解決ffmpeg拉流轉發頻繁丟包問題max delay reached. need to consume packet

軟件：

一、流媒體服務器EasyDarwin-windows-8.1.0-1901141151

二、ffmpeg-20181001-dcbd89e-win64-static

三、直播源：rtsp://192.168.1.168/0

四、流媒體服務器EasyDarwin地址rtsp://192.168.1.28/3

問題現象

[rtsp @ 0000000000122bc0] max delay reached. need to consume packet
[rtsp @ 0000000000122bc0] RTP: missed 33 packets
[rtsp @ 0000000000122bc0] max delay reached. need to consume packet
[rtsp @ 0000000000122bc0] RTP: missed 166 packets
[h264 @ 0000000002f4b280] out of range intra chroma pred mode
[h264 @ 0000000002f4b280] error while decoding MB 27 39
[h264 @ 0000000002f4b280] concealing 502 DC, 502 AC, 502 MV errors in P f
[h264 @ 0000000002ed08c0] Increasing reorder buffer to 8
frame=  964 fps= 30 q=29.0 size=N/A time=00:00:33.26 bitrate=N/A dup=141
frame=  999 fps= 31 q=29.0 size=N/A time=00:00:33.77 bitrate=N/A dup=170
[rtsp @ 0000000000122bc0] max delay reached. need to consume packet
[rtsp @ 0000000000122bc0] RTP: missed 12 packets
[h264 @ 0000000002ed1000] corrupted macroblock 55 32 (total_coeff=-1)
[h264 @ 0000000002ed1000] error while decoding MB 55 32

　　

解決方法：

 

採用如下命令進行拉流轉發，運行了幾個小時，幾乎沒有出現丟包現象

ffmpeg -rtsp_transport tcp -i rtsp://192.168.1.168/0 -vcodec copy -f rtsp rtsp://192.168.1.28/11

 

緣由：這是rtsp協議默認使用udp致使的問題，因此rtsp強制使用tcp方式能夠必定程度避免丟包。

若是是拉rtsp轉rtmp，命令是（沒測試）

ffmpeg -rtsp_transport tcp -i rtsp://admin:password@192.168.1.11:554 -vcodec copy -f flv -an rtmp://localhost/live/tests

 

參考：https://github.com/eguid/FFCH4J/issues/11