_OPT(ts_scale, dbl, ist->ts_scale, ic, st);
涉及到的参数是"itsscale"：解释为"set the input ts scale"
貌似是单位
*/
if (pkt.pts != AV_NOPTS_VALUE)
pkt.pts *= ist->ts_scale;
if (pkt.dts != AV_NOPTS_VALUE)
pkt.dts *= ist->ts_scale;

//做时间基的转换
pkt_dts = av_rescale_q_rnd(pkt.dts, ist->st->time_base, AV_TIME_BASE_Q, (enum AVRounding)(AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX));
if ((ist->dec_ctx->codec_type == AVMEDIA_TYPE_VIDEO ||
ist->dec_ctx->codec_type == AVMEDIA_TYPE_AUDIO) &&
pkt_dts != AV_NOPTS_VALUE && ist->next_dts == AV_NOPTS_VALUE && !copy_ts
&& (is->iformat->flags & AVFMT_TS_DISCONT) && ifile->last_ts != AV_NOPTS_VALUE) {
int64_t delta   = pkt_dts - ifile->last_ts;
if (delta < -1LL*dts_delta_threshold*AV_TIME_BASE ||
delta >  1LL*dts_delta_threshold*AV_TIME_BASE){
ifile->ts_offset -= delta;
av_log(NULL, AV_LOG_DEBUG,
"Inter stream timestamp discontinuity %"PRId64", new offset= %"PRId64"\n",
delta, ifile->ts_offset);
pkt.dts -= av_rescale_q(delta, AV_TIME_BASE_Q, ist->st->time_base);
if (pkt.pts != AV_NOPTS_VALUE)
pkt.pts -= av_rescale_q(delta, AV_TIME_BASE_Q, ist->st->time_base);
}
}

duration = av_rescale_q(ifile->duration, ifile->time_base, ist->st->time_base);
if (pkt.pts != AV_NOPTS_VALUE) {
pkt.pts += duration;
ist->max_pts = FFMAX(pkt.pts, ist->max_pts);
ist->min_pts = FFMIN(pkt.pts, ist->min_pts);
}

if (pkt.dts != AV_NOPTS_VALUE)
pkt.dts += duration;

pkt_dts = av_rescale_q_rnd(pkt.dts, ist->st->time_base, AV_TIME_BASE_Q, (enum AVRounding)(AV_ROUND_NEAR_INF|AV_ROUND_PASS_MINMAX));
if ((ist->dec_ctx->codec_type == AVMEDIA_TYPE_VIDEO ||
ist->dec_ctx->codec_type == AVMEDIA_TYPE_AUDIO) &&
pkt_dts != AV_NOPTS_VALUE && ist->next_dts != AV_NOPTS_VALUE &&
!copy_ts) {
int64_t delta   = pkt_dts - ist->next_dts;
if (is->iformat->flags & AVFMT_TS_DISCONT) {
if (delta < -1LL*dts_delta_threshold*AV_TIME_BASE ||
delta >  1LL*dts_delta_threshold*AV_TIME_BASE ||
pkt_dts + AV_TIME_BASE/10 < FFMAX(ist->pts, ist->dts)) {
ifile->ts_offset -= delta;
av_log(NULL, AV_LOG_DEBUG,
"timestamp discontinuity %"PRId64", new offset= %"PRId64"\n",
delta, ifile->ts_offset);
pkt.dts -= av_rescale_q(delta, AV_TIME_BASE_Q, ist->st->time_base);
if (pkt.pts != AV_NOPTS_VALUE)
pkt.pts -= av_rescale_q(delta, AV_TIME_BASE_Q, ist->st->time_base);
}
} else {
if ( delta < -1LL*dts_error_threshold*AV_TIME_BASE ||
delta >  1LL*dts_error_threshold*AV_TIME_BASE) {
av_log(NULL, AV_LOG_WARNING, "DTS %"PRId64", next:%"PRId64" st:%d invalid dropping\n", pkt.dts, ist->next_dts, pkt.stream_index);
pkt.dts = AV_NOPTS_VALUE;
}
if (pkt.pts != AV_NOPTS_VALUE){
int64_t pkt_pts = av_rescale_q(pkt.pts, ist->st->time_base, AV_TIME_BASE_Q);
delta   = pkt_pts - ist->next_dts;
if ( delta < -1LL*dts_error_threshold*AV_TIME_BASE ||
delta >  1LL*dts_error_threshold*AV_TIME_BASE) {
av_log(NULL, AV_LOG_WARNING, "PTS %"PRId64", next:%"PRId64" invalid dropping st:%d\n", pkt.pts, ist->next_dts, pkt.stream_index);
pkt.pts = AV_NOPTS_VALUE;
}
}
}
}

if (pkt.dts != AV_NOPTS_VALUE)
ifile->last_ts = av_rescale_q(pkt.dts, ist->st->time_base, AV_TIME_BASE_Q);

如果pkt.dts == AV_NOPTS_VALUE 并且pkt.dts == AV_NOPTS_VALUE，以上这些语句都没有起作用。

之后调用如下语句

process_input_packet(ist, &pkt, 0);

process_input_packet

/* pkt = NULL means EOF (needed to flush decoder buffers) */
static int process_input_packet(InputStream *ist, const AVPacket *pkt, int no_eof)
{
int ret = 0, i;
int got_output = 0;

