[x265] [PATCH] created partition size and function pointer for luma filter
Steve Borho
steve at borho.org
Thu Oct 17 20:54:41 CEST 2013
On Thu, Oct 17, 2013 at 10:25 AM, <praveen at multicorewareinc.com> wrote:
> # HG changeset patch
> # User Praveen Tiwari
> # Date 1382023511 -19800
> # Node ID b5375476fdedcce6e60c71c215ed95c923b9df43
> # Parent dfae391107c3d2823adabe284f1c08278669d8f0
> created partition size and function pointer for luma filter
>
> diff -r dfae391107c3 -r b5375476fded source/common/primitives.h
> --- a/source/common/primitives.h Thu Oct 17 16:05:37 2013 +0530
> +++ b/source/common/primitives.h Thu Oct 17 20:55:11 2013 +0530
> @@ -148,6 +148,20 @@
> NUM_CHROMA_PARTITIONS
> };
>
> +//Each luma size of table will be used to genrate 3 partion sizes, except
> few exceptions.
>
Comments need a space after //
> +//WxH will be used for WxH, Wx(H + 1) and (W + 1)x(H), like 4x8 will
> genrate 4x8, 4x9 and 5x9
>
generate
+//exceptions are 16x64, 48x64, 64x16, 64x48 because they don't have
> corresponding odd block sizes i.e (W + 1)xH and Wx(H + 1)
>
> +enum LumaPartions
>
LumaPartitions
> +{
> + LUMA_PARTITION_4x8, LUMA_PARTITION_4x16 = 3,
> LUMA_PARTITION_8x4 = 6, LUMA_PARTITION_8x8 = 9,
> + LUMA_PARTITION_8x16 = 12, LUMA_PARTITION_8x32 = 15,
> LUMA_PARTITION_12x16 = 18, LUMA_PARTITION_16x4 = 21,
> + LUMA_PARTITION_16x8 = 24, LUMA_PARTITION_16x12 = 27,
> LUMA_PARTITION_16x16 = 30, LUMA_PARTITION_16x32 = 33,
> + LUMA_PARTITION_16x64 = 36, LUMA_PARTITION_24x32,
> LUMA_PARTITION_32x8 = 40, LUMA_PARTITION_32x16 = 43,
> + LUMA_PARTITION_32x24 = 46, LUMA_PARTITION_32x32 = 49,
> LUMA_PARTITION_32x64 = 52, LUMA_PARTITION_48x64 = 55,
> + LUMA_PARTITION_64x16, LUMA_PARTITION_64x32,
> LUMA_PARTITION_64x48 = 60, LUMA_PARTITION_64x64,
> + NUM_LUMA_PARTITIONS = 64
> +};
>
I don't see the point to making a new sparse enum that lines up with the
existing Partitions enum. If you're not going to pack these so
NUM_LUMA_PARTITIONS is 24, why don't you just use Partitions directly?
--
Steve Borho
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/x265-devel/attachments/20131017/39ac6549/attachment.html>
More information about the x265-devel
mailing list