error with ociotoicc

[Gresham Lochner]

Hi everyone,

Recently I've tried using the ociotoicc utility with a lut on the
command line. When I try to load the icc file into photoshop, I receive:
"Could not load the RGB working space because the profile is not a valid
RGB working space profile."

I've tried .cube's, 3dl's. I've even tried changing some of the tags
in the icc profile, nothing has worked so far. When I'm looking at the
different tags, it looks like a normal icc profile. I've attached it to
the document.

Anyone else having this issue?

Thanks!

- Gresham

[Marie Fétiveau]

Hello !

I tried your ICC profil in Gimp and it seems to work.

I ask one of my colleague to try it on Photoshop CS2 but as the profil doesn't have a description I can't find it in the list (may be you can add a --description to your ocio2icc cmdline ?).

But I made a quick try with a 3dl :

ocio2icc -lut aces_to_ODT_sRGB_17.3dl --description "pouet" test.icc

(This 3dl was generated with ocioBakeLut and the IIF profil Jeremy provided few weeks ago).

It works too...

Sorry that's not really helpful ;p

Marie

[Malcolm Humphreys]

Hi Gresham,

Where are you these days?

The icc profiles generated from any of the ocio tools can only be used as 'soft-proofing' profiles.

I'm not sure if your interested in the technical details but in brief as ocio is a RGB only system internally (rather than XYZ or Lab). This means that 3D luts can only process in the fwd direction (it doesn't matter which cube format you use).

For a 'working space' icc profile you need 2 transforms a fwd and reverse (in Lab or XYZ) for it to be valid, for this and the above reason this is not possible to generate a profile with ocio and with the RGB cube that you have.

You should be able to load your 'soft-proofing' profile over from the window menu (i think..) in photoshop, I guess we should document this. A 'soft-proofing' profile doesn't change the 'working space' of your file just shows you what your file would look like when viewed/printed on the target device which your profile describes.

.malcolm

[Gresham Lochner]

Hey Marie and Malcolm,

Thanks for the feedback. I was able to add a descriptor argument when
creating the icc profile and it ended up working.

This might be a great opportunity to talk about what people's matte
painting to comp workflows are (photoshop to nuke in particular). I've
found that there are 3 different areas where you can utilize an icc
profile within photoshop.

First is the "working spaces" area in the Edit > Color Settings menu.
Second is imbedding the icc into the document via Edit > Assign Profile.
Third is setting up a proof in the View > Proof Setup > Custom menu.

Each one in that order, overrides the previous.

I've also noticed that using the first and second methods to apply an
icc profile, it also affects the color picker and weights the colors
according to the lut. When I set the icc via the proof setup, it
doesn't affect the color picker.

Currently what I'm doing is applying my lut through the "assign profile"
area and then creating a sRGB icc profile with nuke's sRGB math and
setting that to my proof setup. That way an artist can hit ctrl + Y to
turn on and off the lut baked icc.

Now that I've been thinking about it more, it might make sense to choose
the sRGB icc profile as the default and setting the lut baked icc to the
proof so the matte painter is painting in their....semi-neutral
colorspace and every now and then will hit ctrl+y to check their work
with the lut on.

What are people's thoughts? Do you think most matte painters would want
to work with lut on by default (like most compers do in nuke) or work in
a neutral color space and then check their work every now and then with
a lut on?

Thanks,

- G

[Kai-Uwe Behrmann]

Am 25.11.11, 10:15 -0000 schrieb Malcolm Humphreys:

> The icc profiles generated from any of the ocio tools can only be used as
> 'soft-proofing' profiles.

What profiles class are these ICC profiles? The ocio (git) command gives
me a monitor class profile:
ocio2icc -lut aces_to_ODT_sRGB_17.3dl --description "aces_to_ODT_sRGB_17.3dl" test.icc

ICC profiles of class 'monitor' can be used as editing(working) space
profiles.

> I'm not sure if your interested in the technical details but in brief as ocio
> is a RGB only system internally (rather than XYZ or Lab). This means that 3D
> luts can only process in the fwd direction (it doesn't matter which cube
> format you use).

The ocio2icc contains a A2B0 and a B2A0 cLUT. So forward and backward
tables are pressent.

