dany/libwebp
1
0
Fork 0
mirror of https://github.com/webmproject/libwebp.git synced 2025-07-13 06:24:27 +02:00
Code Issues Actions Packages Projects Releases Wiki Activity

Add YUV420 riskiness metric.

Browse Source 
Basic version for 8 bit only and C only (no simd).

Hidden behind a compile flag because the precompiled table is
fairly large (114kiB) and more than triples the size of the library.

Change-Id: I165c78e863df6a17b32f578bdbffe3adda9ac1d0
This commit is contained in:
Maryla
2023-11-21 18:14:41 +01:00
parent 89c5b91746
commit d7a0506dcc
10 changed files with 6450 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
sharpyuv/sharpyuv.h
View File

@ -53,7 +53,7 @@ extern "C" {
// SharpYUV API version following the convention from semver.org
#define SHARPYUV_VERSION_MAJOR 0
#define SHARPYUV_VERSION_MINOR 4
#define SHARPYUV_VERSION_MINOR 5
#define SHARPYUV_VERSION_PATCH 0
// Version as a uint32_t. The major number is the high 8 bits.
// The minor number is the middle 8 bits. The patch number is the low 16 bits.
@ -164,6 +164,36 @@ SHARPYUV_EXTERN int SharpYuvConvertWithOptions(
int u_stride, void* v_ptr, int v_stride, int yuv_bit_depth, int width,
int height, const SharpYuvOptions* options);
// Computes a score between 0 and 100 which represents the risk of having visual
// quality loss from converting an RGB image to YUV420.
// A low score, typically < 40, means there is a low risk of artifacts from
// chroma subsampling and a simple averaging algorithm can be used instead of
// the more expensive SharpYuvConvert function.
// A medium score, typically >= 40 and < 70, means that simple chroma
// subsampling will produce artifacts and it may be advisable to use the more
// costly SharpYuvConvert for YUV420 conversion.
// A high score, typically >= 70, means there is a very high risk of artifacts
// from chroma subsampling even with SharpYuvConvert, and best results might be
// achieved by using YUV444.
// If not using SharpYuvConvert, a threshold of about 50 can be used to decide
// between (simple averaging) 420 and 444.
// r_ptr, g_ptr, b_ptr: pointers to the source r, g and b channels. Should point
// to uint8_t buffers if rgb_bit_depth is 8, or uint16_t buffers otherwise.
// rgb_step: distance in bytes between two horizontally adjacent pixels on the
// r, g and b channels. If rgb_bit_depth is > 8, it should be a
// multiple of 2.
// rgb_stride: distance in bytes between two vertically adjacent pixels on the
// r, g, and b channels. If rgb_bit_depth is > 8, it should be a
// multiple of 2.
// rgb_bit_depth: number of bits for each r/g/b value. Only a value of 8 is
// currently supported.
// width, height: width and height of the image in pixels
// Returns 0 on failure.
SHARPYUV_EXTERN int SharpYuvEstimate420Risk(
const void* r_ptr, const void* g_ptr, const void* b_ptr, int rgb_step,
int rgb_stride, int rgb_bit_depth, int width, int height,
const SharpYuvOptions* options, float* score);
// TODO(b/194336375): Add YUV444 to YUV420 conversion. Maybe also add 422
// support (it's rarely used in practice, especially for images).