Photoshop render element importer v0.02

[Shimakaze]

remember you have to disable Photoshop's native OpenExr plugin as well.
or else it will keep using it. just rename it to OpenEXR.off or something.
to check if the layers were saved, simply compare the file sizes with and without the layers. a multilayered EXR file is usually huge.

[trixian]

Just discovered the same confusing stuff.
http://www.chaosgroup.com/forums/vbu...t=40995&page=2

[womble]

Hi Shim, first off i'd like to just say thanks for the effort put into your scripts and advice on the render elements post I've found it all very helpful. I just had a few questions that I hoped you would maybe be able to help me with?

1) How do you work the use of .exr files into your general workflow? With the limited 32bit abilities of photoshop CS3 what would you say is the biggest advantage to working in full float linear mode? For example, do you find that being able to adjust the exposure of the various element layers to be of great help? Or could you perhaps give me an example of the changes you might make to a hypothetical image that would make it worthwhile to switch over to 32bit.

2) Do you find it necessary at any point in your workflow to 'downgrade' the image into 8 or 16 bit mode in order to carry out further photoshop work (such as the various adjustments and photoshop work that is not supported when working in a 32 bit environment. When you do reach 'the point of no return' where you lose all of your float data do you have a specific workflow to ensure this information is saved? For example do you save a 32bit float version of your image before 'downgrading' your file.

3) Finally, do you find that there is a huge improvement in your post-production ability when switching to 32bit exr, or is it more of the ability to subtly change the various elements of your render rather than a dramatic ability to alter the image without re-rendering?

Thanks for all of your help, and again, muchos props for your scripts

[Shimakaze]

Well there are really more then one reason to use 32 bit exr files, and one of the big ones has nothing to do with post at all. When using the raw image rendering functionality in VRay you get a whole bunch of advantages for free. If you disable the memory frame buffer and only render to vrimg/exr then it pretty much removes any limitation to picture size and complexity. It also gives you a perfect save of the data. (the compression used when saving to raw exr at rendertime is a lot more effective then the compression used when you save it manually from the frame buffer. Try it.) EXR files also has very good multi layer support, and is the easiest way to get them into Photoshop unless you want to buy the cebas PSD plugin.
Now about your questions:
You already mentioned one of the reasons to use 32bit. Being able to expose pictures after render. However I rarely change exposure in any ways that couldn't be done in a 16bit picture, since I prefer to expose using the VRay Physical Camera. To me the biggest workflow advantages is that (when using the ProEXR photoshop plugin) you can save all the render elements into one file and have Photoshop be able to read them properly, 32bit gives you a lot more information like the brightness values of lights (helpful if you want to add glow for instance) and so on. That being said I usually collapse the picture down to 16bit relatively soon in the workflow. Although I usually keep a duplicate just in case. All in all there really isn't any reason to not save your renders in 32bit other then storage space. Harddrives are cheap these days anyways. And when saving 32bit EXR files using the raw image function the better compression actually makes them smaller then if you were to save a 16bit image from the frame buffer in many cases!
By the way if you have After Effects you suddenly have a lot more 32bit freedom.

[pg1]

Hey Shimakaze,
We have recently started to output all our animations to 32bit exr and have much better control in post but one of the problem is file space as the exr. files are much larger. So I am trying to figure out whats the best way to output my files and elements with the least file space and keep the best quality.

I am just starting to use the vrimg file options for the first time and am trying to figure a few things out. I did a few test last night with a 640x480 image.
When I render to vrimg file the file size is quiet large 35mbs (this is with 8 elements output).

Double clicking on it (thanks to your easy to use .reg file )gets me the elements output to exr. at a small file size of between 300kbs-500kbs per element

Is the rather large difference between the 35mb .vrimg file and much smaller exr. elements just due to .exr compression? As the vrimg file is so large (this was a 640x480 render) do you just delete them after you have extracted them to exr. elements? and how do you deal with this problem with an animation?

in your other post you said

And when saving 32bit EXR files using the raw image function the better compression actually makes them smaller then if you were to save a 16bit image from the frame buffer in many cases!