> For a 'working space' icc profile you need 2 transforms a fwd and reverse (in
> Lab or XYZ) for it to be valid, for this and the above reason this is not
> possible to generate a profile with ocio and with the RGB cube that you have.

Correctly. But what does the input aces_to_ODT_sRGB_17.3dl describe?
Sorry I have no glue about aces or ODT. Are these scene refered working
spaces?

kind regards
Kai-Uwe Behrmann
--
developing for colour management
www.behrmann.name + www.oyranos.org

[Marie Fétiveau]

Hello Gresham,

For example, here we use a display LUT to simulate film on log images.

When working in Photoshop, matte painters only activate the LUT when they need to check if their work is ok.

So Proof Set up + ctrl-Y is what they need.

++

Marie 

[Marie Fétiveau]

Hello Kai-Uwe,

 

Correctly. But what does the input aces_to_ODT_sRGB_17.3dl describe?

Sorry I have no glue about aces or ODT. Are these scene refered working spaces?

aces_to_ODT_sRGB_17.3dl is a LUT I used to make a quick test about Gresham's pb.

It's just an example here (I took the first lut I had at hand :) )

Anyway, to answer your question, it's a display LUT I generated thanks to ociobakelut and the IIF ocio config file.

IIF stands for Image Interchange Framework, it's an architecture for a better color management designed for motion pictures. 

ACES (Academy Colorspace Encoding Specifications) and ODT (Output Device Transform) are part of this framework.

If you're not planning to use IIF workflow (it's quite experimental right now), this LUT isn't very usefull :)

That said, if IIF interests you : here's a soft introduction and a more precise academy document.

++

Marie

 

[Kai-Uwe Behrmann]

Am 28.11.11, 12:32 +0100 schrieb Marie F�tiveau:

>> Correctly. But what does the input aces_to_ODT_sRGB_17.3dl describe?
>> Sorry I have no glue about aces or ODT. Are these scene refered working
>> spaces?
>
> aces_to_ODT_sRGB_17.3dl is a LUT I used to make a quick test about
> Gresham's pb.
> It's just an example here (I took the first lut I had at hand :) )
>
> Anyway, to answer your question, it's a display LUT I generated thanks to
> ociobakelut and the IIF ocio config file.
> IIF stands for Image Interchange

> Framework<http://www.oscars.org/science-technology/council/projects/iif.html>,

> it's an architecture for a better color management designed for motion
> pictures.
> ACES (Academy Colorspace Encoding Specifications) and ODT (Output Device
> Transform) are part of this framework.
> If you're not planning to use IIF workflow (it's quite experimental right
> now), this LUT isn't very usefull :)

Then aces_to_ODT_sRGB_17.3dl defines a colour transformation and not a
colour space.
Does the 3dl container transport CIE*XYZ informations? Otherwise I do not
yet understand how the reference to the PCS (CIE*XYZ) inside the profile
come in existence.

But probably I miss something. You motion picture people use a very
different terminology :-)

> That said, if IIF interests you : here's a soft

> introduction<http://mikemost.com/?p=235>and a more precise academy
> document<http://www.digitalpreservation.gov/partners/ampas/high/IIF_Overview_August_2010.pdf>

thanks
Do you have a link to the ACES and RRT�s specifications. I found only soft
descriptions via my internet search.

kind regards
Kai-Uwe

[Malcolm Humphreys]

On 28 Nov, 2011,at 09:42 PM, Kai-Uwe Behrmann <ku...@gmx.de> wrote:

The ocio2icc contains a A2B0 and a B2A0 cLUT. So forward and backward
tables are pressent.

 

I know but the BToA0Tag (PCS to Device space) is empty (this is so the profile will load).

So they will work only in one direction so you might be able to assign them (associate the current pixel values with a profile), but converting to and from will give you some pretty wrong results.

I would only recommend using these profiles to soft-proof as they can easily have quantization and space compression errors due to the way they are created.

.malcolm

[Marie Fétiveau]

Do you have a link to the ACES and RRT´s specifications. I found only soft descriptions via my internet search.

I'm not sure full specifications are publics yet... :/

You may have to register here ( http://iif.stcatp.org/cgi-bin/apply.cgi) to access to documentations.

 

And yes, aces_to_ODT_sRGB_17 defines a color transformation from ACES colorspace to sRGB colorspace (and it embeds the RRT + ODT sRGB transforms). 

++

Marie