diff options
author | Ronald S. Bultje <rbultje@google.com> | 2013-03-26 16:46:09 -0700 |
---|---|---|
committer | Ronald S. Bultje <rbultje@google.com> | 2013-03-26 16:46:09 -0700 |
commit | 790fb132157aab73b20c94a223dbe4470036d832 (patch) | |
tree | 3b12459d77a97373b4e4c9240a4d135a4ef874f5 /vp9/encoder/vp9_tokenize.h | |
parent | 646616602dca641f8fe38ca4c7adbb77df5ae8e3 (diff) | |
download | libvpx-790fb132157aab73b20c94a223dbe4470036d832.tar libvpx-790fb132157aab73b20c94a223dbe4470036d832.tar.gz libvpx-790fb132157aab73b20c94a223dbe4470036d832.tar.bz2 libvpx-790fb132157aab73b20c94a223dbe4470036d832.zip |
Use above/left (instead of previous in scan-order) as token context.
Pearson correlation for above or left is significantly higher than for
previous-in-scan-order (absolute values depend on position in scan, but
in general, we gain about 0.1-0.2 by using either above or left; using
both basically just makes this even better). For eob branch skipping,
we continue to use the previous token in scan order.
This helps about 0.9% on derf after re-training on a limited data set.
Full re-training and results on larger-resolution clips are pending.
Note that this commit breaks trellis, so we can probably get further
gains out of it by fixing trellis at some later point.
Change-Id: Iead68e296fc3a105cca746b5e3da9555d6010cfe
Diffstat (limited to 'vp9/encoder/vp9_tokenize.h')
-rw-r--r-- | vp9/encoder/vp9_tokenize.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/vp9/encoder/vp9_tokenize.h b/vp9/encoder/vp9_tokenize.h index 464d7caf6..907f814ac 100644 --- a/vp9/encoder/vp9_tokenize.h +++ b/vp9/encoder/vp9_tokenize.h @@ -29,7 +29,7 @@ typedef struct { } TOKENEXTRA; typedef int64_t vp9_coeff_accum[REF_TYPES][COEF_BANDS][PREV_COEF_CONTEXTS] - [MAX_ENTROPY_TOKENS]; + [MAX_ENTROPY_TOKENS + 1]; int vp9_mby_is_skippable_4x4(MACROBLOCKD *xd); int vp9_mbuv_is_skippable_4x4(MACROBLOCKD *xd); |