Attaching some slightly larger examples :)


On 28/01/2011, at 5:46 PM, Malcolm Humphreys wrote:

Added ociobuildicc app which will build a soft-proofing icc profile for a given working space.
(added LCMS2 into ext)

Commits:
https://github.com/malcolmhumphreys/OpenColorIO/commit/1b0d39fb10db10f4edbf1fb0fed124b548b836d9

Attached is an example of two profiles in action.

From the left
* sRGB reference image (ocio cpu)
* matte paint allocation space
* log space image.

The visual difference in the log image is most likely caused by 16bit quantization errors in the
CLUT of the AToB0Tag. These errors are less with the matte paint allocation as the distance
to travel is less.

I did play around with supporting the new MPE D2Bx and B2Dx tags which should make this
problem go away (http://www.color.org/ICCSpecRevision_02_11_06_Float.pdf) but couldn't
get it too work reliably so I pulled it out till I can.

.malcolm



<ocioicc_softp_off.jpeg><ocioicc_softp_on.jpeg>

<ocioicc_softp_off.jpeg><ocioicc_softp_on.jpeg>

Added ociobuildicc app which will build a soft-proofing icc profile for a given working space.
(added LCMS2 into ext)

Commits:
https://github.com/malcolmhumphreys/OpenColorIO/commit/1b0d39fb10db10f4edbf1fb0fed124b548b836d9

Attached is an example of two profiles in action.

From the left
* sRGB reference image (ocio cpu)
* matte paint allocation space
* log space image.

The visual difference in the log image is most likely caused by 16bit quantization errors in the
CLUT of the AToB0Tag. These errors are less with the matte paint allocation as the distance
to travel is less.

I did play around with supporting the new MPE D2Bx and B2Dx tags which should make this
problem go away (http://www.color.org/ICCSpecRevision_02_11_06_Float.pdf) but couldn't
get it too work reliably so I pulled it out till I can.

.malcolm



<ocioicc_softp_off.jpeg><ocioicc_softp_on.jpeg>

Oh,
https://github.com/dbr/OpenColorIO/commit/0247f16ae202eeb5316f695c1310dac523272a0e
On 23/01/2011, at 4:41 AM, Jeremy Selan wrote:

Nice job on the docs.  By all means, check them in!  (Even partial
docs are better than nothing). :)

I wonder, could the cccid's be presented in a drop-down, similar to the

colour-space listings?

I like this, but I'm not sure that there's a nuke UI widget rich
enough for this behavior.   What we want is a popdown that lets you
choose from a bunch of pre-defined options, but also lets you type in
new values (such as envvars).  Im not sure that nuke popdowns can
allow text entry capability.  The other difficulty is that in nuke if
you use the popdowns, the underlying value is an int index rather than
a string.  This sucks because if you were to update the array, change
the underlying value, etc, the only precious data for nuke is what
order it was in the list. (Whereas in our case the value is precious).
So I think it's safer to keep it a string for now.   But... what if
we added a button that when pressed provided a popup for the user?
(Calling out to OCIO to enumerate the options).  That would make it
convenient + keep data value 'native'.

I like the use of the invisible flag. I'll investigate setting it up.

-- Jeremy

Thanks!

I like the intent, but have two minor issues:

First, the trivial: I'd prefer to never allow NULL values as return
types from const char * functions. Please use an empty string ""
instead.

The reason being, OCIO expects the client to copy the returned string
value, if the client intends to hang onto it. So it's often convenient
in your code to do a directly assign the function result to a
std::string,

std::string rolename = config->getRoleNameByIndex(i);

If the API promises to never return a NULL value, this will work
great. However, if OCIO can return a null pointer than the above code
will crash. Thus for all char * return functions the user would be
required to explicitly add a null check. This is probably counter to
user expectations, and I expect that a few would slip through and
result in occasionally crashy client code. (This would be likely
because most of the time it *would* work fine. Only on invalid
configurations, or other corner cases, would this sharp corner be run
into. I'd like to keep OCIO as friendly as reasonable.)

(See http://github.com/imageworks/OpenColorIO/issues#issue/20 for more
thoughts on this topic).

Sure that sounds great, i'll fix that.

The second issue is that you've exposed two helper functions
(getRoleColorSpaceName), that are not quite as helpful as they appear.
Is the client supposed to use these at runtime to determine which
colorspace to use for a given role?

