Added vp9_short_idct1_32x32_c

and called this function in vp9_dequant_idct_add_32x32_c when
eob == 1.  For the test clip used, the decoder performance improved
by 21+%.  Based on Yaowu's 16 point idct work.

Change-Id: Ib579a90fed531d45777980e04bf0c9b23c093c43

1 files changed, 16 insertions, 7 deletions

diff --git a/vp9/decoder/vp9_dequantize.c b/vp9/decoder/vp9_dequantize.c
index e46be3ac7..18d4e59c7 100644
--- a/vp9/decoder/vp9_dequantize.c
+++ b/vp9/decoder/vp9_dequantize.c
@@ -349,13 +349,22 @@ void vp9_dequant_idct_add_32x32_c(int16_t *input, const int16_t *dq,
   int i;
 
   if (eob) {
-    input[0]= input[0] * dq[0] / 2;
-    for (i = 1; i < 1024; i++)
-      input[i] = input[i] * dq[1] / 2;
-    vp9_short_idct32x32_c(input, output, 64);
-    vpx_memset(input, 0, 2048);
-
-    add_residual(output, pred, pitch, dest, stride, 32, 32);
+    input[0] = input[0] * dq[0] / 2;
+#if !CONFIG_DWTDCTHYBRID
+    if (eob == 1) {
+      vp9_short_idct1_32x32_c(input, output);
+      add_constant_residual(output[0], pred, pitch, dest, stride, 32, 32);
+      input[0] = 0;
+    } else {
+#endif
+      for (i = 1; i < 1024; i++)
+        input[i] = input[i] * dq[1] / 2;
+      vp9_short_idct32x32_c(input, output, 64);
+      vpx_memset(input, 0, 2048);
+      add_residual(output, pred, pitch, dest, stride, 32, 32);
+#if !CONFIG_DWTDCTHYBRID
+    }
+#endif
   }
 }