My application uses FFmpeg to demux and decode video streams from IP cameras. Video is encoded with either H.264, H.265, or MJPEG. Formats used include RTSP, and MPEG-PS. Currently I am using avformat_find_stream_info
function to init format context when opening stream, and it works OK, but it introduces some delay at the beginning. I was able to reduce it by changing probe size, but I can't make it less than 2 seconds without making it unreliable. I would like to get rid of this delay.
I can get stream information from camera before opening it, so I wanted to initialize stream information in AVFormatContext
without probing it. Is it possible with FFmpeg API?
question from:
https://stackoverflow.com/questions/66067104/ffmpeg-initiating-codec-parameters-manually 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…