On Mar 6, 2005, at 2:22 AM, Shel Belinkoff wrote:
OK, that makes sense. But then, by the same token, wouldn't it make just
as much sense to scan using a profile with an even wider gamut?
You can do this if you want, but there are tradeoffs involved...
There's no point using a colour space that greatly exceeds the capabilities of your scanner as you end up being able represent colours that your scanner can't even produce in the first place.
As a result, a wider gamut working space means worse tonality for the same bit-depth as the same range of numbers have to represent a larger range of colours, so the discrete colour change from one value to the next is larger. Practically speaking, this is only really important if you work in 24-bit colour. 48-bit colour is a good thing despite the increase in CPU, memory and disk space requirements for processing and archiving.
Most raw photos will fit within sRGB. If you're making adjustments that affect the saturation you may need to be careful about gamut clipping. If you keep today's printing technology in mind then you'll be limited if a new process comes along... but only if you actually need highly saturated colours.
Also bear in mind that not many monitors are capable of displaying much outside of sRGB.
Choosing the perfect working space is a real can of worms and it depends on your equipment and your long-term intentions. Luckily, for most people, it doesn't make a big difference as long as the process works (ie calibrated screens, files saved with embedded profiles and the lab being able to cope with them).
Cheers,
- Dave
http://www.digistar.com/~dmann/
