Hi, One of the things which I couldn't mention in my previous mail was that in my application. I've two video decoder one is based on ffms2 (a wrapper over ffmpeg/libav) and the other is based on ffmpeg. Initially I had just ffmpeg based video loader but since I was facing issues with seeking I shifted to ffms but ffmpeg is also. I can chose between the two whenever I want to play a video file. The problem which I have describe is specifically with ffms2 use of sws_scale function. I've modified the code of ffms2 so that code for both ffms and ffmpeg base decoder are same for playing video. After analysing the timings of two decoder, I found that sws_scale used in ffms is taking 15milisec whereas ffmpeg use of sws_scale take 2milisec, all the parameters in both cases are same for same video file without playing any audio. (audio decoding is taking constant irrespective of frame size can be enabled or disabled) The only difference which I think is that ffms2 is loading as dll whereas in case of ffmpeg it is statically. But I think this shouldn't matter. Please share your thoughts.
Thanks and Regards, Pradeep -- View this message in context: http://libav-users.943685.n4.nabble.com/Frame-decoding-and-sw-scale-timings-for-HD-video-tp4658429p4658436.html Sent from the libav-users mailing list archive at Nabble.com. _______________________________________________ Libav-user mailing list [email protected] http://ffmpeg.org/mailman/listinfo/libav-user
