ffmpeg 源代码简单分析 ： av_read_frame()
此前写了好几篇ffmpeg源代码分析文章，列表如下：
图解FFMPEG打开媒体的函数avformat_open_input
ffmpeg 源代码简单分析 ： av_register_all()
ffmpeg 源代码简单分析 ： avcodec_register_all()
ffmpeg 源代码简单分析 ： av_read_frame()
ffmpeg 源代码简单分析 ： avcodec_decode_video2()
============================
ffmpeg中的av_read_frame()的作用是读取码流中的音频若干帧或者视频一帧。例如，解码视频的时候，每解码一个视频帧，需要先调用 av_read_frame()获得一帧视频的压缩数据，然后才能对该数据进行解码（例如H.264中一帧压缩数据通常对应一个NAL）。
对该函数源代码的分析是很久之前做的了，现在翻出来，用博客记录一下。
上代码之前，先参考了其他人对av_read_frame()的解释，在此做一个参考：
通过av_read_packet(***)，读取一个包，需要说明的是此函数必须是包含整数帧的，不存在半帧的情况，以ts流为例，是读取一个完整的PES包（一个完整pes包包含若干视频或音频es包），读取完毕后，通过av_parser_parse2(***)分析出视频一帧（或音频若干帧），返回，下次进入循环的时候，如果上次的数据没有完全取完，则st = s->cur_st;不会是NULL，即再此进入av_parser_parse2(***)流程，而不是下面的av_read_packet（**）流程，这样就保证了，如果读取一次包含了N帧视频数据（以视频为例），则调用av_read_frame（***）N次都不会去读数据，而是返回第一次读取的数据，直到全部解析完毕。
av_read_frame()的源代码如下：
//获取一个AVPacket
/*
* av_read_frame - 新版本的ffmpeg用的是av_read_frame，而老版本的是av_read_packet
* 。区别是av_read_packet读出的是包，它可能是半帧或多帧，不保证帧的完整性。av_read_frame对
* av_read_packet进行了封装，使读出的数据总是完整的帧
*/
int av_read_frame(AVFormatContext *s, AVPacket *pkt)
{
const int genpts = s->flags & AVFMT_FLAG_GENPTS;
int eof = 0;
if (!genpts)
/**
* This buffer is only needed when packets were already buffered but
* not decoded, for example to get the codec parameters in MPEG
* streams.
* 一般情况下会调用read_frame_internal(s, pkt)
* 直接返回
*/
return s->packet_buffer ? read_from_packet_buffer(s, pkt) :
read_frame_internal(s, pkt);
for (;;) {
int ret;
AVPacketList *pktl = s->packet_buffer;
if (pktl) {
AVPacket *next_pkt = &pktl->pkt;
if (next_pkt->dts != AV_NOPTS_VALUE) {
int wrap_bits = s->streams[next_pkt->stream_index]->pts_wrap_bits;
while (pktl && next_pkt->pts == AV_NOPTS_VALUE) {
if (pktl->pkt.stream_index == next_pkt->stream_index &&
(av_compare_mod(next_pkt->dts, pktl->pkt.dts, 2LL << (wrap_bits - 1)) < 0) &&
av_compare_mod(pktl->pkt.pts, pktl->pkt.dts, 2LL << (wrap_bits - 1))) { //not b frame
next_pkt->pts = pktl->pkt.dts;
}
pktl = pktl->next;
}
pktl = s->packet_buffer;
}
/* read packet from packet buffer, if there is data */
if (!(next_pkt->pts == AV_NOPTS_VALUE &&
next_pkt->dts != AV_NOPTS_VALUE && !eof))
return read_from_packet_buffer(s, pkt);
}
ret = read_frame_internal(s, pkt);
if (ret < 0) {
if (pktl && ret != AVERROR(EAGAIN)) {
eof = 1;
continue;
} else
return ret;
}
if (av_dup_packet(add_to_pktbuf(&s->packet_buffer, pkt,
&s->packet_buffer_end)) < 0)
return AVERROR(ENOMEM);
}
}
一般情况下，av_read_frame()会调用read_frame_internal()，其代码如下所示：
//av_read_frame对他进行了封装
static int read_frame_internal(AVFormatContext *s, AVPacket *pkt)
{
AVStream *st;
int len, ret, i;
//初始化
av_init_packet(pkt);
for(;;) {
/* 选择当前的 input stream */
st = s->cur_st;
if (st) {
//不需要解析。不清楚哪些数据属于这类
if (!st->need_parsing || !st->parser) {
/* no parsing needed: we just output the packet as is */
/* raw data support */
