RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-28 02:07:35 +00:00
Code Issues Activity Actions

LibVideo/VP9: Pre-calculate the quantizers at the start of each frame

Browse source 
Quantizers are a constant for the whole frame, except when segment
features override them, in which case they are a constant per segment
ID. We take advantage of this by pre-calculating those after reading
the quantization parameters and segmentation features for a frame.
This results in a small 1.5% improvement (~12.9s -> ~12.7s).
This commit is contained in:
Zaggy1024 2023-04-17 00:27:13 -05:00 • committed by Tim Flynn
parent 094b0d8a78
commit 90c0e1ad8f
6 changed files with 104 additions and 70 deletions
Download patch file Download diff file Expand all files Collapse all files

6
Userland/Libraries/LibVideo/VP9/Decoder.h
View file

@ -73,11 +73,11 @@ private:
/* (8.6) Reconstruction and Dequantization */
// Returns the quantizer index for the current block
static u8 get_base_quantizer_index(BlockContext const&);
static u8 get_base_quantizer_index(SegmentFeatureStatus alternative_quantizer_feature, bool should_use_absolute_segment_base_quantizer, u8 base_quantizer_index);
// Returns the quantizer value for the dc coefficient for a particular plane
static u16 get_dc_quantizer(BlockContext const&, u8 plane);
static u16 get_dc_quantizer(u8 bit_depth, u8 base, i8 delta);
// Returns the quantizer value for the ac coefficient for a particular plane
static u16 get_ac_quantizer(BlockContext const&, u8 plane);
static u16 get_ac_quantizer(u8 bit_depth, u8 base, i8 delta);
// (8.6.2) Reconstruct process
DecoderErrorOr<void> reconstruct(u8 plane, BlockContext const&, u32 transform_block_x, u32 transform_block_y, TransformSize transform_block_size, TransformSet);