Comment Re: Bad summary (Score 1) 130
32-bit per channel graphics is a bit of a misnomer. It uses a 32-bit float for each channel instead of the usual 8/10 bit unsigned integer. Furthermore, the value must be between 0 and 1 in order to display as anything other than 0% or 100% of that color. Its real utility is in high dynamic range situations, where overshooting the mark actually means something. In ordinary situations you just get the equivalent of an unsigned integer with the same bits as the mantissa, which for single-precision (32-bit) floats is 23 bits. For half-precision (16 bit) floats, you get 10 bits. Half-precision floats are much more common than single-precision right now in graphics, but give it time. Windows 10 and 11 default to half-precision float per channel graphics, IIRC. Graphics cards are tuned for doing half-precision calculations, as well.