Allow me to answer... "No!"

:)

ConstColorSpaceRcPtr getColorSpace(const char * name) const;
already does this in the 'correct' way, and this should be the *only*
allowed mechanism for getting a ColorSpace from a name. It
encapsulates a bit of extra mojo your new function does not, such as
looking at the 'strictParsing', and if allowed returning the DEFAULT
colorspace role in fall-through situations. It also resolves the
ambiguity where a role name and colorspace name overlap, and having
this resolution only defined in a single place seems preferable.

What if we renamed your functions to make it super clear that they are
not intended for color-space resolution, but instead as configuration
setup/ui purposes?

I.e.
setRole(const char * role, const char * colorSpaceName);
const char * getRoleName(int index);
const char * getRoleValue(int index);

I don't remind removing them altogether.

I was really just adding them to match the old functionality of getRoleNameByIndex() which returned the colorspace name not the role name, I wasn't sure if you were needing the colorspace name.

Calling getColorSpace(getRoleName(0)) seems nice an clean to me.

.malcolm

Even though the role 'value' is a colorspace name, I think the
function would be far less tempting for the naive API user to
accidentally call.

If we'd like a convenience function to do the resolution, we could add
const char * getColorSpaceName(const char * name) const;

which would internally rely on config::getColorSpace.

-- Jeremy

On Sun, Jan 16, 2011 at 8:22 PM, Malcolm Humphreys
<malcolmh...@...> wrote:

When building some UI menus, I wanted to get a list of defined roles. Config::getRoleNameByIndex() only returned the resolved colorspace name.

--snip--
Replaced config->getRoleNameByIndex (which returned the colorspace name, not the role name) with the following:

- getRoleName(int): get the role name at index, this will return values like 'scene_linear', 'compositing_log'. Returns NULL if index is out of range

- getRoleColorSpaceName(int): get the actual colorspace name at role index. Returns NULL when index is out of range

- getRoleColorSpaceName(char*): get the actual colorspace name from a role name. Returns NULL if roles is not defined
--snip--

https://github.com/malcolmhumphreys/OpenColorIO/commit/d7c8094f6aad57f980060e87d0f547ff36d844b7

.malcolm

Commit:
http://github.com/jeremyselan/OpenColorIO/commit/d37b316cd8354216b4d89b85c06ab08108b5d1cc

A pretty trivial change, include python.h first when building pyglue.

Some linux installations are adamant about including python.h first,
or you get warnings about duplicate #DEFINES for _POSIX_C_SOURCE,
_XOPEN_SOURCE, etc. This cleans it up.

-- Jeremy

Yes, I like it. Just haven't had a chance to update the code and
re-submit yet. Sorry...

-- Jeremy

On Wed, Jan 26, 2011 at 4:08 PM, Malcolm Humphreys
<malcolmh...@...> wrote:

Did you have any thoughts on this?

On 26/01/2011, at 12:12 PM, Malcolm Humphreys wrote:

I'm thinking of an edge case where you want to query the defalt displays of a color configuration independently of the active displays. The example I could think of is writing a script to check to see if a display is supported by default and creating one if it is not.

Could this not be part of the API? the core could just keep around the list that was in the profile as the default list and then the active would either be a copy of this or what was ever dynamically set.

Looking a little more closely I don't understand the need to have multiple defaults in the active_views
active_views: [Film, Log, Raw]
Unless you want to explictly exclude a defined view as a default.

This might answer jeremy's questions as well.

I was thinking the order is important in active_displays and active_views.

I guess the use case for all this is:

Say we have 3 groups of machines.
group1: dual sRGB heads
group2: sRGB and one Dreamcolor (DCI-P3 emulation) heads
group3: sRGB and a DCI projector head

In the profile it would be nice to model that you want both sRGB and Dream as the visible displays (say in the nuke node) if you load the profile directly with sRGB being the default. eg. active_displays: [sRGB, Dream]

For group2 you would might want the Dream display to be the default so you could set the environment var $OCIO_ACTIVE_DISPLAYS = Dream,sRGB

And the same for group3, $OCIO_ACTIVE_DISPLAYS = DCI,sRGB (and the Dream display would not be visible)

The active_views would be identify the default as the first item and then the order the api would return them in. This would work similarly in being able to override the order for different machines.


