Skip to content

floatToRGBAEncoding.md

Latest commit

 

History

History
197 lines (167 loc) · 4.68 KB

floatToRGBAEncoding.md

File metadata and controls

197 lines (167 loc) · 4.68 KB

Float to RGBA encoding/Decoding methods test

This document show thes results of various tested methods to encoding/decoding float to RGBA value

ENTRY VALUE : 
double PID = 3.1415926535897932384626433832795028841971693993751058209749445923078164062;

Gary Ruddock (Skytiger) approach

Link

ARGB32

Double

entry 			:3.1415927410125732
retreived 		:3.1415927409952498
norm entry 		:0.5
norm retreived 	:0.49999999999724287
ARGB value 		:2139062143

Float

entry 			:0.45663998
retreived 		:0.45663992
norm entry 		:0.45663998
norm retreived 	:0.45663992
ARGB value 		:2138337538

ARGB24

Double

entry 			:3.1415927410125732
retreived 		:3.1415925522730985
norm entry 		:0.5
norm retreived 	:0.49999996996118046
ARGB value 		:-8421505

Float

entry 			:0.45663998
retreived 		:0.45663995
norm entry 		:0.45663998
norm retreived 	:0.45663995
ARGB value 		:-9146109

ARGB24 Method 2

Double

entry 			:3.1415927410125732
retreived 		:3.141592554493359
norm entry 		:0.5
norm retreived 	:0.4999999703145459
ARGB value 		:-8388609

Float

entry 			:0.45663998
retreived 		:0.45663992
norm entry 		:0.45663998
norm retreived 	:0.45663992
ARGB value 		:-9116070

ARGB16 Method (custom implementation)

Double

entry 			:3.1415927410125732
retreived 		:3.1416414342823544
norm entry 		:0.5
norm retreived 	:0.5000077497743654
ARGB value 		:-8421377

Float

entry 			:0.45663998
retreived 		:0.4566552
norm Entry 		:0.45663998
norm retreived 	:0.4566552
ARGB value 		:-9145857

Aras Pranckevičius approach

Link

ARGB32

Double

entry 			:3.1415927410125732
retreived 		:3.116855791764753
norm entry 		:0.5
norm retreived 	:0.4960629921232488
ARGB value 		:2122219134

Float

entry 			:0.45663998
retreived 		:0.45663992
norm entry 		:0.45663998
norm retreived 	:0.45663992
ARGB value 		:2138337538

Chris Wellons approach

This approach is useful when you need to encode an 2D x,y position into en RGBA texture where x >> RG et y >> BA Link

ARGB16 (two value encoded int two channels)

Double

entry 			:3.1415927410125732
retreived 		:3.129224825983652
norm entry 		:0.5
norm retreived 	:0.4980315852421828
ARGB value 		:-16744449

Float

entry 			:0.45663998
retreived 		:0.45489427
norm entry 		:0.45663998
norm retreived 	:0.45489427
ARGB value 		:-16747265

Conclusion

  1. Gary Ruddock's methods seems to be the more accurate by encoding into the 32 bits (8bits per channel) and using a factor and mask approach.
  2. Aras Pranckevičius' approach seems quite inefficient fo JAVA by losing values at its first digit
  3. Chris Wellons' approach seems also inefficient but quite usefull if we want to store 1 value per 2 channel such as x >> RG and y >> BA

Test1 on Double

PID = 3.1415927410125732
Entry = PID/TWO_PI (norm PID)
Retreived = decode * TWO_PI
NAME VALUES
ENTRY VALUE (double) 3.1415927410125732
Gary Ruddock ARGB32 3.1415927409952498
Gary Ruddock ARGB24-1 3.1415925522730985
Gary Ruddock ARGB24-2 3.141592554493359
Gary Ruddock ARGB16 3.1416414342823544
Aras Pranckevičiu ARGB32 3.116855791764753
Chris Wellons ARGB16 3.129224825983652

Test2 on Double

PID = 98742.0/216236.0 = 0.45663997530937195
Entry = PID
Retreived = decode
NAME VALUES
ENTRY VALUE (double) 0.45663997530937195
Gary Ruddock ARGB32 0.4566399754117086
Gary Ruddock ARGB24-1 0.4566399349694513
Gary Ruddock ARGB24-2 0.45663991322699965
Gary Ruddock ARGB16 0.4566551930526745
Aras Pranckevičiu ARGB32 0.452702967774219
Chris Wellons ARGB16 0.454894325238035

Test2 on Float

PID = 98742.0/216236.0 = 0.45663998
Entry = PID
Retreived = decode
NAME VALUES
ENTRY VALUE (double) 0.45663998
Gary Ruddock ARGB32 0.45663992
Gary Ruddock ARGB24-1 0.45663995
Gary Ruddock ARGB24-2 0.45663992
Gary Ruddock ARGB16 0.4566552
Aras Pranckevičiu ARGB32 0.45663992
Chris Wellons ARGB16 0.45489427

References :

https://www.gamedev.net/forums/topic/442138-packing-a-float-into-a-a8r8g8b8-texture-shader/#2936108 http://diaryofagraphicsprogrammer.blogspot.com/2009/10/bitmasks-packing-data-into-fp-render.html