h264编解码末尾丢帧问题原因和解决

相关

x264编码
ffmpeg解码
ffmpeg编码

问题

编解码h264流时，会发现末尾丢帧。以ffmpeg为例，调用如下接口
int avcodec_encode_video2 ( AVCodecContext * avctx, AVPacket * avpkt, const AVFrame * frame, int * got_packet_ptr )

Encode a frame of video.
Takes input raw video data from frame and writes the next output packet, if available, to avpkt. The output packet does not necessarily contain data for the most recent frame, as encoders can delay and reorder input frames internally as needed.

int avcodec_decode_video2 ( AVCodecContext * avctx, AVFrame * picture, int * got_picture_ptr, const AVPacket * avpkt )

Decode the video frame of size avpkt->size from avpkt->data into picture.
Some decoders may support multiple frames in a single AVPacket, such decoders would then just decode the first frame.

官网api的提到：

Takes input raw video data from frame and writes the next output packet, if available, to avpkt. The output packet does not necessarily contain data for the most recent frame, as encoders can delay and reorder input frames internally as needed.

原因：

拿编码来说，丢帧形式如下图：

这被称为编码延迟。延迟原因又分为两种，一是计算延迟，二是缓存延迟。
所以：
h->frames.i_delay =
param->i_sync_lookahead + // 前向考虑帧数
max ( param->i_bframe, // B帧数量
param->rc.i_lookahead) + // 码率控制前向考虑帧数
另外编码时，通过设置参数将编码和获取的帧间隔缩小到0，参考“zerolatency"的tune值中参数的设置。

解决

编码延迟就需要在编码所有yuv数据之后必须flush。
int vflush_encoder(AVFormatContext *fmt_ctx, unsigned int stream_index) { int ret = 0; int got_frame; AVPacket enc_pkt; if (!(fmt_ctx->streams[stream_index]->codec->codec->capabilities & CODEC_CAP_DELAY)) return 0; av_init_packet(&enc_pkt); while (IS_GOING) { enc_pkt.data = NULL; enc_pkt.size = 0; ret = avcodec_encode_video2(fmt_ctx->streams[stream_index]->codec, &enc_pkt, NULL, &got_frame); if (ret < 0) break; if (!got_frame) { ret = 0; break; } ret = av_write_frame(fmt_ctx, &enc_pkt); printf("Flush Encoder: Succeed to encode 1 frame!\tsize:%5d\n", enc_pkt.size); av_free_packet(&enc_pkt); if (ret < 0) break; } return ret; }

解码同理，在av_read_frame到没有avpacket可以读取之后，要继续调用avcodec_encode_video2，将视频末尾帧解码出来。