Mercurial > libavcodec.hg
diff mpegvideo.h @ 1505:010f76d07a27 libavcodec
use lagrange multipler instead of qp for ratecontrol, this may break some things, tell me ASAP if u notice anything broken
quality which was 1..31 float is now a 1..FF_LAMBDA_MAX int, and FF_QP2LAMBDA * qp can be used to convert to the new range
author | michaelni |
---|---|
date | Tue, 07 Oct 2003 11:32:40 +0000 |
parents | 0355f2b3519a |
children | ad5461bad820 |
line wrap: on
line diff
--- a/mpegvideo.h Sun Oct 05 21:48:16 2003 +0000 +++ b/mpegvideo.h Tue Oct 07 11:32:40 2003 +0000 @@ -354,7 +354,9 @@ uint8_t *edge_emu_buffer; ///< points into the middle of allocated_edge_emu_buffer int qscale; ///< QP - float frame_qscale; ///< qscale from the frame level rc FIXME remove + int lambda; ///< lagrange multipler used in rate distortion + int lambda2; ///< (lambda*lambda) >> FF_LAMBDA_SHIFT + int *lambda_table; int adaptive_quant; ///< use adaptive quantization int dquant; ///< qscale difference to prev qscale int pict_type; ///< I_TYPE, P_TYPE, B_TYPE, ...