Hi,
I am working on a directfb video output for a mpeg2 video player.
I'm using libavcodec/libavformat/libavutil to mpeg2 decoding process and
I would like to optimize the frame display routine.
I read the libsoftmpeg code and I wrote a display routine where there are memcpys from
AVFrame structure to YUV420 DirectFB Surface. Then, the data are blitted from YUV420 DirectFB
Surface to RBG DirectFB Surface.
My question is: Are the memcpys really necessary? Is it possible to avoid memcpys?
Is it possible to use preallocated data field to avoid memcpys?
Thanks in advance,
I am working on a directfb video output for a mpeg2 video player.
I'm using libavcodec/libavformat/libavutil to mpeg2 decoding process and
I would like to optimize the frame display routine.
I read the libsoftmpeg code and I wrote a display routine where there are memcpys from
AVFrame structure to YUV420 DirectFB Surface. Then, the data are blitted from YUV420 DirectFB
Surface to RBG DirectFB Surface.
My question is: Are the memcpys really necessary? Is it possible to avoid memcpys?
Is it possible to use preallocated data field to avoid memcpys?
Thanks in advance,
Peter.
_______________________________________________ directfb-users mailing list [email protected] http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
