[Openicc] ProPhoto, ICCv4, was: meta data in test chart

[Chris Murphy]

On Jan 27, 2011, at 6:08 PM, Graeme Gill wrote:

> Chris Murphy wrote:
> 
>> values are imaginary colors. But yes, I'm understanding what you're saying. While I
>> still think it's correct to say one is rendering into ProPhoto RGB using any Raw
>> processing application, what hinders this from being fully utilized is we're making
>> these rendering choices based on a display that is distinctly not capable of showing us
>> everything in ProPhoto RGB (or even much of it).
> 
> Nothing will ever be able to render the whole of ProPhoto, any more than the
> whole of ICC L*a*b* or scRGB can be displayed. It has a blue primary
> with an XYZ of 3.1339 0.0091 82.4895. Look at the Y :- that means it's black!

Yes I know. By "being fully utilized" I just mean we can't really see how we're rendering a Raw image into ProPhoto when out view into that space is typically an sRGB gamut display, or at best an Adobe RGB gamut display. ProPhoto's shape has some negatives, as you mention, but the positives are that it's still gray balanced, and is quite uniform. Ordinary images don't come close to being encoded in the extreme areas of ProPhoto RGB - those extreme areas exist by the choosing of primaries that allow encoding of desired complementary colors. The negatives are a side effect, not a design goal.

The idea of the PRMG application to ProPhoto is to dispense with those areas of the color space that are not only imaginary but just not likely at all to encode real objects captured by a digital camera.


> So given a calibrated/profiled wide gamut display, the author of an image
> can choose whatever appearance they like within that gamut. In the past,
> they would be forced to render it down to a small gamut such as sRGB, or
> a printer gamut if they were going to deliver it as a print. But now,
> as long as it is delivered electronically, and as long as they are using
> a wide gamut encoding space, nothing is forcing them to deal with that.
> 
> But if they choose to create a very wide gamut image, somewhere down
> the track it's going to be squeezed into smaller gamut output devices.
> If they don't create versions of the image that they've hand crafted to
> fit within those smaller gamuts, it will be done for them by the
> technology, either crudely (ie. clipping), wrongly (gamut mapped
> from the wrong/too large source gamut), or more cleverly (gamut mapped
> to just fit that image in the destination).

Yes, all true. That's why I teach to deliver output capable TIFFs and JPEGs, not full rez, 16bpc ProPhoto RGB TIFFs let alone the Raw file itself.


Chris