Anyone using the "reference" role?

117 views
Skip to first unread message

Matt Plec

unread,
Nov 2, 2015, 6:23:35 PM11/2/15
to OpenColorIO Users
We're debating how to set the reference role, but although it appears to be intended for reference material -- ie stuff you find on the internet  --  in the config we define to_reference and from_reference sections in color space definitions which would imply it's the space that the colorspaces are defined with reference to (which is typically not screen-ready srgb).

Is anyone aware of any 3rd party (commercial or open source) tools using the 'reference' role and assuming a particular meaning?

Anyone willing to 'fess up with what you're setting it to in your own configs and if you've had any unpleasant surprises doing that?

Thanks,
Matt

Steve Agland

unread,
Nov 4, 2015, 2:30:58 PM11/4/15
to ocio-...@googlegroups.com

It's intended as the latter meaning: a reference space against which the other spaces are defined. We use the linear ACES2065-1 space. I don't know of any tools "using" it for any particular purpose other than that.

You can define your own roles if you need to, so you define something else - say, "web" - to be the assumed color space of reference images found online.

Cheers

Steve

--
You received this message because you are subscribed to the Google Groups "OpenColorIO Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ocio-users+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Kevin Wheatley

unread,
Nov 4, 2015, 3:58:57 PM11/4/15
to ocio-...@googlegroups.com


Sent on the go...

On 4 Nov 2015, at 19:30, Steve Agland <sag...@gmail.com> wrote:

It's intended as the latter meaning: a reference space against which the other spaces are defined. We use the linear ACES2065-1 space. I don't know of any tools "using" it for any particular purpose other than that.


We do the same but with ACEScg (or historically linearised camera space). Though depending on your list of applications and their level of OCIO support and version the actual reference space might not make a huge difference, the more complete the support the less difference it makes, with simplistic support we found ACEScg worked better (less conversions are needed)

We also have some custom roles used for specifying grading colour spaces, etc. We do this as we have grades provided by our clients as well as our own internal ones and these are not always in the same space.

Kevin


Matt Plec

unread,
Nov 4, 2015, 4:23:59 PM11/4/15
to ocio-...@googlegroups.com
Ah, brilliant, thanks guys.

In the config roles section (http://opencolorio.org/userguide/config_syntax.html#roles) it says: "reference - Colorspace used for reference imagery (e.g sRGB images from the internet)". I probably won't be the first or last to be confused. I'll submit a pull request with a revision to that. Anyone in particular pushing updates to the opencolorio.org website?



denser...@gmail.com

unread,
Nov 10, 2015, 3:54:04 PM11/10/15
to OpenColorIO Users
Hmm, the OCIO 1.0 config does not use the reference colorspace (ACES2065-1) for the reference role; it uses Utility - Raw. Nuke-default uses linear, and spi-vfx uses lnf (linear) which are both the reference colorspace. So that would imply the OCIO config is using an odd definition for the value. Shouldn't the OCIO be 2065-1 (or even something like sRGB for "reference images"?)

denser...@gmail.com

unread,
Nov 12, 2015, 11:43:19 AM11/12/15
to OpenColorIO Users
*ACES 1.0 config (not OCIO 1.0)

Matt Plec

unread,
Nov 12, 2015, 1:22:03 PM11/12/15
to ocio-...@googlegroups.com
Utility raw, at least in HPD's aces_1.0.1, is a no-op so it's effectively the same thing. From those and the previous comments it seems that in typical practice it's the to/from reference, not "stuff I found on google images for reference". I've submitted a pull request with a change to the docs to reflect that.


On Thu, Nov 12, 2015 at 8:43 AM, <denser...@gmail.com> wrote:
*ACES 1.0 config (not OCIO 1.0)

--
Reply all
Reply to author
Forward
0 new messages