OpenEXR support

[esi]

Hi Vlado,

I would imagine that you are very busy. However as our studio is moving onto some Feature film work I would very much like to use the OpenEXR format, however my TD (Dmitry who you know), says that it will not render this format out of VRAY. Is this true? If so is it difficult to implement this into VRAY?

I only ask as it would be a very nice addition to our toolset.

Thanks very much

Mark Rasmussen
VFX Supervisor / Founder
Enigma Studios Inc

[Dynedain]

It's because Max7 doesn't have full floating point operations. Supposedly Max8 supports openexr "natively" which might mean that it has full floating point internal operations. The other option is to render to vrayimg format and ask Vlado for the app they have that will convert vrayimg to openexr.

[cpnichols]

I render to EXR with Vray. In Max7 you need the Splutterfish plugin (free), and in Max 8 it is built in. I will support EXR in float unclamped as well.

This tutorial was done with exrs:

http://www.chaosgroup.com/forum/phpB...ic.php?t=13818

[Morbid Angel]

but vlado mentioned that vray does not support direct rendering into open exr?

[czoog]

but vlado mentioned that vray does not support direct rendering into open exr?

Not sure about that, but I though throb had mentioned that vlado had written a vray image format to .exr converter

[panthon]

I´ve been rendering to openexr with max 8 with no problems. The only drawback is that combustion won´t read the extra channels, such as z-depth or velocity, which rpf will. I don´t know if it´s due to bad implementation of exr in max or in combustion...
What´s the benefit of using openexr with 32 bit per channel instead of rpf 32 bit?

[Morbid Angel]

ok lets straghten things out:
max 7 and vray do not render to open exr directly, thats why there is a converter, so one must first render to vrayimg then convert the files into exr. Max 8 is a different story all together.

[trixian]

Actually, I have been having problems with this as of lately, in max8. Seems even with unchecking "clamp output" and enabling real rgb in the gbuffer part, my exr images saved as 32-bit float do not contain the hdr information in combustion. I have had a hard time figuring this out. I was then playing with the curves in the Vray framebuffer getting a nice exposure I wanted to recreate in combustion, but could not for the life of me recreate it. Can anyone give me any hints?

[Morbid Angel]

As far as I know (little experience with max 8 and exr/vray) there has been no update for vray to support this feature in max 8, so one can assume this would not work. Rather use vrimg to exr converter.

[cpnichols]

What´s the benefit of using openexr with 32 bit per channel instead of rpf 32 bit?

One huge advantage is diskspace. EXR lossless compression is awesome, which can save a huge amount of money if you are doing a feature film.

[Morbid Angel]

chris what about this:
Would this be a benefit if an exr file would contain not only gbuffer info and float, but also all of the pases i.e. specular, reflection, shadow, etc etc?
would this be a benefit to store all of these pases in one file as opposed to separate ones....as it would make compositors life much easier not having to load the pases and having to sort out 20+ different pases.

[cpnichols]

chris what about this:
Would this be a benefit if an exr file would contain not only gbuffer info and float, but also all of the pases i.e. specular, reflection, shadow, etc etc?
would this be a benefit to store all of these pases in one file as opposed to separate ones....as it would make compositors life much easier not having to load the pases and having to sort out 20+ different pases.

Yeap.. everything in float would be great.... not sure that it isn't right now. pretty sure that all the data I wrote out was in float.
The one file thing would be great as well.
I would also be curious about making the vrimg to exr converter more public as well.

[trixian]

Definately need that vrimg to exr converter. Do I just email Count Vladola and ask for it? Or must one be in a secret coven of devine raytracers to be privy to it?

[studioDIM]

RPF already stores everything (Z, velocities, and so on) as 32bpc.
That's one of the reasons why it tends to be bigger.
In fact, it has a lossless compression of its own.

Edit: i stand corrected by myself
EXRs with RLE are more than half the size of RPFs, at 32bpc
At lower range, though, RPF is smaller than an RLE Tga.
It seems compaction is switched off at higher depths...
Anyone with a Max SDK at hand to confirm this?

The other reason is the number of extended channels you may want to embed.
While EXR has to have them custom coded, RPF ones are just a tick of a box away.

Combustion does clamp to 1.000 the brightness of 32bpc images when an alpha premultiplication is done in the footage properties page, rendering pretty useless any CC you might attempt afterwards.
Try using the image as Straight and matte the image through the comp, rather than in the footage properties.

hope it helps...

[vlado]

As far as I know (little experience with max 8 and exr/vray) there has been no update for vray to support this feature in max 8, so one can assume this would not work.

Newer V-Ray builds produce proper floating point information in 3dsmax 8, so this issue should be taken care of.

Best regards,
Vlado