Assert no 8x4/4x8 partition for scaled references
This commit adds a new configure option:
--enable-better-hw-compatibility
The purpose of the configure option is to provide information on known
hardware decoder implementation bugs, so encoder implementers may
choose to implement their encoders in a way to avoid triggering these
decoder bugs.
The WebM team were made aware of that a number of hardware decoders
have trouble in handling the combination of scaled frame reference
frame and 8x4 or 4x8 partitions. This commit added asserts to vp9
decoder, so when built with above configure option, the decoder can
assert if an input bitstream triggers such decoder bug.
Change-Id: I386204cfa80ed16b50ebde57f886121ed76200bf
diff --git a/configure b/configure
index 24992c4..2053769 100755
--- a/configure
+++ b/configure
@@ -35,6 +35,9 @@
${toggle_debug_libs} in/exclude debug version of libraries
${toggle_static_msvcrt} use static MSVCRT (VS builds only)
${toggle_vp9_highbitdepth} use VP9 high bit depth (10/12) profiles
+ ${toggle_better_hw_compatibility}
+ enable encoder to produce streams with better
+ hardware decoder compatibility
${toggle_vp8} VP8 codec support
${toggle_vp9} VP9 codec support
${toggle_vp10} VP10 codec support
@@ -319,6 +322,7 @@
vp9_temporal_denoising
coefficient_range_checking
vp9_highbitdepth
+ better_hw_compatibility
experimental
size_limit
${EXPERIMENT_LIST}
@@ -377,6 +381,7 @@
temporal_denoising
vp9_temporal_denoising
coefficient_range_checking
+ better_hw_compatibility
vp9_highbitdepth
experimental
"
diff --git a/vp9/common/vp9_reconinter.c b/vp9/common/vp9_reconinter.c
index d8c14ec..3eb19b1 100644
--- a/vp9/common/vp9_reconinter.c
+++ b/vp9/common/vp9_reconinter.c
@@ -187,6 +187,10 @@
const int is_scaled = vp9_is_scaled(sf);
if (is_scaled) {
+#if CONFIG_BETTER_HW_COMPATIBILITY
+ assert(xd->mi[0]->mbmi.sb_type != BLOCK_4X8 &&
+ xd->mi[0]->mbmi.sb_type != BLOCK_8X4);
+#endif
// Co-ordinate of containing block to pixel precision.
const int x_start = (-xd->mb_to_left_edge >> (3 + pd->subsampling_x));
const int y_start = (-xd->mb_to_top_edge >> (3 + pd->subsampling_y));