#[repr(i32)]pub enum ColorType {
Show 27 variants
Unknown = 0,
Alpha8 = 1,
RGB565 = 2,
ARGB4444 = 3,
RGBA8888 = 4,
RGB888x = 5,
BGRA8888 = 6,
RGBA1010102 = 7,
BGRA1010102 = 8,
RGB101010x = 9,
BGR101010x = 10,
BGR101010xXR = 11,
BGRA10101010XR = 12,
RGBA10x6 = 13,
Gray8 = 14,
RGBAF16Norm = 15,
RGBAF16 = 16,
RGBF16F16F16x = 17,
RGBAF32 = 18,
R8G8UNorm = 19,
A16Float = 20,
R16G16Float = 21,
A16UNorm = 22,
R16G16UNorm = 23,
R16G16B16A16UNorm = 24,
SRGBA8888 = 25,
R8UNorm = 26,
}Expand description
Describes how pixel bits encode color. A pixel may be an alpha mask, a grayscale, RGB, or ARGB.
The names of each variant implicitly define the channel ordering and size in memory. Due to historical reasons the names do not follow 100% identical convention, but are typically labeled from least significant to most significant.
Unless specified otherwise, a channel’s value is treated as an unsigned integer with a range of [0, 2^N-1] and this is mapped uniformly to a floating point value of [0.0, 1.0]. Some color types instead store data directly in 32-bit floating point (assumed to be IEEE), or in 16-bit “half” floating point values.
Note: By default, Skia operates with the assumption of a little-Endian system. The bit patterns shown in the documentation assume LE byte order.
Variants§
Unknown = 0
Unknown or unrepresentable as an SkColorType.
Alpha8 = 1
Single channel data (8-bit) interpreted as an alpha value. RGB are 0. Bits: [A:7..0]
RGB565 = 2
Three channel BGR data (5 bits red, 6 bits green, 5 bits blue) packed into a LE 16-bit word. Bits: [R:15..11 G:10..5 B:4..0]
ARGB4444 = 3
Four channel ABGR data (4 bits per channel) packed into a LE 16-bit word. Bits: [R:15..12 G:11..8 B:7..4 A:3..0]
RGBA8888 = 4
Four channel RGBA data (8 bits per channel) packed into a LE 32-bit word. Bits: [A:31..24 B:23..16 G:15..8 R:7..0]
RGB888x = 5
Three channel RGB data (8 bits per channel) packed into a LE 32-bit word. The remaining bits are ignored and alpha is forced to opaque. Bits: [x:31..24 B:23..16 G:15..8 R:7..0]
BGRA8888 = 6
Four channel BGRA data (8 bits per channel) packed into a LE 32-bit word. R and B are swapped relative to RGBA8888. Bits: [A:31..24 R:23..16 G:15..8 B:7..0]
RGBA1010102 = 7
Four channel RGBA data (10 bits per color, 2 bits for alpha) packed into a LE 32-bit word. Bits: [A:31..30 B:29..20 G:19..10 R:9..0]
BGRA1010102 = 8
Four channel BGRA data (10 bits per color, 2 bits for alpha) packed into a LE 32-bit word. R and B are swapped relative to RGBA1010102. Bits: [A:31..30 R:29..20 G:19..10 B:9..0]
RGB101010x = 9
Three channel RGB data (10 bits per channel) packed into a LE 32-bit word. The remaining bits are ignored and alpha is forced to opaque. Bits: [x:31..30 B:29..20 G:19..10 R:9..0]
BGR101010x = 10
Three channel BGR data (10 bits per channel) packed into a LE 32-bit word. The remaining bits are ignored and alpha is forced to opaque. R and B are swapped relative to RGB101010x. Bits: [x:31..30 R:29..20 G:19..10 B:9..0]
BGR101010xXR = 11
Three channel BGR data (10 bits per channel) packed into a LE 32-bit word. The remaining bits are ignored and alpha is forced to opaque. Instead of normalizing [0, 1023] to [0.0, 1.0] the color channels map to an extended range of [-0.752941, 1.25098]. Bits: [x:31..30 R:29..20 G:19..10 B:9..0]
BGRA10101010XR = 12
Four channel BGRA data (10 bits per channel) packed into a LE 64-bit word. Each channel is preceded by 6 bits of padding. Instead of normalizing [0, 1023] to [0.0, 1.0] the color and alpha channels map to an extended range of [-0.752941, 1.25098]. Bits: [A:63..54 x:53..48 R:47..38 x:37..32 G:31..22 x:21..16 B:15..6 x:5..0]
RGBA10x6 = 13
Four channel RGBA data (10 bits per channel) packed into a LE 64-bit word. Each channel is preceded by 6 bits of padding. Bits: [A:63..54 x:53..48 B:47..38 x:37..32 G:31..22 x:21..16 R:15..6 x:5..0]
Gray8 = 14
Single channel data (8-bit) interpreted as a grayscale value (e.g. replicated to RGB). Bits: [G:7..0]
RGBAF16Norm = 15
Four channel RGBA data (16-bit half-float per channel) packed into a LE 64-bit word. Values are assumed to be in [0.0,1.0] range, unlike RGBAF16. Bits: [A:63..48 B:47..32 G:31..16 R:15..0]
RGBAF16 = 16
Four channel RGBA data (16-bit half-float per channel) packed into a LE 64-bit word. This has extended range compared to RGBAF16Norm. Bits: [A:63..48 B:47..32 G:31..16 R:15..0]
RGBF16F16F16x = 17
Three channel RGB data (16-bit half-float per channel) packed into a LE 64-bit word. The last 16 bits are ignored and alpha is forced to opaque. Bits: [x:63..48 B:47..32 G:31..16 R:15..0]
RGBAF32 = 18
Four channel RGBA data (32-bit float per channel) packed into a LE 128-bit word. Bits: [A:127..96 B:95..64 G:63..32 R:31..0]
R8G8UNorm = 19
Two channel RG data (8 bits per channel). Blue is forced to 0, alpha is forced to opaque. Bits: [G:15..8 R:7..0]
A16Float = 20
Single channel data (16-bit half-float) interpreted as alpha. RGB are 0. Bits: [A:15..0]
R16G16Float = 21
Two channel RG data (16-bit half-float per channel) packed into a LE 32-bit word. Blue is forced to 0, alpha is forced to opaque. Bits: [G:31..16 R:15..0]
A16UNorm = 22
Single channel data (16 bits) interpreted as alpha. RGB are 0. Bits: [A:15..0]
R16G16UNorm = 23
Two channel RG data (16 bits per channel) packed into a LE 32-bit word. B is forced to 0, alpha is forced to opaque. Bits: [G:31..16 R:15..0]
R16G16B16A16UNorm = 24
Four channel RGBA data (16 bits per channel) packed into a LE 64-bit word. Bits: [A:63..48 B:47..32 G:31..16 R:15..0]
SRGBA8888 = 25
Four channel RGBA data (8 bits per channel) packed into a LE 32-bit word. The RGB values are assumed to be encoded with the sRGB transfer function. Bits: [A:31..24 B:23..16 G:15..8 R:7..0]
R8UNorm = 26
Single channel data (8 bits) interpreted as red. G and B are forced to 0, alpha is forced to opaque. Bits: [R:7..0]