<<However, if you're going to take it all back down to 8-bit in the end then you might as well just work in 8-bit all the way.>> Err, Herschel, didn't we just have a post from someone for whom that was patently not the case? Having decided to allow the camera to throw away everything more than 8-bits (ie: shoot jpeg) there is no way back. ;o) I'm closer to agreeing with you as a practical solution for in-computer-processing ... but a fundamental principle of data collection for me has always been to round off as late as possible. I had learned that mantra by the time I reached my teens. Current display hardware can only show 256 OR LESS levels per colour channel. Current printer hardware can, by and large, only handle 24-bit images. Does that mean we should stick in the past? One HUGE advantage of HDR storage may well render <G> the current array of color spaces redundant. Permanently changing image DATA -could be a thing of the past. Leave the numbers pretty much untouched and instead code in only how they are to be handled. Posterisation: gone. When you decide you want to print an image: at that point a copy of the data is transformed to the numbers the printer can handle. If it's for web display, ditto ... HDR workflow removes the compromises - it's just a little slower. <<I tell my students to stick to a level slightly higher than the weakest link in the processing chain.>> Good advice for the present <<And I agree that the sooner we get with DNG, the better for all concerned.>> I don't see that happening because there are too many vested interests. You don't own your image data: you licence the right to decode it from Nikon, Canon, Fuji ... You BUY their software or at the least BUY software from someone else who is paying them for the right to decode it :o) Bob -- Whatever you Wanadoo: http://www.wanadoo.co.uk/time/ This email has been checked for most known viruses - find out more at: http://www.wanadoo.co.uk/help/id/7098.htm