AVPacket avpkt;
//如果是文件的第一次读取，就处理
if (!ist->saw_first_ts) {
//有B帧的情况下dts和pts是不同的。
ist->dts = ist->st->avg_frame_rate.num ? - ist->dec_ctx->has_b_frames * AV_TIME_BASE / av_q2d(ist->st->avg_frame_rate) : 0;
ist->pts = 0;
if (pkt && pkt->pts != AV_NOPTS_VALUE && !ist->decoding_needed) {
ist->dts += av_rescale_q(pkt->pts, ist->st->time
4000
_base, AV_TIME_BASE_Q);
ist->pts = ist->dts; //unused but better to set it to a value thats not totally wrong
}
ist->saw_first_ts = 1;
}
//此时在ist中dts和pts都是有值的，第一次进来是初始值，之后进来的值
//初始化next_dts和next_pts
if (ist->next_dts == AV_NOPTS_VALUE)
ist->next_dts = ist->dts;
if (ist->next_pts == AV_NOPTS_VALUE)
ist->next_pts = ist->pts;
/*如果pkt为NULL，表示流结束了。给avpkt赋值为此后在调用decode_video()->avcodec_decode_video2()时会将剩余的帧都返回。
在process_input()函数的出错处理里面有调用process_input_packet(ist, NULL, 0);就在此处处理
*/
if (!pkt) {
/* EOF handling */
av_init_packet(&avpkt);
avpkt.data = NULL;
avpkt.size = 0;
goto handle_eof;
} else {
avpkt = *pkt;
}

if (pkt->dts != AV_NOPTS_VALUE) {
ist->next_dts = ist->dts = av_rescale_q(pkt->dts, ist->st->time_base, AV_TIME_BASE_Q);
if (ist->dec_ctx->codec_type != AVMEDIA_TYPE_VIDEO || !ist->decoding_needed)
ist->next_pts = ist->pts = ist->dts;
}

// while we have more to decode or while the decoder did output something on EOF
/*pkt里为视频是为一帧，为音频是可能为多帧。所以要一个循环*/
while (ist->decoding_needed && (avpkt.size > 0 || (!pkt && got_output))) {
int duration;
handle_eof:
//
ist->pts = ist->next_pts;
ist->dts = ist->next_dts;

//没什么用。
if (avpkt.size && avpkt.size != pkt->size &&
!(ist->dec->capabilities & AV_CODEC_CAP_SUBFRAMES)) {
av_log(NULL, ist->showed_multi_packet_warning ? AV_LOG_VERBOSE : AV_LOG_WARNING,
"Multiple frames in a packet from stream %d\n", pkt->stream_index);
ist->showed_multi_packet_warning = 1;
}
//解码一个pkt
switch (ist->dec_ctx->codec_type) {
case AVMEDIA_TYPE_AUDIO:
ret = decode_audio    (ist, &avpkt, &got_output);
break;
case AVMEDIA_TYPE_VIDEO:
ret = decode_video    (ist, &avpkt, &got_output);
if (avpkt.duration) {
//返回的pkt中已经计算出了时长
duration = av_rescale_q(avpkt.duration, ist->st->time_base, AV_TIME_BASE_Q);
} else if(ist->dec_ctx->framerate.num != 0 && ist->dec_ctx->framerate.den != 0) {
//根据帧率计算时长
int ticks= av_stream_get_parser(ist->st) ? av_stream_get_parser(ist->st)->repeat_pict+1 : ist->dec_ctx->ticks_per_frame;
duration = ((int64_t)AV_TIME_BASE *
ist->dec_ctx->framerate.den * ticks) /
ist->dec_ctx->framerate.num / ist->dec_ctx->ticks_per_frame;
} else
duration = 0;
//dts和pts加时长，就是下一个帧的dts和pts，注意pkt中的都是临时的。只是在解码的时候用，过程当中的时间都保持在ist中。
if(ist->dts != AV_NOPTS_VALUE && duration) {
ist->next_dts += duration;
}else
ist->next_dts = AV_NOPTS_VALUE;

if (got_output)
ist->next_pts += duration; //FIXME the duration is not correct in some cases
break;
case AVMEDIA_TYPE_SUBTITLE:
ret = transcode_subtitles(ist, &avpkt, &got_output);
break;
default:
return -1;
}

if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Error while decoding stream #%d:%d: %s\n",
ist->file_index, ist->st->index, av_err2str(ret));
if (exit_on_error)
exit_program(1);
break;
}

avpkt.dts=
avpkt.pts= AV_NOPTS_VALUE;

