You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* direct calling of opengl without going through sdl for video, audio
* use arm assembly code for copying pixels between arrays
* skip bidrectional frames -code available
* 16bit or 24bit rendering based on video lag condition
* multi threaded single frame decoding
* multi threaded - parallel frames decoding -ffmpeg
* Time slicing by OS for each thread(Read packets, Decode Video, Audio Thread, Put Frames to Screen thread) needs to be verified.(Whether they are given equal time slices, concurrently)
* Process specific optimization during compilation
* Java overhead for layout, display
* Using surface flinger in native ??
* Use profiler and convert heavily used methods to assembly arm/x86
* Look at gstreamer, vlc ->looks like, they are using surface flinger
* optimized arm compilation using the compiler from arm.com
* Use libmedia, libsurfaceflinger_client based output of audio and video
* H/w decoding (stagefright framework/OMX)
Details required
* use inbuilt decoder for known file formats (avi)
* use hardware acceleration - not available
* using gstreamer framework ??
* using libstagefright
Known questions and possible answers
* what to do if the decoding of a frame itself is >40ms, for a frame rate of 25FPS ? skip frames, disable audio, convert the file to a supported native file format with less FPS ??
Suggestions
* Send your suggestions or any possible solutions to [email protected]