....
I gota run guys to Australian Day celebrations.

Happy Australia Day everyone! (get a beer into you)

.malcolm

-Joseph


-----Original Message-----
From: ocio...@... [mailto:ocio...@...] On Behalf Of Malcolm Humphreys
Sent: Tuesday, January 25, 2011 3:49 PM
To: ocio...@...
Subject: Re: [ocio-dev] Review: DisplayTransform interface update

Hi Joesph,

I like your layout. I think that calling the

active_displays: [sRGB, DCIP3] # or active_devices:

default_displays:[sRGB, DCIP3]

The name change will allow the opportunity to have a list of default display that can be different from what the active displays are.
I think that this can be important given the ability to dynamically generate colorspaces and display devices.

Oh I was thinking in the profile the list of active_displays would be the defaults. Which then you could dynamically override with $OCIO_ACTIVE_DISPLAYS (which was called $OCIO_DEVICE_MASK).

.malcolm

-Joseph

-----Original Message-----
From: ocio...@... [mailto:ocio...@...] On
Behalf Of Malcolm Humphreys
Sent: Tuesday, January 25, 2011 3:23 PM
To: OpenColorIO Developers
Subject: Re: [ocio-dev] Review: DisplayTransform interface update

Hi,

Here's a chunk from spi-vfx with the new options:

displays:
- !<Display> {device: sRGB, alias: Raw, colorspace: nc10}
- !<Display> {device: sRGB, alias: Log, colorspace: lg10}
- !<Display> {device: sRGB, alias: Film, colorspace: srgb8}
- !<Display> {device: DCIP3, alias: Raw, colorspace: nc10}
- !<Display> {device: DCIP3, alias: Log, colorspace: lg10}
- !<Display> {device: DCIP3, alias: Film, colorspace: p3dci8}
device_defaults: [sRGB, DCIP3]
alias_defaults: [Film, Log, Raw]

I feel that the term 'alias' is simpler to understand, but then again
I wrote it so if you have a better term I'm open to ideas.

Ok thats a bit clearer. But I think having this as a flat list is not so nice.

After playing around with this a bit I came up with this layout. This makes it clearer that you have a set of displays and a set of views on each of them.

displays:
sRGB:
-!<View> {name: Raw, colorspace: nc10}
-!<View> {name: Log, colorspace: lg10}
-!<View> {name: Film, colorspace: srgb8}
DCIP3:
-!<View> {name: Raw, colorspace: nc10}
-!<View> {name: Log, colorspace: lg10}
-!<View> {name: Film, colorspace: p3dci8}
active_displays: [sRGB, DCIP3] # or active_devices:
active_views: [Film, Log, Raw]

You'll see that the default value is also explicitly specified now.
Under common circumstances, only the 1st value will be returned as
default, but a full search order is allowed for corner cases where
not all aliases are defined for all devices, or there are
OCIO_DEVICE_MASK specified that disallow particular device options.

The idea behind OCIO_DEVICE_MASK is that your config will contain the
master list of all devices in use on a show, but still allow the
flexibility (through the use of envvars) to control which an end user
sees. This allows for example, a facility to roll out code to make
sure the user only has the "sRGB" option on desktop machines. (Which
DLP connected machines may have all the options: sRGB, DCIP3, etc).

I think this is great!

.malcolm

-- Jeremy

On Sat, Jan 22, 2011 at 9:40 PM, Malcolm Humphreys
<malcolmh...@...> wrote:

Hi,

Can you post an example profile with a few more display lines. The code for this looks fine, I'm just not too sure of the structure in the yaml and how clear the names device and alias are as terminology in this context.

eg, We have a mixture displays of different make and models and each of these I would call a device (we do group these into manageable sets and calibrate them to an 'ideal' device). I keep wanting to call these a 'target' or 'target-device'.

But I feel like I'm maybe missing something so an example would be good.