*pkt = st->cur_pkt;
st->cur_pkt.data= NULL;
st->cur_pkt.side_data_elems = 0;
st->cur_pkt.side_data = NULL;
compute_pkt_fields(s, st, NULL, pkt);
s->cur_st = NULL;
if ((s->iformat->flags & AVFMT_GENERIC_INDEX) &&
(pkt->flags & AV_PKT_FLAG_KEY) && pkt->dts != AV_NOPTS_VALUE) {
ff_reduce_index(s, st->index);
av_add_index_entry(st, pkt->pos, pkt->dts, 0, 0, AVINDEX_KEYFRAME);
}
break;
} //需要解析
else if (st->cur_len > 0 && st->discard < AVDISCARD_ALL) {
//解析
len = av_parser_parse2(st->parser, st->codec, &pkt->data, &pkt->size,
st->cur_ptr, st->cur_len,
st->cur_pkt.pts, st->cur_pkt.dts,
st->cur_pkt.pos);
st->cur_pkt.pts = AV_NOPTS_VALUE;
st->cur_pkt.dts = AV_NOPTS_VALUE;
/* increment read pointer */
st->cur_ptr += len;
st->cur_len -= len;
/* return packet if any */
if (pkt->size) {
got_packet:
pkt->duration = 0;
pkt->stream_index = st->index;
pkt->pts = st->parser->pts;
pkt->dts = st->parser->dts;
pkt->pos = st->parser->pos;
if(pkt->data == st->cur_pkt.data && pkt->size == st->cur_pkt.size){
s->cur_st = NULL;
pkt->destruct= st->cur_pkt.destruct;
st->cur_pkt.destruct= NULL;
st->cur_pkt.data = NULL;
assert(st->cur_len == 0);
}else{
pkt->destruct = NULL;
}
compute_pkt_fields(s, st, st->parser, pkt);
if((s->iformat->flags & AVFMT_GENERIC_INDEX) && pkt->flags & AV_PKT_FLAG_KEY){
int64_t pos= (st->parser->flags & PARSER_FLAG_COMPLETE_FRAMES) ? pkt->pos : st->parser->frame_offset;
ff_reduce_index(s, st->index);
av_add_index_entry(st, pos, pkt->dts,
0, 0, AVINDEX_KEYFRAME);
}
break;
}
} else {
/* free packet */
av_free_packet(&st->cur_pkt);
s->cur_st = NULL;
}
} else {
AVPacket cur_pkt;
/* read next packet */
//读取AVPacket,老版本里只有av_read_packet，现在被封装了
ret = av_read_packet(s, &cur_pkt);
if (ret < 0) {
if (ret == AVERROR(EAGAIN))
return ret;
/* return the last frames, if any */
for(i = 0; i < s->nb_streams; i++) {
st = s->streams[i];
if (st->parser && st->need_parsing) {
av_parser_parse2(st->parser, st->codec,
&pkt->data, &pkt->size,
NULL, 0,
AV_NOPTS_VALUE, AV_NOPTS_VALUE,
AV_NOPTS_VALUE);
if (pkt->size)
goto got_packet;
}
}
/* no more packets: really terminate parsing */
return ret;
}
st = s->streams[cur_pkt.stream_index];
st->cur_pkt= cur_pkt;
if(st->cur_pkt.pts != AV_NOPTS_VALUE &&
st->cur_pkt.dts != AV_NOPTS_VALUE &&
st->cur_pkt.pts < st->cur_pkt.dts){
av_log(s, AV_LOG_WARNING, "Invalid timestamps stream=%d, pts=%"PRId64", dts=%"PRId64", size=%d\n",
st->cur_pkt.stream_index,
st->cur_pkt.pts,
st->cur_pkt.dts,
st->cur_pkt.size);
// av_free_packet(&st->cur_pkt);
// return -1;
}
if(s->debug & FF_FDEBUG_TS)
av_log(s, AV_LOG_DEBUG, "av_read_packet stream=%d, pts=%"PRId64", dts=%"PRId64", size=%d, duration=%d, flags=%d\n",
st->cur_pkt.stream_index,
st->cur_pkt.pts,
st->cur_pkt.dts,
st->cur_pkt.size,
st->cur_pkt.duration,
st->cur_pkt.flags);
s->cur_st = st;
st->cur_ptr = st->cur_pkt.data;
