Support Adobe's DCP which expect ACR curve

[Beep6581]

Originally reported on Google Code with ID 2836

Adobe's own profiles typically lack a curve and then expect the Adobe Camera Raw default
curve to be applied.

Small isolated patch, done and committed.

It will add the ACR default curve to Adobe's own profiles and then it can be activated
normally using the tone curve check box.

Reported by torger@ludd.ltu.se on 2015-07-09 19:49:25

[Beep6581]

Cool! Could you add a note about it to RawPedia? Else I can do it next week as I'll
be away till Sunday.

I'm curious, is this ACR curve always the same? How did you get it?

Reported by entertheyoni on 2015-07-09 19:58:40

• Status changed: NeedsDocumentation

[Beep6581]

The ACR curve is found in the DNG reference code. It's always the same, but not all
of Adobe profiles use it. If they don't the curve is included in the profile though.
So AFAIK the only implicit curve is the ACR default and that's the one I just added
to dcp.cc.

I think the rawpedia docs on third-party profiles needs to be restructured a bit with
proper headlines, now it's just a big block of text. We need to add description of
the new DCP checkboxes too, and the baseline exposure stuff if I just get that in.

Reported by torger@ludd.ltu.se on 2015-07-09 20:45:52

[Beep6581]

I've refreshed the rawpedia docs a bit now.

Reported by torger@ludd.ltu.se on 2015-07-09 21:25:43

• Status changed: FixedPendingConfirmation

[Beep6581]

Would it be possible to make the curve available for profiles other than Adobe DCPs
as well? I think the DCPs created by X-Rite software (i.e. most of those found in RT)
probably also expect this ACR curve, as the software makers probably expect their profiles
to be used with Lightroom.

Maybe a checkbox "use default curve" that is auto-activated if it detects Adobe's own
profiles but that can still be activated manually for other profiles would be a solution?

Reported by stefan.ittner on 2015-07-10 06:56:21

[Beep6581]

You can always add your own curve via the curve tool.

I don't know how X-rite designs their profiles, but I would expect they are designed
for a linear curve, and adding an S-curve is just a "look" to them.

If the profile is specifically designed for a curve then colors wouldn't match when
used with a linear curve and AFAIK people have tested X-rite profiles for accuracy
and got good results?

If RT has profiles that actually are designed for ACR default curve it's really against
the design of RT, which in it's native mode is intended to be used with colorimetric
profiles. With recent patches it can be used properly with DCP's that provide a look
too though.

I consider adding-a-curve-despite-the-profile-lacks-it is one of those Adobe-specific
hacks Adobe has made to their "standard" DNG/DCP formats and we only support it because
Adobe is so dominant in their position and people are likely to use Adobe profiles.
When it comes to making profiles for RT I think we should be able to design them in
the way RT wants them.

Reported by torger@ludd.ltu.se on 2015-07-10 12:31:34

[Beep6581]

Of course we can add our own curves, but it would be nice to have the de-facto standard
base curve as a reference option. I wouldn't know how to re-create it exactly with
RT's tools.
I think the X-Rite software creates the profiles with LR usage in mind, at least it
defaults to saving them in the respective LR folder. Wouldn't LR use the same standard
base curve with those X-Rite DCPs as well as it would with the Adobe DCPs?

Reported by stefan.ittner on 2015-07-10 13:32:22

[Beep6581]

Yes it probably will. I haven't analyzed the X-rite software though. If I would guess
I'd say it makes a profile from CC24 target which will be slightly desaturated by nature
(as the target is matte) and then applying an Adobe curve will look sort of good and
therefore they have not cared to make any pre-processing for the curve.

I have an adobe.rtc on a different computer which quite well matches their curve. I
can post it here when I get to it.

Reported by torger@ludd.ltu.se on 2015-07-11 07:23:01

[Beep6581]

Reported by torger@ludd.ltu.se on 2015-08-05 20:41:23

• Status changed: Fixed