From: Jingning Han <jingning@google.com>
Date: Wed, 10 Jul 2013 22:45:34 +0000 (-0700)
Subject: Fix tx_type bug in intra4x4 rd loop
X-Git-Tag: v1.3.0~896^2
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=18803f9cc4386e8631090471a948a3baad3fb3c0;p=libvpx

Fix tx_type bug in intra4x4 rd loop

This commit fixed the mis-use of the tx_type for inverse transform
in intra4x4 rate-distortion optimization loop. It improves the
overall coding performance.

Change-Id: I7fe9953175b74890357dbcee33c138573766e980
---

diff --git a/vp9/encoder/vp9_rdopt.c b/vp9/encoder/vp9_rdopt.c
index 4cb38f7a9..f42467969 100644
--- a/vp9/encoder/vp9_rdopt.c
+++ b/vp9/encoder/vp9_rdopt.c
@@ -1288,9 +1288,9 @@ static int64_t rd_pick_intra4x4block(VP9_COMP *cpi, MACROBLOCK *x, int ib,
                                                           block, 16),
                                       16, &ssz) >> 2;
 
-        if (best_tx_type != DCT_DCT)
+        if (tx_type != DCT_DCT)
           vp9_short_iht4x4_add(BLOCK_OFFSET(pd->dqcoeff, block, 16),
-                               dst, pd->dst.stride, best_tx_type);
+                               dst, pd->dst.stride, tx_type);
         else
           xd->inv_txm4x4_add(BLOCK_OFFSET(pd->dqcoeff, block, 16),
                              dst, pd->dst.stride);