.malcolm

On 21/01/2011, at 10:01 AM, Jeremy Selan wrote:

Commit:
http://github.com/jeremyselan/OpenColorIO/commit/2ac16ae5e55949e297
e
d0d14baa585f04ff4d452

Issues:
https://github.com/imageworks/OpenColorIO/issues#issue/25
https://github.com/imageworks/OpenColorIO/issues#issue/21
https://github.com/imageworks/OpenColorIO/issues#issue/22

The API to get/set display devices has been updated. What previous
was 'TransformName' is now called a 'DisplayAlias', to minimize
confusion with new users. (The old Transforms has nothing to do
with OCIO::Transform, so this name was misleading). A new
environment variable is available, $OCIO_DEVICE_MASK, which allows
for optional runtime speicification of connected display devices.
The config format has been updated to reflect these modified names.
Also available are two new entries, device_defaults +
alias_defaults which allow for a specification of defaults for devices and alises.

This breaks compatibility with both the API and the .config format.

FYI, Pretty soon we hope to have a stable branch locked off!
We're going to be switching Imageworks this month to exclusively
use the public OCIO for all upcoming productions.

-- Jeremy

I'm thinking of an edge case where you want to query the defalt displays of a color configuration independently of the active displays. The example I could think of is writing a script to check to see if a display is supported by default and creating one if it is not.

Could this not be part of the API? the core could just keep around the list that was in the profile as the default list and then the active would either be a copy of this or what was ever dynamically set.

Looking a little more closely I don't understand the need to have multiple defaults in the active_views
active_views: [Film, Log, Raw]
Unless you want to explictly exclude a defined view as a default.

This might answer jeremy's questions as well.

I was thinking the order is important in active_displays and active_views.

I guess the use case for all this is:

Say we have 3 groups of machines.
group1: dual sRGB heads
group2: sRGB and one Dreamcolor (DCI-P3 emulation) heads
group3: sRGB and a DCI projector head

In the profile it would be nice to model that you want both sRGB and Dream as the visible displays (say in the nuke node) if you load the profile directly with sRGB being the default. eg. active_displays: [sRGB, Dream]

For group2 you would might want the Dream display to be the default so you could set the environment var $OCIO_ACTIVE_DISPLAYS = Dream,sRGB

And the same for group3, $OCIO_ACTIVE_DISPLAYS = DCI,sRGB (and the Dream display would not be visible)

The active_views would be identify the default as the first item and then the order the api would return them in. This would work similarly in being able to override the order for different machines.


....
I gota run guys to Australian Day celebrations.

Happy Australia Day everyone! (get a beer into you)

.malcolm

-Joseph


-----Original Message-----
From: ocio...@... [mailto:ocio...@...] On Behalf Of Malcolm Humphreys
Sent: Tuesday, January 25, 2011 3:49 PM
To: ocio...@...
Subject: Re: [ocio-dev] Review: DisplayTransform interface update

Hi Joesph,

I like your layout. I think that calling the

active_displays: [sRGB, DCIP3] # or active_devices:

default_displays:[sRGB, DCIP3]

The name change will allow the opportunity to have a list of default display that can be different from what the active displays are.
I think that this can be important given the ability to dynamically generate colorspaces and display devices.

Oh I was thinking in the profile the list of active_displays would be the defaults. Which then you could dynamically override with $OCIO_ACTIVE_DISPLAYS (which was called $OCIO_DEVICE_MASK).

.malcolm

-Joseph

-----Original Message-----
From: ocio...@... [mailto:ocio...@...] On
Behalf Of Malcolm Humphreys
Sent: Tuesday, January 25, 2011 3:23 PM
To: OpenColorIO Developers
Subject: Re: [ocio-dev] Review: DisplayTransform interface update

Hi,

Here's a chunk from spi-vfx with the new options:

displays:
- !<Display> {device: sRGB, alias: Raw, colorspace: nc10}
- !<Display> {device: sRGB, alias: Log, colorspace: lg10}
- !<Display> {device: sRGB, alias: Film, colorspace: srgb8}
- !<Display> {device: DCIP3, alias: Raw, colorspace: nc10}
- !<Display> {device: DCIP3, alias: Log, colorspace: lg10}
- !<Display> {device: DCIP3, alias: Film, colorspace: p3dci8}
device_defaults: [sRGB, DCIP3]
alias_defaults: [Film, Log, Raw]

I feel that the term 'alias' is simpler to understand, but then again
I wrote it so if you have a better term I'm open to ideas.

Ok thats a bit clearer. But I think having this as a flat list is not so nice.

