Documentation
¶
Overview ¶
Package color provides color space types and conversions for gg.
Package color provides fast color space conversion using lookup tables.
The lookup tables (LUT) provide O(1) sRGB ↔ Linear conversions, replacing expensive math.Pow calls with simple array lookups. This is critical for performance in alpha blending operations.
sRGB is the standard color space for images and displays, but blending operations should be performed in linear space for physically correct results.
References:
- sRGB specification: https://www.w3.org/Graphics/Color/sRGB
- GPU Gems 3, Chapter 24: https://developer.nvidia.com/gpugems/gpugems3/part-iv-image-effects/chapter-24-importance-being-linear
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func LinearToSRGB ¶
LinearToSRGB converts a linear component to sRGB (OETF - Opto-Electronic Transfer Function). Formula: if l <= 0.0031308: l*12.92; else: 1.055*pow(l, 1/2.4)-0.055 Input and output are in range [0,1].
func LinearToSRGBFast ¶ added in v0.5.0
LinearToSRGBFast converts linear float32 to sRGB byte using lookup table.
Input is clamped to [0.0, 1.0] range automatically. Uses 12-bit precision (4096 entries) which is more than sufficient for 8-bit sRGB output.
Example:
s := LinearToSRGBFast(0.5) // 188 (not 128!)
func LinearToSRGBSlow ¶ added in v0.5.0
LinearToSRGBSlow converts linear float32 to sRGB byte using math.Pow.
This is the reference implementation, ~15x slower than the LUT version. Used for testing and verification only.
func SRGBToLinear ¶
SRGBToLinear converts an sRGB component to linear (EOTF - Electro-Optical Transfer Function). Formula: if s <= 0.04045: s/12.92; else: pow((s+0.055)/1.055, 2.4) Input and output are in range [0,1].
func SRGBToLinearFast ¶ added in v0.5.0
SRGBToLinearFast converts sRGB byte to linear float32 using lookup table.
This is ~20x faster than computing with math.Pow for each pixel. Used in blend operations that require linear color space.
Example:
r := SRGBToLinearFast(128) // ~0.2159 (not 0.5!)
func SRGBToLinearSlow ¶ added in v0.5.0
SRGBToLinearSlow converts sRGB byte to linear float32 using math.Pow.
This is the reference implementation, ~20x slower than the LUT version. Used for testing and verification only.
Types ¶
type ColorF32 ¶
type ColorF32 struct {
R, G, B, A float32
}
ColorF32 represents a color with float32 components in [0,1]. RGB components are in the color space indicated by context. Alpha is always linear (never gamma-encoded).
func LinearToSRGBColor ¶
LinearToSRGBColor converts a full color from linear to sRGB space. Only RGB components are converted; alpha remains linear (never gamma-encoded).
func SRGBToLinearColor ¶
SRGBToLinearColor converts a full color from sRGB to linear space. Only RGB components are converted; alpha remains linear (never gamma-encoded).
type ColorSpace ¶
type ColorSpace uint8
ColorSpace represents a color space.
const ( // ColorSpaceSRGB represents the standard sRGB color space. ColorSpaceSRGB ColorSpace = iota // ColorSpaceLinear represents the linear RGB color space. ColorSpaceLinear )
Source Files