When I do this I get a small exr of about 1.5 mb (much better file size but how unless there is compression?) and am able to access all the layers of the exr using pro Exr in afterfxs which is perfect but now I got to get my boss to pay up for proExr so I access the exr layers in photoshop for any still images.

So what I am trying to figure out is since you use proexr why/how do you use the scripts.
Do you output to a large vrimg, use your scripts and how would this work for animations?

or do you output to multilayered raw exr and then use proexr in which case why do you need the scripts?

[Shimakaze]

I'm not certain about this (Vlado can probably answer better then me) but since vrimg files contain the raw render data they might not be compressed at all which would account for the difference in file size. they might also contain additional info which isn't used in exr files.

If you use ProEXR you don't really have any needs for my scripts. I wrote them as a cheap alternative (free cheap enough?)
some of the Photoshop scripts are still useful though. in fact I might post some more later. Got some quite helpful ones for doing textures.
Which workflow you go for really depends on whether you like your render elements to be in a single file or separated. both has their advantages. about keeping vrimg files, ifyou can then do. But if you are short on space there's no danger in deleting them. just make sure you get all your rendered elements out first

[pg1]

Cheers Shimakaze that clears up the workflow for me and thanks for the scripts will definitely find good use for them the next time I am doing stills now that I just installed CS3 extended

[pg1]

Shimakaze you said that

It also gives you a perfect save of the data. (the compression used when saving to raw exr at rendertime is a lot more effective then the compression used when you save it manually from the frame buffer. Try it.)

Can you explain this a bit better. I can output an image through the frame buffer and its 3.3mb and output the same image as a raw exr. and it is 1.2 mb and now I can access all the layers in the exr.

When I output through the frame buffer I save with zip 16 scanline blocks compression (the pro exr manual said that this is almost always the best compression format for cg)
I dont get a compression dialogue when I output to raw exr.
What is happening when I chose to output to a raw exr, is it using a different compression or change to image to half float?
How am I getting more (i.e layered file) at a smaller filesize

It seems too good to be true ,makes me suspicious

[Shimakaze]

It is indeed full float and I must admit I have no idea why it works that way but it does. Just something I've stumbled across really. From what I can tell theres no quality loss. Maybe Vlado can clear it up?

[m_hinks]

I was recently advised that a 16 bit float exr would be more than enought, with full control still to re-expose the image. 32 bit is not really required for most work. This was said by someone who works heavily with Nuke and exr's all the time. Im not yet happy to change my render settings, but maybe someone can explain on here? The file size will be much better with 16 but.

[tct70]

Works great, thanks!

[pg1]

Hey M_hinks

This is copied and pasted from the pro exr plugin manual. Im not sure I fully understand it myself but hope it helps

OpenEXR images are always floating point. But because full 32-bit floating point image data tends
to be unwieldy and compresses poorly, OpenEXR uses a 16-bit (“Half ”) floating point format that
matches what is found in some modern graphics cards. For digital images which generally have
some degree of noise or sampling error, the extra precision from a 32-bit float file is really meaningless.
Of course, OpenEXR does support full 32-bit float if you need it, as well as integer pixels
for non-image data like Object ID.
To store pixel data more effectively, OpenEXR utilizes image compression, typically lossless to
avoid any kind of image degradation, yet still provide a significant reduction in file size. When image
fidelity is less crucial, lossy compression and chroma subsampling are available to further
crush file sizes.

[m_hinks]

nice, thanks pg1. Maybe Ill have to try half float in the future. Could anyone add anything further to this subject?

[Shimakaze]

Aaaah Half float!! Didn't think to check that. When you save EXR files from the RAW section in VRay you get a half float picture, not full. Explains the size difference.
And yes, you don't need full float except in extreme cases.

[Chris]

I know I said it works but, It still does not work 4 me.

The only way I get all my render elements out of the vrimg file is when I convert to exr and open it with the exr pro plug-in for photoshop. What am I doing wrong? or is that a stupid question?