EXR features.

[Dubbie99]

I am missing a few EXR features in Solaris that I would like to have.

1: DWAB compression.
2: Data window auto-crop.
3: Camera in the metadata.

[bdancer]

V-Ray does not do any image writing - all is done by Hydra. You should request those from SideFX.

[Dubbie99]

Ah understood. Will do!

[stehrani3d]

Does the Render Product LOP provide the functionality you need? I see a few options relating to DWA-B Compression, Data Window and Metadata

[Dubbie99]

Ah nice one stehrani3d, that takes care of the compression and the metadata. Because only the renderer knows the pixel bounds for computing the data window, I don’t know if there is a way for vray to communicate that to hydra to set that automatically? I’ll ask side effects.

[LeoYfver]

All of these are handled by hydra/husk

1. Dwaa and dwab is set on the renderproduct. Though it does only filter the beauty and no other aov for me. I have asked sidefx and no response so if anybody has managed please enlighten me . I want to do it on all except data passes.

2. This is controlled by a husk command. --autocrop rgba for example. This was added in 19.5. https://www.sidefx.com/docs/houdini/ref/utils/husk.html

3. Is set on renderproduct too. If you check my gbk_aovs there should be example code for how to do this.

[Dubbie99]

Thanks Leo, I have learned a lot about this Seems like I can do everything I need.

I did read that it is possible as of 19.5 to set the compression per AOV, but with a quick look I couldn't see how.

[LeoYfver]

A week ago i sent them a bug report that seems to be fixed now. I havent tried it out yet though

Basically the rules are based on openEXR recommendation. So when the channel names are uppercase it gets compressed and lowercase they dont.

So after this bugfix when you export a color rendervar it will be layer.RGBA (compressed) and if you export rendervar as point or vector it will be layer.xyz (not compressed).



Regards Alexis

[Dubbie99]

Good to know. Thanks Alexis!

[Dubbie99]

Hey Alexis,

what code did you use exactly to write the camera xForm into the metadata. Not seeing anything in your HDA besides a field to add metadata… maybe that’s what you meant?

[LeoYfver]

There should be a drodown next to the parameter field. There you can select cameras. It will create a python if script that takes all the cameras and list them. Are we talking about the gbk_aovs?

[Dubbie99]

I see that, but as far as I can tell, it just gets the name of the camera, not its transform.
Yes, talking about gbk_aovs.
Am I missing something?

[LeoYfver]

ah, yes. The position was only done for the lights i think. We export cameras as alembic so the compositors have access to the position but not for the lights so that probably why.