Author: marco Date: Fri Aug 31 23:12:26 2007 New Revision: 1274 Log: Update TODO
Modified: dirac/TODO Modified: dirac/TODO ============================================================================== --- dirac/TODO (original) +++ dirac/TODO Fri Aug 31 23:12:26 2007 @@ -27,13 +27,13 @@ Optimizations: - Write IDWT in assembler - Perhaps write arithmetic decoding in assembler - Perhaps think about multithreading? -- Merge the loops for halfpel interpolation. -- The code has not been profiled a lot +- Merge the loops for halfpel interpolation (border case checks are + not required anymore) +- Use multiples of 16 pixels for width of refframes - Do not use GetBitContext/PutBitContext for aritmetic (de)coding - Instead of using `struct dirac_blockmotion', it might be faster to save the values separately. - Avoid coeff_posx/posy whenever possible - Avoid malloc during the IDWT/DWT - Interleave IDWT/DWT stages -- Pre-calculate the spation weighting matrices -- Try to pre-calculate the weighting values for qpel/eighthpel +- Pre-calculate the spation weighting matrices for non-common cases _______________________________________________ FFmpeg-soc mailing list [email protected] http://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-soc
