Further reduce memory to decode lossy+alpha images

Earlier such images were using roughly 9 * width * height bytes for
decoding. Now, they take 6 * width * height memory.

Change-Id: Ie4a681ca5074d96d64f30b2597fafdca648dd8f7
This commit is contained in:
Urvang Joshi
2013-05-13 16:24:49 -07:00
parent 5199eab516
commit 64c844863a
4 changed files with 253 additions and 166 deletions

View File

@ -33,6 +33,13 @@ void VP8LInverseTransform(const struct VP8LTransform* const transform,
int row_start, int row_end,
const uint32_t* const in, uint32_t* const out);
// Similar to the static method ColorIndexInverseTransform() that is part of
// lossless.c, but used only for alpha decoding. It takes uint8_t (rather than
// uint32_t) arguments for 'src' and 'dst'.
void VP8LColorIndexInverseTransformAlpha(
const struct VP8LTransform* const transform, int y_start, int y_end,
const uint8_t* src, uint8_t* dst);
// Subtracts green from blue and red channels.
void VP8LSubtractGreenFromBlueAndRed(uint32_t* argb_data, int num_pixs);