Functions to rescale the color values by dividing by 255. More...
Collaboration diagram for Rescale:Enumerations | |
| enum | RescaleType { Div255_Round, Div255_Round_Approx, Div255_Floor, Div256_Round, Div256_Floor } |
| How to carry out the rescaling after performing the alpha overlay. More... | |
Functions | |
| uint8x8_t | div256_floor (uint16x8_t x) |
| This is a flooring division by 256, which is close enough to 255, but the result may be one bit too small. More... | |
| uint8x8_t | div256_round (uint16x8_t x) |
| This is a rounding division by 256, which is close enough to 255, but the result may be one bit too small. More... | |
| uint8x8_t | div255_floor (uint16x8_t x) |
| This is an exact flooring division by 255, this is the correct divisor, but requires a little bit more code, and doesn't round the result. More... | |
| uint8x8_t | div255_round_approx (uint16x8_t x) |
| This is an approximation of a rounding division by 255, this is the correct divisor, and it rounds the result correctly 99.80621337890625% of the time. More... | |
| uint8x8_t | div255_round (uint16x8_t x) |
| This is an exact rounding division by 255, this is the correct divisor, and the result is rounded correctly in 100% of the cases. More... | |
| uint8_t | div256_floor (uint16_t x) |
| This is a flooring division by 256, which is close enough to 255, but the result may be one bit too small. More... | |
| uint8_t | div256_round (uint16_t x) |
| This is a rounding division by 256, which is close enough to 255, but the result may be one bit too small. More... | |
| uint8_t | div255_floor (uint16_t x) |
| This is an exact flooring division by 255, this is the correct divisor, but requires a little bit more code, and doesn't round the result. More... | |
| uint8_t | div255_round (uint16_t x) |
| This is an exact rounding division by 255, this is the correct divisor, and the result is rounded correctly in 100% of the cases. More... | |
| uint8_t | div255_round_approx (uint16_t x) |
| Uses div255_round, and is exact. More... | |
| template<RescaleType rescale_type = RescaleType::Div255_Round> | |
| uint8x8_t | rescale (uint16x8_t x) |
| Rescale the 16-bit color product by dividing by 255 or an approximation thereof. More... | |
| template<RescaleType rescale_type = RescaleType::Div255_Round> | |
| uint8_t | rescale (uint16_t x) |
| Rescale the 16-bit color product by dividing by 255 or an approximation thereof. More... | |
Detailed Description
Functions to rescale the color values by dividing by 255.
Enumeration Type Documentation
◆ RescaleType
|
strong |
How to carry out the rescaling after performing the alpha overlay.
The product has to be divided by 255 to be exact, but other divisions can be used for (small) performance gains.
| Enumerator | |
|---|---|
| Div255_Round | Exact rounding division by 255. See div255_round |
| Div255_Round_Approx | Approximate rounding division by 255, correct in 99.806% of the cases. |
| Div255_Floor | Exact flooring division by 255. See div255_floor |
| Div256_Round | Fast, rounding division by 256. See div256_round |
| Div256_Floor | Fast, flooring division by 256. See div256_floor |
Definition at line 12 of file rescale_type.hpp.
Function Documentation
◆ div256_floor() [1/2]
|
inline |
This is a flooring division by 256, which is close enough to 255, but the result may be one bit too small.
Definition at line 16 of file rescale.hpp.
◆ div256_round() [1/2]
|
inline |
This is a rounding division by 256, which is close enough to 255, but the result may be one bit too small.
Definition at line 22 of file rescale.hpp.
◆ div255_floor() [1/2]
|
inline |
This is an exact flooring division by 255, this is the correct divisor, but requires a little bit more code, and doesn't round the result.
Definition at line 37 of file rescale.hpp.
◆ div255_round_approx() [1/2]
|
inline |
This is an approximation of a rounding division by 255, this is the correct divisor, and it rounds the result correctly 99.80621337890625% of the time.
That is, the result is one unit too large in 127 out of all 2¹⁶ possible input values.
This function is just as cheap as the div255_floor function.
Definition at line 57 of file rescale.hpp.
◆ div255_round() [1/2]
|
inline |
This is an exact rounding division by 255, this is the correct divisor, and the result is rounded correctly in 100% of the cases.
(Compared to std::round(x / 255.0))
The result is incorrect for very large numbers where the quotient would overflow an 8-bit integer.
Definition at line 76 of file rescale.hpp.
◆ div256_floor() [2/2]
|
inline |
This is a flooring division by 256, which is close enough to 255, but the result may be one bit too small.
Definition at line 153 of file rescale.hpp.
◆ div256_round() [2/2]
|
inline |
This is a rounding division by 256, which is close enough to 255, but the result may be one bit too small.
Definition at line 159 of file rescale.hpp.
◆ div255_floor() [2/2]
|
inline |
This is an exact flooring division by 255, this is the correct divisor, but requires a little bit more code, and doesn't round the result.
Definition at line 165 of file rescale.hpp.
◆ div255_round() [2/2]
|
inline |
This is an exact rounding division by 255, this is the correct divisor, and the result is rounded correctly in 100% of the cases.
(Compared to std::round(x / 255.0))
Definition at line 175 of file rescale.hpp.
◆ div255_round_approx() [2/2]
|
inline |
Uses div255_round, and is exact.
Definition at line 184 of file rescale.hpp.
◆ rescale() [1/2]
|
inline |
Rescale the 16-bit color product by dividing by 255 or an approximation thereof.
- Parameters
-
x The 16-bit product to rescale.
- Template Parameters
-
rescale_type The rescaling mode to use. See RescaleType.
Definition at line 200 of file rescale.hpp.
◆ rescale() [2/2]
|
inline |
Rescale the 16-bit color product by dividing by 255 or an approximation thereof.
- Parameters
-
x The 16-bit product to rescale.
- Template Parameters
-
rescale_type The rescaling mode to use. See RescaleType.
Definition at line 214 of file rescale.hpp.