// touch data and size only if not EOF
if (pkt) {
/*如果是视频帧，avpkt.size==0，ret==0，循环结束，
如果是音频帧，avpkt可能包含多帧，avpkt.size处理完可能不为0，
可能需要循环几次*/
if(ist->dec_ctx->codec_type != AVMEDIA_TYPE_AUDIO)
ret = avpkt.size;
avpkt.data += ret;
avpkt.size -= ret;
}
if (!got_output) {
continue;
}
if (got_output && !pkt)
break;
}
//pkt 传入是NULL时，给滤镜传入EOF信息
/* after flushing, send an EOF on all the filter inputs attached to the stream */
/* except when looping we need to flush but not to send an EOF */
if (!pkt && ist->decoding_needed && !got_output && !no_eof) {
int ret = send_filter_eof(ist);
if (ret < 0) {
av_log(NULL, AV_LOG_FATAL, "Error marking filters as finished\n");
exit_program(1);
}
}
/*如果流不需要重新编解码的情况，这里处理dts和next_dts*/
/* handle stream copy */
if (!ist->decoding_needed) {
ist->dts = ist->next_dts;
switch (ist->dec_ctx->codec_type) {
case AVMEDIA_TYPE_AUDIO:
ist->next_dts += ((int64_t)AV_TIME_BASE * ist->dec_ctx->frame_size) /
ist->dec_ctx->sample_rate;
break;
case AVMEDIA_TYPE_VIDEO:
//优先按帧率计算
if (ist->framerate.num) {
// TODO: Remove work-around for c99-to-c89 issue 7
AVRational time_base_q = AV_TIME_BASE_Q;
//ist中的时间基是ffmpeg的内部时间基time_base_q，将next_dts转换为流本身的时间基
int64_t next_dts = av_rescale_q(ist->next_dts, time_base_q, av_inv_q(ist->framerate));
//next_dts为以帧率为时间基，+1就为下一帧的时间
ist->next_dts = av_rescale_q(next_dts + 1, av_inv_q(ist->framerate), time_base_q);
} else if (pkt->duration) {
//pkt->duration的时间基为流本身的ist->st->time_base，将它进行转换，并加到ist->next_dts，
//含义是现在的dts加上此帧经过的时长，表示下一帧的dts
ist->next_dts += av_rescale_q(pkt->duration, ist->st->time_base, AV_TIME_BASE_Q);
} else if(ist->dec_ctx->framerate.num != 0) {
/*
AVCodecParserContext中repeat_pict的注释如下：
* This field is used for proper frame duration computation in lavf.
* It signals, how much longer the frame duration of the current frame
* is compared to normal frame duration.
*
* frame_duration = (1 + repeat_pict) * time_base
*
* It is used by codecs like H.264 to display telecined material.
* -------------------------------------------
* AVCodecContext中ticks_per_frame的注释如下：
* For some codecs, the time base is closer to the field rate than the frame rate.
* Most notably, H.264 and MPEG-2 specify time_base as half of frame duration
* if no telecine is used ...
*
* Set to time_base ticks per frame. Default 1, e.g., H.264/MPEG-2 set it to 2.
*/
/*这里之后再分析吧。大概也是根据ist->dec_ctx->framerate算出一帧的时间加到next_dts上就是下一帧开始时间*/
int ticks= av_stream_get_parser(ist->st) ? av_stream_get_parser(ist->st)->repeat_pict + 1 : ist->dec_ctx->ticks_per_frame;
ist->next_dts += ((int64_t)AV_TIME_BASE *
ist->dec_ctx->framerate.den * ticks) /
ist->dec_ctx->framerate.num / ist->dec_ctx->ticks_per_frame;
}
break;
}
ist->pts = ist->dts;
ist->next_pts = ist->next_dts;
}
for (i = 0; pkt && i < nb_output_streams; i++) {
OutputStream *ost = output_streams[i];

if (!check_output_constraints(ist, ost) || ost->encoding_needed)
continue;
//如果不需要重新编解码，就走到这里。ost->encoding_needed判断。
do_streamcopy(ist, ost, pkt);
}

return got_output;
}

decode_video

针对时间的处理只有一句

pkt->dts  = av_rescale_q(ist->dts, AV_TIME_BASE_Q, ist->st->time_base);

将在ist中保存的dts给pkt中

static int decode_video(InputStream *ist, AVPacket *pkt, int *got_output)
{
AVFrame *decoded_frame, *f;
int i, ret = 0, err = 0, resample_changed;
int64_t best_effort_timestamp;
AVRational *frame_sample_aspect;

if (!ist->decoded_frame && !(ist->decoded_frame = av_frame_alloc()))
return AVERROR(ENOMEM);
if (!ist->filter_frame && !(ist->filter_frame = av_frame_alloc()))
return AVERROR(ENOMEM);
//解码后的帧存放处
decoded_frame = ist->decoded_frame;
//开始给pkt中的时间赋值。只对dts赋值。
pkt->dts  = av_rescale_q(ist->dts, AV_TIME_BASE_Q, ist->st->time_base);

update_benchmark(NULL);
//实际的解码函数，got_output不为0表示解码出了帧，解码出的帧放到decoded_frame中。
ret = avcodec_decode_video2(ist->dec_ctx,
decoded_frame, got_output, pkt);
update_benchmark("decode_video %d.%d", ist->file_index, ist->st->index);

// The following line may be required in some cases where there is no parser
// or the parser does not has_b_frames correctly
if (ist->st->codec->has_b_frames < ist->dec_ctx->has_b_frames) {
if (ist->dec_ctx->codec_id == AV_CODEC_ID_H264) {
ist->st->codec->has_b_frames = ist->dec_ctx->has_b_frames;
} else
av_log(ist->dec_ctx, AV_LOG_WARNING,
"has_b_frames is larger in decoder than demuxer %d > %d.\n"
"If you want to help, upload a sample "
"of this file to ftp://upload.ffmpeg.org/incoming/ "
"and contact the ffmpeg-devel mailing list. (ffmpeg-devel@ffmpeg.org)",
ist->dec_ctx->has_b_frames,
ist->st->codec->has_b_frames);
}
//判断输出结果，涉及到参数"xerror"。
check_decode_result(ist, got_output, ret);
//如果解码返回成功，做个判断，看是否解码后的长宽、像素格式和最初分析的一样，不一样就输出。没啥用
if (*got_output && ret >= 0) {
if (ist->dec_ctx->width  != decoded_frame->width ||
ist->dec_ctx->height != decoded_frame->height ||
ist->dec_ctx->pix_fmt != decoded_frame->format) {
av_log(NULL, AV_LOG_DEBUG, "Frame parameters mismatch context %d,%d,%d != %d,%d,%d\n",
decoded_frame->width,
decoded_frame->height,
decoded_frame->format,
ist->dec_ctx->width
13eab
,
ist->dec_ctx->height,
ist->dec_ctx->pix_fmt);
}
}

