rv40dsp: implement prescaled versions for biweight.

Quite often, the original weights are multiple of 512. By prescaling them
by 1/512 when they are computed (once per frame), no intermediate shifting
is needed, and no prescaling on each call either.

The x86 code already used that trick.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

1 files changed, 6 insertions, 1 deletions

diff --git a/libavcodec/rv34dsp.h b/libavcodec/rv34dsp.h
index c70194cc20..58da59f038 100644
--- a/libavcodec/rv34dsp.h
+++ b/libavcodec/rv34dsp.h
@@ -58,7 +58,12 @@ typedef struct RV34DSPContext {
     qpel_mc_func avg_pixels_tab[4][16];
     h264_chroma_mc_func put_chroma_pixels_tab[3];
     h264_chroma_mc_func avg_chroma_pixels_tab[3];
-    rv40_weight_func rv40_weight_pixels_tab[2];
+    /**
+     * Biweight functions, first dimension is transform size (16/8),
+     * second is whether the weight is prescaled by 1/512 to skip
+     * the intermediate shifting.
+     */
+    rv40_weight_func rv40_weight_pixels_tab[2][2];
     rv34_inv_transform_func rv34_inv_transform;
     rv34_inv_transform_func rv34_inv_transform_dc;
     rv34_idct_add_func rv34_idct_add;