After playing around with this a bit I came up with this layout. This makes it clearer that you have a set of displays and a set of views on each of them.

displays:
sRGB:
-!<View> {name: Raw, colorspace: nc10}
-!<View> {name: Log, colorspace: lg10}
-!<View> {name: Film, colorspace: srgb8}
DCIP3:
-!<View> {name: Raw, colorspace: nc10}
-!<View> {name: Log, colorspace: lg10}
-!<View> {name: Film, colorspace: p3dci8}
active_displays: [sRGB, DCIP3] # or active_devices:
active_views: [Film, Log, Raw]

You'll see that the default value is also explicitly specified now.
Under common circumstances, only the 1st value will be returned as
default, but a full search order is allowed for corner cases where
not all aliases are defined for all devices, or there are
OCIO_DEVICE_MASK specified that disallow particular device options.

The idea behind OCIO_DEVICE_MASK is that your config will contain the
master list of all devices in use on a show, but still allow the
flexibility (through the use of envvars) to control which an end user
sees. This allows for example, a facility to roll out code to make
sure the user only has the "sRGB" option on desktop machines. (Which
DLP connected machines may have all the options: sRGB, DCIP3, etc).

I think this is great!

.malcolm

-- Jeremy

On Sat, Jan 22, 2011 at 9:40 PM, Malcolm Humphreys
<malcolmh...@...> wrote:

Hi,

Can you post an example profile with a few more display lines. The code for this looks fine, I'm just not too sure of the structure in the yaml and how clear the names device and alias are as terminology in this context.

eg, We have a mixture displays of different make and models and each of these I would call a device (we do group these into manageable sets and calibrate them to an 'ideal' device). I keep wanting to call these a 'target' or 'target-device'.

But I feel like I'm maybe missing something so an example would be good.

.malcolm

On 21/01/2011, at 10:01 AM, Jeremy Selan wrote:

Commit:
http://github.com/jeremyselan/OpenColorIO/commit/2ac16ae5e55949e297
e
d0d14baa585f04ff4d452

Issues:
https://github.com/imageworks/OpenColorIO/issues#issue/25
https://github.com/imageworks/OpenColorIO/issues#issue/21
https://github.com/imageworks/OpenColorIO/issues#issue/22

The API to get/set display devices has been updated. What previous
was 'TransformName' is now called a 'DisplayAlias', to minimize
confusion with new users. (The old Transforms has nothing to do
with OCIO::Transform, so this name was misleading). A new
environment variable is available, $OCIO_DEVICE_MASK, which allows
for optional runtime speicification of connected display devices.
The config format has been updated to reflect these modified names.
Also available are two new entries, device_defaults +
alias_defaults which allow for a specification of defaults for devices and alises.

This breaks compatibility with both the API and the .config format.

FYI, Pretty soon we hope to have a stable branch locked off!
We're going to be switching Imageworks this month to exclusively
use the public OCIO for all upcoming productions.

-- Jeremy

Would this syntax preserve Display order?  I think that's important.

-- Jeremy

displays:
  sRGB:
     -!<View> {name: Raw, colorspace: nc10}
     -!<View> {name: Log, colorspace: lg10}
     -!<View> {name: Film, colorspace: srgb8}
  DCIP3:
     -!<View> {name: Raw, colorspace: nc10}
     -!<View> {name: Log, colorspace: lg10}
     -!<View> {name: Film, colorspace: p3dci8}

Would this syntax preserve Display order?  I think that's important.

-- Jeremy

displays:
  sRGB:
     -!<View> {name: Raw, colorspace: nc10}
     -!<View> {name: Log, colorspace: lg10}
     -!<View> {name: Film, colorspace: srgb8}
  DCIP3:
     -!<View> {name: Raw, colorspace: nc10}
     -!<View> {name: Log, colorspace: lg10}
     -!<View> {name: Film, colorspace: p3dci8}

displays:
  sRGB:
     -!<View> {name: Raw, colorspace: nc10}
     -!<View> {name: Log, colorspace: lg10}
     -!<View> {name: Film, colorspace: srgb8}
  DCIP3:
     -!<View> {name: Raw, colorspace: nc10}
     -!<View> {name: Log, colorspace: lg10}
     -!<View> {name: Film, colorspace: p3dci8}

I like your layout. I think that calling the

active_displays: [sRGB, DCIP3] # or active_devices:

default_displays:[sRGB, DCIP3]

The name change will allow the opportunity to have a list of default display that can be different from what the active displays are.
I think that this can be important given the ability to dynamically generate colorspaces and display devices.

