> From: Tom Baker [mailto:tbaker1328@...] > > If one were 'adding' the values together, you'd be on the right > track. However, it is per color channel that these values are > applied. So, 8 bits is 256 values per channel. However, if you > have a value of 220 in a channel and convert it to 16bit, you > still don't have any more real data than you did before. You > only have a 16 bit value with the same real value. From a strictly numeric point of view, eight bits is eight bits. The original poster thought he could somehow get 24 bits of "resolution". My point was that you do indeed get a little bit more resolution, perceptually, because the amount of quantization noise you perceive in an 8bpc color image is probably better than a 9-bit B&W image, for the reason I explained. A more extreme example of this is that an inkjet print, even a B&W one, made with color inks will have less visible noise than a black ink only print, even though each ink "channel" is still only one bit. -- Ciao, Paul D. DeRocco Paul mailto:pderocco@...
Message
RE: [Digital BW] 8 bit color -> 16 bit grayscale conversion
2003-12-12 by Paul D. DeRocco
Attachments
- No local attachments were found for this message.