if (!*got_output || ret < 0)
return ret;
//对应参数"top"，和场有关
if(ist->top_field_first>=0)
decoded_frame->top_field_first = ist->top_field_first;
//记录解码的帧数
ist->frames_decoded++;
//和硬件相关
if (ist->hwaccel_retrieve_data && decoded_frame->format == ist->hwaccel_pix_fmt) {
err = ist->hwaccel_retrieve_data(ist->dec_ctx, decoded_frame);
if (err < 0)
goto fail;
}
ist->hwaccel_retrieved_pix_fmt = decoded_frame->format;
//此句很重要，是解码计算出的帧时间戳
best_effort_timestamp= av_frame_get_best_effort_timestamp(decoded_frame);
if(best_effort_timestamp != AV_NOPTS_VALUE) {
//ist中的时间单位是AV_TIME_BASE_Q，decoded_frame中的时间单位是ist->st->time_base
int64_t ts = av_rescale_q(decoded_frame->pts = best_effort_timestamp, ist->st->time_base, AV_TIME_BASE_Q);

if (ts != AV_NOPTS_VALUE)
ist->next_pts = ist->pts = ts;
}

if (debug_ts) {
av_log(NULL, AV_LOG_INFO, "decoder -> ist_index:%d type:video "
"frame_pts:%s frame_pts_time:%s best_effort_ts:%"PRId64" best_effort_ts_time:%s keyframe:%d frame_type:%d time_base:%d/%d\n",
ist->st->index, av_ts2str(decoded_frame->pts),
av_ts2timestr(decoded_frame->pts, &ist->st->time_base),
best_effort_timestamp,
av_ts2timestr(best_effort_timestamp, &ist->st->time_base),
decoded_frame->key_frame, decoded_frame->pict_type,
ist->st->time_base.num, ist->st->time_base.den);
}
/*表示视频就一帧。解完这次就不解了。处理音频时就没有这一句。在process_input_packet中会对此值做处理。
如果处理完还有值就需要在解码*/
pkt->size = 0;

if (ist->st->sample_aspect_ratio.num)
decoded_frame->sample_aspect_ratio = ist->st->sample_aspect_ratio;
//判断判断为真，需要重新设置滤镜链表。
resample_changed = ist->resample_width   != decoded_frame->width  ||
ist->resample_height  != decoded_frame->height ||
ist->resample_pix_fmt != decoded_frame->format;
if (resample_changed) {
av_log(NULL, AV_LOG_INFO,
"Input stream #%d:%d frame changed from size:%dx%d fmt:%s to size:%dx%d fmt:%s\n",
ist->file_index, ist->st->index,
ist->resample_width,  ist->resample_height,  av_get_pix_fmt_name(ist->resample_pix_fmt),
decoded_frame->width, decoded_frame->height, av_get_pix_fmt_name(decoded_frame->format));

ist->resample_width   = decoded_frame->width;
ist->resample_height  = decoded_frame->height;
ist->resample_pix_fmt = decoded_frame->format;

for (i = 0; i < nb_filtergraphs; i++) {
if (ist_in_filtergraph(filtergraphs[i], ist) && ist->reinit_filters &&
configure_filtergraph(filtergraphs[i]) < 0) {
av_log(NULL, AV_LOG_FATAL, "Error reinitializing filters!\n");
exit_program(1);
}
}
}

frame_sample_aspect= av_opt_ptr(avcodec_get_frame_class(), decoded_frame, "sample_aspect_ratio");
for (i = 0; i < ist->nb_filters; i++) {
if (!frame_sample_aspect->num)
*frame_sample_aspect = ist->st->sample_aspect_ratio;

if (i < ist->nb_filters - 1) {
f = ist->filter_frame;
err = av_frame_ref(f, decoded_frame);
if (err < 0)
break;
} else
f = decoded_frame;
//放到滤镜里面
ret = av_buffersrc_add_frame_flags(ist->filters[i]->filter, f, AV_BUFFERSRC_FLAG_PUSH);
if (ret == AVERROR_EOF) {
ret = 0; /* ignore */
} else if (ret < 0) {
av_log(NULL, AV_LOG_FATAL,
"Failed to inject frame into filter network: %s\n", av_err2str(ret));
exit_program(1);
}
}

fail:
av_frame_unref(ist->filter_frame);
av_frame_unref(decoded_frame);
return err < 0 ? err : ret;
}

reap_filters

/**
* Get and encode new output from any of the filtergraphs, without causing
* activity.
*
* @return  0 for success, <0 for severe errors
*/
static int reap_filters(int flush)
{
AVFrame *filtered_frame = NULL;
int i;

/* Reap all buffers present in the buffer sinks */
for (i = 0; i < nb_output_streams; i++) {
OutputStream *ost = output_streams[i];
OutputFile    *of = output_files[ost->file_index];
AVFilterContext *filter;
AVCodecContext *enc = ost->enc_ctx;
int ret = 0;

if (!ost->filter)
continue;
filter = ost->filter->filter;

if (!ost->filtered_frame && !(ost->filtered_frame = av_frame_alloc())) {
return AVERROR(ENOMEM);
}
//从滤镜处理后的帧存放在这里
filtered_frame = ost->filtered_frame;

while (1) {
double float_pts = AV_NOPTS_VALUE; // this is identical to filtered_frame.pts but with higher precision
//从滤镜获取帧
ret = av_buffersink_get_frame_flags(filter, filtered_frame,
AV_BUFFERSINK_FLAG_NO_REQUEST);
if (ret < 0) {
if (ret != AVERROR(EAGAIN) && ret != AVERROR_EOF) {
av_log(NULL, AV_LOG_WARNING,
"Error in av_buffersink_get_frame_flags(): %s\n", av_err2str(ret));
} else if (flush && ret == AVERROR_EOF) {
//没有数据了。已经到了滤镜的结束。
if (filter->inputs[0]->type == AVMEDIA_TYPE_VIDEO)
do_video_out(of->ctx, ost, NULL, AV_NOPTS_VALUE);
}
break;
}
//如果设置了此值，获取到的帧丢弃
if (ost->finished) {
av_frame_unref(filtered_frame);
continue;
}
//下面这段不知道啥意思
if (filtered_frame->pts != AV_NOPTS_VALUE) {
int64_t start_time = (of->start_time == AV_NOPTS_VALUE) ? 0 : of->start_time;
AVRational tb = enc->time_base;
int extra_bits = av_clip(29 - av_log2(tb.den), 0, 16);

