[Openicc] meta data in test chart

[Graeme Gill]

Alastair M. Robinson wrote:

> The last two variants are easy to handle: tagged images should be
> converted to the printer's profile, while DeviceXXX images should be
> passed through unmolested (provided they match the driver's colourspace
> - if not the job needs to fail; silently converting would make
> diagnosing problems much harder.)

In my experience, this doesn't often meet user expectations. Maybe
applications have improved over recent years, but all too often
in the past applications have created "RGB" and "CMYK" data
that is labelled "DeviceXXX", when they really mean "a normal
device", not the actual device you have. "normal" means
"like a display" or "like a printing press". So the only
way of meeting this user expectation is to translate
DeviceXXX into some reasonable spaces (possibly user
selected), such as sRGB, SWOP etc. This reality is reflected
in the PostScript and PDF standards, PostScript with the "UseCIEColor"
flag, and PDF with the equivalent "DefaultXXXX" dictionaries.

This means that "DeviceXXX" is not a way of getting profiling
chart values, or "pre-matched" values through to the device.
Some other mechanism is required. You might be tempted by
Separation or DeviceN space, but note that these are subject
to PS/PDF transfer functions, and even these may be modified
by a RIP if it's operating in a "proofing" or "preview" mode
(ie. it's emulating a different output device).

Graeme Gill.