Fixed Point vs Floating Point
What does ‘Fixed 0.8’ mean???
- Fixed means that the decimal point is assumed fixed in one position
- out of N bits, K bits used for integer portion, M bits used for fractional portion (K.M)
- Can be signed or unsigned, also use a bit for S for sign information (SK.M)
- View color information as a number with 0 bits for magnitude, 8 bits for fraction
- Color ranges from [0, 1)
- “11111111” => .99609 (255/256)
- “00000001” => .00390 (1/256)