tb.den <<= extra_bits;
float_pts =
av_rescale_q(filtered_frame->pts, filter->inputs[0]->time_base, tb) -
av_rescale_q(start_time, AV_TIME_BASE_Q, tb);
float_pts /= 1 << extra_bits;
// avoid exact midoints to reduce the chance of rounding differences, this can be removed in case the fps code is changed to work with integers
float_pts += FFSIGN(float_pts) * 1.0 / (1<<17);

filtered_frame->pts =
av_rescale_q(filtered_frame->pts, filter->inputs[0]->time_base, enc->time_base) -
av_rescale_q(start_time, AV_TIME_BASE_Q, enc->time_base);
}
//if (ost->source_index >= 0)
//    *filtered_frame= *input_streams[ost->source_index]->decoded_frame; //for me_threshold

switch (filter->inputs[0]->type) {
case AVMEDIA_TYPE_VIDEO:
//设置宽高比
if (!ost->frame_aspect_ratio.num)
enc->sample_aspect_ratio = filtered_frame->sample_aspect_ratio;

if (debug_ts) {
av_log(NULL, AV_LOG_INFO, "filter -> pts:%s pts_time:%s exact:%f time_base:%d/%d\n",
av_ts2str(filtered_frame->pts), av_ts2timestr(filtered_frame->pts, &enc->time_base),
float_pts,
enc->time_base.num, enc->time_base.den);
}
//编码并输出到文件
do_video_out(of->ctx, ost, filtered_frame, float_pts);
break;
case AVMEDIA_TYPE_AUDIO:
if (!(enc->codec->capabilities & AV_CODEC_CAP_PARAM_CHANGE) &&
enc->channels != av_frame_get_channels(filtered_frame)) {
av_log(NULL, AV_LOG_ERROR,
"Audio filter graph output is not normalized and encoder does not support parameter changes\n");
break;
}
do_audio_out(of->ctx, ost, filtered_frame);
break;
default:
// TODO support subtitle filters
av_assert0(0);
}

av_frame_unref(filtered_frame);
}
}

return 0;
}

do_video_out

static void do_video_out(AVFormatContext *s,
OutputStream *ost,
AVFrame *next_picture,
double sync_ipts)
{
int ret, format_video_sync;
AVPacket pkt;
AVCodecContext *enc = ost->enc_ctx;
AVCodecContext *mux_enc = ost->st->codec;
int nb_frames, nb0_frames, i;
double delta, delta0;
double duration = 0;
int frame_size = 0;
InputStream *ist = NULL;
AVFilterContext *filter = ost->filter->filter;

if (ost->source_index >= 0)
ist = input_streams[ost->source_index];
/*求duration的值*/
if (filter->inputs[0]->frame_rate.num > 0 &&
filter->inputs[0]->frame_rate.den > 0)
duration = 1/(av_q2d(filter->inputs[0]->frame_rate) * av_q2d(enc->time_base));

if(ist && ist->st->start_time != AV_NOPTS_VALUE && ist->st->first_dts != AV_NOPTS_VALUE && ost->frame_rate.num)
duration = FFMIN(duration, 1/(av_q2d(ost->frame_rate) * av_q2d(enc->time_base)));

if (!ost->filters_script &&
!ost->filters &&
next_picture &&
ist &&
lrintf(av_frame_get_pkt_duration(next_picture) * av_q2d(ist->st->time_base) / av_q2d(enc->time_base)) > 0) {
duration = lrintf(av_frame_get_pkt_duration(next_picture) * av_q2d(ist->st->time_base) / av_q2d(enc->time_base));
}

if (!next_picture) {
//reap_filters函数中在从滤镜中取数据结束后调用do_video_out(of->ctx, ost, NULL, AV_NOPTS_VALUE);会走到这里
//end, flushing
nb0_frames = nb_frames = mid_pred(ost->last_nb0_frames[0],
ost->last_nb0_frames[1],
ost->last_nb0_frames[2]);
} else {
delta0 = sync_ipts - ost->sync_opts; // delta0 is the "drift" between the input frame (next_picture) and where it would fall in the output.
delta  = delta0 + duration;

/* by default, we output a single frame */
nb0_frames = 0; // tracks the number of times the PREVIOUS frame should be duplicated, mostly for variable framerate (VFR)
nb_frames = 1;

format_video_sync = video_sync_method;
if (format_video_sync == VSYNC_AUTO) {
if(!strcmp(s->oformat->name, "avi")) {
format_video_sync = VSYNC_VFR;
} else
format_video_sync = (s->oformat->flags & AVFMT_VARIABLE_FPS) ? ((s->oformat->flags & AVFMT_NOTIMESTAMPS) ? VSYNC_PASSTHROUGH : VSYNC_VFR) : VSYNC_CFR;
if (   ist
&& format_video_sync == VSYNC_CFR
&& input_files[ist->file_index]->ctx->nb_streams == 1
&& input_files[ist->file_index]->input_ts_offset == 0) {
format_video_sync = VSYNC_VSCFR;
}
if (format_video_sync == VSYNC_CFR && copy_ts) {
format_video_sync = VSYNC_VSCFR;
}
}
ost->is_cfr = (format_video_sync == VSYNC_CFR || format_video_sync == VSYNC_VSCFR);

