The way I am currently using is soft decoding, ffmpeg version is out of personal guess is that the CPU after being used, decoding the cumulative delay caused by not in time,
Reconnection flow can eliminate the time delay, but I want to know more direct optimization scheme, because can't put into exile in practical use with with oneself reconnection,
Looking forward to have Daniel can help,
CodePudding user response:
Live class should play after caton or rapid display to the latest progress, accelerate the buffer or abandon the card immediately did not play the streaming data,
I don't know you live to deal, say,
CodePudding user response:
Send lost frames, I have lost to the next frame come so far, I have clear the cache, the current preliminary as to is an audio problem over there, I feel is a question of resampling over there,CodePudding user response:
Through my own screening, later found to be the CPU capacity, audio output to the sound card change is slow, and my program set up until the completion of a frame audio playback is played a frame of audio, audio and video synchronization with audio and video synchronization, so lead to this problem, I write down the reason, the hope can give later encountered similar problems under referenceCodePudding user response:
Add a cache,