-Joseph

-----Original Message-----
From: ocio...@... [mailto:ocio...@...] On
Behalf Of Malcolm Humphreys
Sent: Tuesday, January 25, 2011 3:23 PM
To: OpenColorIO Developers
Subject: Re: [ocio-dev] Review: DisplayTransform interface update

Hi,

Here's a chunk from spi-vfx with the new options:

displays:
- !<Display> {device: sRGB, alias: Raw, colorspace: nc10}
- !<Display> {device: sRGB, alias: Log, colorspace: lg10}
- !<Display> {device: sRGB, alias: Film, colorspace: srgb8}
- !<Display> {device: DCIP3, alias: Raw, colorspace: nc10}
- !<Display> {device: DCIP3, alias: Log, colorspace: lg10}
- !<Display> {device: DCIP3, alias: Film, colorspace: p3dci8}
device_defaults: [sRGB, DCIP3]
alias_defaults: [Film, Log, Raw]

I feel that the term 'alias' is simpler to understand, but then again
I wrote it so if you have a better term I'm open to ideas.

Ok thats a bit clearer. But I think having this as a flat list is not so nice.

After playing around with this a bit I came up with this layout. This makes it clearer that you have a set of displays and a set of views on each of them.

displays:
sRGB:
-!<View> {name: Raw, colorspace: nc10}
-!<View> {name: Log, colorspace: lg10}
-!<View> {name: Film, colorspace: srgb8}
DCIP3:
-!<View> {name: Raw, colorspace: nc10}
-!<View> {name: Log, colorspace: lg10}
-!<View> {name: Film, colorspace: p3dci8}
active_displays: [sRGB, DCIP3] # or active_devices:
active_views: [Film, Log, Raw]

You'll see that the default value is also explicitly specified now.
Under common circumstances, only the 1st value will be returned as
default, but a full search order is allowed for corner cases where
not all aliases are defined for all devices, or there are
OCIO_DEVICE_MASK specified that disallow particular device options.

The idea behind OCIO_DEVICE_MASK is that your config will contain the
master list of all devices in use on a show, but still allow the
flexibility (through the use of envvars) to control which an end user
sees. This allows for example, a facility to roll out code to make
sure the user only has the "sRGB" option on desktop machines. (Which
DLP connected machines may have all the options: sRGB, DCIP3, etc).

I think this is great!

.malcolm

-- Jeremy

On Sat, Jan 22, 2011 at 9:40 PM, Malcolm Humphreys
<malcolmh...@...> wrote:

Hi,

Can you post an example profile with a few more display lines. The code for this looks fine, I'm just not too sure of the structure in the yaml and how clear the names device and alias are as terminology in this context.

eg, We have a mixture displays of different make and models and each of these I would call a device (we do group these into manageable sets and calibrate them to an 'ideal' device). I keep wanting to call these a 'target' or 'target-device'.

But I feel like I'm maybe missing something so an example would be good.

.malcolm

On 21/01/2011, at 10:01 AM, Jeremy Selan wrote:

Commit:
http://github.com/jeremyselan/OpenColorIO/commit/2ac16ae5e55949e297
e
d0d14baa585f04ff4d452

Issues:
https://github.com/imageworks/OpenColorIO/issues#issue/25
https://github.com/imageworks/OpenColorIO/issues#issue/21
https://github.com/imageworks/OpenColorIO/issues#issue/22

The API to get/set display devices has been updated. What previous
was 'TransformName' is now called a 'DisplayAlias', to minimize
confusion with new users. (The old Transforms has nothing to do
with OCIO::Transform, so this name was misleading). A new
environment variable is available, $OCIO_DEVICE_MASK, which allows
for optional runtime speicification of connected display devices.
The config format has been updated to reflect these modified names.
Also available are two new entries, device_defaults +
alias_defaults which allow for a specification of defaults for devices and alises.

This breaks compatibility with both the API and the .config format.

FYI, Pretty soon we hope to have a stable branch locked off!
We're going to be switching Imageworks this month to exclusively
use the public OCIO for all upcoming productions.

-- Jeremy