if (delta0 < 0 &&
delta > 0 &&
format_video_sync != VSYNC_PASSTHROUGH &&
format_video_sync != VSYNC_DROP) {
if (delta0 < -0.6) {
av_log(NULL, AV_LOG_WARNING, "Past duration %f too large\n", -delta0);
} else
av_log(NULL, AV_LOG_DEBUG, "Clipping frame in rate conversion by %f\n", -delta0);
sync_ipts = ost->sync_opts;
duration += delta0;
delta0 = 0;
}

switch (format_video_sync) {
case VSYNC_VSCFR:
if (ost->frame_number == 0 && delta0 >= 0.5) {
av_log(NULL, AV_LOG_DEBUG, "Not duplicating %d initial frames\n", (int)lrintf(delta0));
delta = duration;
delta0 = 0;
ost->sync_opts = lrint(sync_ipts);
}
case VSYNC_CFR:
// FIXME set to 0.5 after we fix some dts/pts bugs like in avidec.c
if (frame_drop_threshold && delta < frame_drop_threshold && ost->frame_number) {
nb_frames = 0;
} else if (delta < -1.1)
nb_frames = 0;
else if (delta > 1.1) {
nb_frames = lrintf(delta);
if (delta0 > 1.1)
nb0_frames = lrintf(delta0 - 0.6);
}
break;
case VSYNC_VFR:
if (delta <= -0.6)
nb_frames = 0;
else if (delta > 0.6)
ost->sync_opts = lrint(sync_ipts);
break;
case VSYNC_DROP:
case VSYNC_PASSTHROUGH:
ost->sync_opts = lrint(sync_ipts);
break;
default:
av_assert0(0);
}
}
//之前的都是在算包含几个帧
nb_frames = FFMIN(nb_frames, ost->max_frames - ost->frame_number);
nb0_frames = FFMIN(nb0_frames, nb_frames);

//缓存有三个nb0_frames在last_nb0_frames中，每次[0]是最新的。写之前都向后移。
memmove(ost->last_nb0_frames + 1,
ost->last_nb0_frames,
sizeof(ost->last_nb0_frames[0]) * (FF_ARRAY_ELEMS(ost->last_nb0_frames) - 1));
ost->last_nb0_frames[0] = nb0_frames;

if (nb0_frames == 0 && ost->last_dropped) {
nb_frames_drop++;
av_log(NULL, AV_LOG_VERBOSE,
"*** dropping frame %d from stream %d at ts %"PRId64"\n",
ost->frame_number, ost->st->index, ost->last_frame->pts);
}
if (nb_frames > (nb0_frames && ost->last_dropped) + (nb_frames > nb0_frames)) {
if (nb_frames > dts_error_threshold * 30) {
av_log(NULL, AV_LOG_ERROR, "%d frame duplication too large, skipping\n", nb_frames - 1);
nb_frames_drop++;
return;
}
nb_frames_dup += nb_frames - (nb0_frames && ost->last_dropped) - (nb_frames > nb0_frames);
av_log(NULL, AV_LOG_VERBOSE, "*** %d dup!\n", nb_frames - 1);
}
ost->last_dropped = nb_frames == nb0_frames && next_picture;

/* duplicates frame if needed */
for (i = 0; i < nb_frames; i++) {
AVFrame *in_picture;
av_init_packet(&pkt);
pkt.data = NULL;
pkt.size = 0;

if (i < nb0_frames && ost->last_frame) {
in_picture = ost->last_frame;
} else
in_picture = next_picture;

if (!in_picture)
return;

in_picture->pts = ost->sync_opts;

#if 1
if (!check_recording_time(ost))
#else
if (ost->frame_number >= ost->max_frames)
#endif
return;

#if FF_API_LAVF_FMT_RAWPICTURE
if (s->oformat->flags & AVFMT_RAWPICTURE &&
enc->codec->id == AV_CODEC_ID_RAWVIDEO) {
/* raw pictures are written as AVPicture structure to
avoid any copies. We support temporarily the older
method. */
if (in_picture->interlaced_frame)
mux_enc->field_order = in_picture->top_field_first ? AV_FIELD_TB:AV_FIELD_BT;
else
mux_enc->field_order = AV_FIELD_PROGRESSIVE;
pkt.data   = (uint8_t *)in_picture;
pkt.size   =  sizeof(AVPicture);
pkt.pts    = av_rescale_q(in_picture->pts, enc->time_base, ost->st->time_base);
pkt.flags |= AV_PKT_FLAG_KEY;

write_frame(s, &pkt, ost);
} else
#endif
{
int got_packet, forced_keyframe = 0;
double pts_time;

if (enc->flags & (AV_CODEC_FLAG_INTERLACED_DCT | AV_CODEC_FLAG_INTERLACED_ME) &&
ost->top_field_first >= 0)
in_picture->top_field_first = !!ost->top_field_first;

if (in_picture->interlaced_frame) {
//对场帧的处理
if (enc->codec->id == AV_CODEC_ID_MJPEG)
mux_enc->field_order = in_picture->top_field_first ? AV_FIELD_TT:AV_FIELD_BB;
else
mux_enc->field_order = in_picture->top_field_first ? AV_FIELD_TB:AV_FIELD_BT;
} else
mux_enc->field_order = AV_FIELD_PROGRESSIVE;

in_picture->quality = enc->global_quality;
in_picture->pict_type = 0;