st->cur_len = st->cur_pkt.size;
if (st->need_parsing && !st->parser && !(s->flags & AVFMT_FLAG_NOPARSE)) {
st->parser = av_parser_init(st->codec->codec_id);
if (!st->parser) {
av_log(s, AV_LOG_VERBOSE, "parser not found for codec "
"%s, packets or times may be invalid.\n",
avcodec_get_name(st->codec->codec_id));
/* no parser available: just output the raw packets */
st->need_parsing = AVSTREAM_PARSE_NONE;
}else if(st->need_parsing == AVSTREAM_PARSE_HEADERS){
st->parser->flags |= PARSER_FLAG_COMPLETE_FRAMES;
}else if(st->need_parsing == AVSTREAM_PARSE_FULL_ONCE){
st->parser->flags |= PARSER_FLAG_ONCE;
}
}
}
}
if(s->debug & FF_FDEBUG_TS)
av_log(s, AV_LOG_DEBUG, "read_frame_internal stream=%d, pts=%"PRId64", dts=%"PRId64", size=%d, duration=%d, flags=%d\n",
pkt->stream_index,
pkt->pts,
pkt->dts,
pkt->size,
pkt->duration,
pkt->flags);
return 0;
}
一般的码流都需要解析，这是需要调用av_paser_parse2()，它的代码如下所示：
//解析。例如解析264里的NAL等等
int av_parser_parse2(AVCodecParserContext *s,
AVCodecContext *avctx,
uint8_t **poutbuf, int *poutbuf_size,
const uint8_t *buf, int buf_size,
int64_t pts, int64_t dts,
int64_t pos)
{
int index, i;
uint8_t dummy_buf[FF_INPUT_BUFFER_PADDING_SIZE];
if(!(s->flags & PARSER_FLAG_FETCHED_OFFSET)) {
s->next_frame_offset =
s->cur_offset = pos;
s->flags |= PARSER_FLAG_FETCHED_OFFSET;
}
if (buf_size == 0) {
/* padding is always necessary even if EOF, so we add it here */
memset(dummy_buf, 0, sizeof(dummy_buf));
buf = dummy_buf;
} else if (s->cur_offset + buf_size !=
s->cur_frame_end[s->cur_frame_start_index]) { /* skip remainder packets */
/* add a new packet descriptor */
i = (s->cur_frame_start_index + 1) & (AV_PARSER_PTS_NB - 1);
s->cur_frame_start_index = i;
s->cur_frame_offset[i] = s->cur_offset;
s->cur_frame_end[i] = s->cur_offset + buf_size;
s->cur_frame_pts[i] = pts;
s->cur_frame_dts[i] = dts;
s->cur_frame_pos[i] = pos;
}
if (s->fetch_timestamp){
s->fetch_timestamp=0;
s->last_pts = s->pts;
s->last_dts = s->dts;
s->last_pos = s->pos;
ff_fetch_timestamp(s, 0, 0);
}
/* WARNING: the returned index can be negative */
//H264里对应的就是parser_parse=h264_parse,
index = s->parser->parser_parse(s, avctx, (const uint8_t **)poutbuf, poutbuf_size, buf, buf_size);
//av_log(NULL, AV_LOG_DEBUG, "parser: in:%"PRId64", %"PRId64", out:%"PRId64", %"PRId64", in:%d out:%d id:%d\n", pts, dts, s->last_pts, s->last_dts, buf_size, *poutbuf_size, avctx->codec_id);
/* update the file pointer */
if (*poutbuf_size) {
/* fill the data for the current frame */
s->frame_offset = s->next_frame_offset;
/* offset of the next frame */
s->next_frame_offset = s->cur_offset + index;
s->fetch_timestamp=1;
}
if (index < 0)
index = 0;
s->cur_offset += index;
return index;
}
从index = s->parser->parser_parse(s, avctx, (const uint8_t **)poutbuf, poutbuf_size, buf, buf_size);这句代码可以看出，最终调用了相应解码器的parser_parse()函数。
有点累了，先不做详细分析，以后有机会再补上。