pts_time = in_picture->pts != AV_NOPTS_VALUE ?
in_picture->pts * av_q2d(enc->time_base) : NAN;
//对应参数"force_key_frames"，指定时间戳强制I帧
if (ost->forced_kf_index < ost->forced_kf_count &&
in_picture->pts >= ost->forced_kf_pts[ost->forced_kf_index]) {
ost->forced_kf_index++;
forced_keyframe = 1;
} else if (ost->forced_keyframes_pexpr) {
double res;
ost->forced_keyframes_expr_const_values[FKF_T] = pts_time;
res = av_expr_eval(ost->forced_keyframes_pexpr,
ost->forced_keyframes_expr_const_values, NULL);
ff_dlog(NULL, "force_key_frame: n:%f n_forced:%f prev_forced_n:%f t:%f prev_forced_t:%f -> res:%f\n",
ost->forced_keyframes_expr_const_values[FKF_N],
ost->forced_keyframes_expr_const_values[FKF_N_FORCED],
ost->forced_keyframes_expr_const_values[FKF_PREV_FORCED_N],
ost->forced_keyframes_expr_const_values[FKF_T],
ost->forced_keyframes_expr_const_values[FKF_PREV_FORCED_T],
res);
if (res) {
forced_keyframe = 1;
ost->forced_keyframes_expr_const_values[FKF_PREV_FORCED_N] =
ost->forced_keyframes_expr_const_values[FKF_N];
ost->forced_keyframes_expr_const_values[FKF_PREV_FORCED_T] =
ost->forced_keyframes_expr_const_values[FKF_T];
ost->forced_keyframes_expr_const_values[FKF_N_FORCED] += 1;
}

ost->forced_keyframes_expr_const_values[FKF_N] += 1;
} else if (   ost->forced_keyframes
&& !strncmp(ost->forced_keyframes, "source", 6)
&& in_picture->key_frame==1) {
forced_keyframe = 1;
}

if (forced_keyframe) {
in_picture->pict_type = AV_PICTURE_TYPE_I;
av_log(NULL, AV_LOG_DEBUG, "Forced keyframe at time %f\n", pts_time);
}

update_benchmark(NULL);
if (debug_ts) {
av_log(NULL, AV_LOG_INFO, "encoder <- type:video "
"frame_pts:%s frame_pts_time:%s time_base:%d/%d\n",
av_ts2str(in_picture->pts), av_ts2timestr(in_picture->pts, &enc->time_base),
enc->time_base.num, enc->time_base.den);
}

ost->frames_encoded++;

ret = avcodec_encode_video2(enc, &pkt, in_picture, &got_packet);
update_benchmark("encode_video %d.%d", ost->file_index, ost->index);
if (ret < 0) {
av_log(NULL, AV_LOG_FATAL, "Video encoding failed\n");
exit_program(1);
}
//got_packet为1表示有数据编出。不是每次调用编码器都有帧输出
if (got_packet) {
if (debug_ts) {
av_log(NULL, AV_LOG_INFO, "encoder -> type:video "
"pkt_pts:%s pkt_pts_time:%s pkt_dts:%s pkt_dts_time:%s\n",
av_ts2str(pkt.pts), av_ts2timestr(pkt.pts, &enc->time_base),
av_ts2str(pkt.dts), av_ts2timestr(pkt.dts, &enc->time_base));
}

if (pkt.pts == AV_NOPTS_VALUE && !(enc->codec->capabilities & AV_CODEC_CAP_DELAY))
pkt.pts = ost->sync_opts;
//转换时间基
av_packet_rescale_ts(&pkt, enc->time_base, ost->st->time_base);

if (debug_ts) {
av_log(NULL, AV_LOG_INFO, "encoder -> type:video "
"pkt_pts:%s pkt_pts_time:%s pkt_dts:%s pkt_dts_time:%s\n",
av_ts2str(pkt.pts), av_ts2timestr(pkt.pts, &ost->st->time_base),
av_ts2str(pkt.dts), av_ts2timestr(pkt.dts, &ost->st->time_base));
}

frame_size = pkt.size;
write_frame(s, &pkt, ost);

/* if two pass, output log */
if (ost->logfile && enc->stats_out) {
fprintf(ost->logfile, "%s", enc->stats_out);
}
}
}
ost->sync_opts++;
/*
* For video, number of frames in == number of packets out.
* But there may be reordering, so we can't throw away frames on encoder
* flush, we need to limit them here, before they go into encoder.
*/
ost->frame_number++;

if (vstats_filename && frame_size)
do_video_stats(ost, frame_size);
}

if (!ost->last_frame)
ost->last_frame = av_frame_alloc();
av_frame_unref(ost->last_frame);
if (next_picture && ost->last_frame)
av_frame_ref(ost->last_frame, next_picture);
else
av_frame_free(&ost->last_frame);
}

write_frame

void Cffmpeg::write_frame(AVFormatContext *s, AVPacket *pkt, OutputStream *ost)
{
AVBitStreamFilterContext *bsfc = ost->bitstream_filters;
AVCodecContext          *avctx = ost->encoding_needed ? ost->enc_ctx : ost->st->codec;
int ret;

if (!ost->st->codec->extradata_size && ost->enc_ctx->extradata_size) {
ost->st->codec->extradata = (uint8_t *)av_mallocz(ost->enc_ctx->extradata_size + AV_INPUT_BUFFER_PADDING_SIZE);
if (ost->st->codec->extradata) {
memcpy(ost->st->codec->extradata, ost->enc_ctx->extradata, ost->enc_ctx->extradata_size);
ost->st->codec->extradata_size = ost->enc_ctx->extradata_size;
}
}

if ((avctx->codec_type == AVMEDIA_TYPE_VIDEO && video_sync_method == VSYNC_DROP) ||
(avctx->codec_type == AVMEDIA_TYPE_AUDIO && audio_sync_method < 0))
pkt->pts = pkt->dts = AV_NOPTS_VALUE;

/*
* Audio encoders may split the packets --  #frames in != #packets out.
* But there is no reordering, so we can limit the number of output packets
* by simply dropping them here.
* Counting encoded video frames needs to be done separately because of
* reordering, see do_video_out()
*/
if (!(avctx->codec_type == AVMEDIA_TYPE_VIDEO && avctx->codec)) {
if (ost->frame_number >= ost->max_frames) {
av_packet_unref(pkt);
return;
}
ost->frame_number++;
}
if (avctx->codec_type == AVMEDIA_TYPE_VIDEO) {
int i;
uint8_t *sd = av_packet_get_side_data(pkt, AV_PKT_DATA_QUALITY_STATS,
NULL);
ost->quality = sd ? AV_RL32(sd) : -1;
ost->pict_type = sd ? sd[4] : AV_PICTURE_TYPE_NONE;

for (i = 0; i<FF_ARRAY_ELEMS(ost->error); i++) {
if (sd && i < sd[5])
ost->error[i] = AV_RL64(sd + 8 + 8*i);
else
ost->error[i] = -1;
}

if (ost->frame_rate.num && ost->is_cfr) {
if (pkt->duration > 0)
av_log(NULL, AV_LOG_WARNING, "Overriding packet duration by frame rate, this should not happen\n");
pkt->duration = av_rescale_q(1, av_inv_q(ost->frame_rate),
ost->st->time_base);
}
}
//对应参数"bsf"
if (bsfc)
av_packet_split_side_data(pkt);
//对应参数"bsf"
if ((ret = av_apply_bitstream_filters(avctx, pkt, bsfc)) < 0) {
print_error("", ret);
if (exit_on_error)
exit_program(1);
}
if (pkt->size == 0 && pkt->side_data_elems == 0)
return;
if (!ost->st->codecpar->extradata && avctx->extradata) {
ost->st->codecpar->extradata = (uint8_t *)av_malloc(avctx->extradata_size + FF_INPUT_BUFFER_PADDING_SIZE);
if (!ost->st->codecpar->extradata) {
av_log(NULL, AV_LOG_ERROR, "Could not allocate extradata buffer to copy parser data.\n");
exit_program(1);
}
ost->st->codecpar->extradata_size = avctx->extradata_size;
memcpy(ost->st->codecpar->extradata, avctx->extradata, avctx->extradata_size);
}

if (!(s->oformat->flags & AVFMT_NOTIMESTAMPS)) {
if (pkt->dts != AV_NOPTS_VALUE &&
pkt->pts != AV_NOPTS_VALUE &&
pkt->dts > pkt->pts) {
av_log(s, AV_LOG_WARNING, "Invalid DTS: %"PRId64" PTS: %"PRId64" in output stream %d:%d, replacing by guess\n",
pkt->dts, pkt->pts,
ost->file_index, ost->st->index);
pkt->pts =
pkt->dts = pkt->pts + pkt->dts + ost->last_mux_dts + 1
- FFMIN3(pkt->pts, pkt->dts, ost->last_mux_dts + 1)
- FFMAX3(pkt->pts, pkt->dts, ost->last_mux_dts + 1);
}
if(
(avctx->codec_type == AVMEDIA_TYPE_AUDIO || avctx->codec_type == AVMEDIA_TYPE_VIDEO) &&
pkt->dts != AV_NOPTS_VALUE &&
!(avctx->codec_id == AV_CODEC_ID_VP9 && ost->stream_copy) &&
ost->last_mux_dts != AV_NOPTS_VALUE) {
int64_t max = ost->last_mux_dts + !(s->oformat->flags & AVFMT_TS_NONSTRICT);
if (pkt->dts < max) {
int loglevel = max - pkt->dts > 2 || avctx->codec_type == AVMEDIA_TYPE_VIDEO ? AV_LOG_WARNING : AV_LOG_DEBUG;
av_log(s, loglevel, "Non-monotonous DTS in output stream "
"%d:%d; previous: %"PRId64", current: %"PRId64"; ",
ost->file_index, ost->st->index, ost->last_mux_dts, pkt->dts);
if (exit_on_error) {
av_log(NULL, AV_LOG_FATAL, "aborting.\n");
exit_program(1);
}
av_log(s, loglevel, "changing to %"PRId64". This may result "
"in incorrect timestamps in the output file.\n",
max);
if(pkt->pts >= pkt->dts)
pkt->pts = FFMAX(pkt->pts, max);
pkt->dts = max;
}
}
}
ost->last_mux_dts = pkt->dts;

ost->data_size += pkt->size;
ost->packets_written++;

pkt->stream_index = ost->index;

if (debug_ts) {
av_log(NULL, AV_LOG_INFO, "muxer <- type:%s "
"pkt_pts:%s pkt_pts_time:%s pkt_dts:%s pkt_dts_time:%s size:%d\n",
av_get_media_type_string(ost->enc_ctx->codec_type),
av_ts2str(pkt->pts), av_ts2timestr(pkt->pts, &ost->st->time_base),
av_ts2str(pkt->dts), av_ts2timestr(pkt->dts, &ost->st->time_base),
pkt->size
);
}

ret = av_interleaved_write_frame(s, pkt);
if (ret < 0) {
print_error("av_interleaved_write_frame()", ret);
main_return_code = 1;
close_all_output_streams(ost, (OSTFinished)(MUXER_FINISHED | ENCODER_FINISHED), ENCODER_FINISHED);
}
av_packet_unref(pkt);
}