OPENexr, Still not sure about the benefits for still images

[TomD_Arch]

So I've recently started to experiment with OPENexr format. I've done the research, got the alpha thing working in photoshop, I can even get render elements applied to the exr before hitting render. I open the file and have a really nice 32-bit image. What next? Do I just go straight to the HDR Toning adjustments? How do the render elements play into the file I just saved? Am I supposed to still save those and use them in post production for compositing, or were they just for the rendering calculation and now they are all influencing the .exr file I saved? I can't apply them with filters unless I save the image down to 16-bit. I keep reading about how everyone is using OPENexr and all the benefits it presents, so I feel like if I'm not using it...I'm an idiot! I get it, they store a ton more information. I am just not sure how to leverage it. Can someone please help me out of idiotness of the first degree when it comes to OPENexr's?

[Chris MacDonald]

It's no different to saving all your render elements into one file (that happens to be 32 bit). How you use those render elements is up to you - how did you use them before?

 

I don't use .EXR because Adobe seem hell bent on not offering full support for multichannel exr.

[Scott Dombrowski]

My main use for 32bit .exr is being able to not worry too much about exposure in Max. I want to get it close, of course, but if the raw render is not exposed correctly, it's no big deal to slap an exposure adjustment layer on in Photoshop and dial it in. You can do that with 16bit as well, so the real benefit of 32bit for me is with multiple light groups, especially for interiors. If you split each group of lights into separate 32bit passes, you can reassemble them in Photoshop using the linear dodge (add) blending mode and then put an exposure adjustment layer on each. Now you can adjust each light group's exposure just like it was on a dimmer switch. That's a lot easier than tweaking it in Max.

[TomD_Arch]

How you use those render elements is up to you - how did you use them before?

 

I don't use .EXR because Adobe seem hell bent on not offering full support for multichannel exr.

 

So far in experimenting and render elements, I've added the usual suspects: Diffuse, Output -DDI, -DII, -Reflection, -Specualar, -Transp. and -Transl.) I saved all of them as 32-bit individually and set them to 16-bit with in the actual saved render.exp output file.

 

At this point if I take just the .exr render and open it in Photoshop I then go to the HDR Toning and adjust the settings from there for my color correction.

 

I have also toyed around with bringing the individual elements into photoshop and compositing them using the "Linear Dodge(add)" filter on each of those elements applied to the diffuse, but then I have to turn them so far down, like 5 or 10% to get to anywhere decent looking, it seems to negate the ability to adjust the "Range" aspect of the HDR.

[Francisco Penaloza]

EXR file was build to work in large production, there was a need of a compression friendly file, a few options how to do that also with the option to output high bit depth and also being able to put separate passes within a single file for easy streaming on networks. Also a flexible format that keep growing in features.

 

Thankfully this was created as open source so now we all can enjoy of those options.

For me the major advantage is the large amount of light information and the use of saving several passes within a single file. When you work in animations streaming a single file work better than loading hundred of frames for each pass.

 

As mentioned by Scott, saving at 16 bit or 32 bit give you a lot of flexibility to do lighting changes in post. Usually you can easily resolve those issues that your exterior is too bright in a interior shot or vice versa, when your interiors looks too dark in a exterior shot.

 

Also when you work on night or dusk shots, you can nicely "fake" interior lighting with the refraction pass.

 

Sadly Adobe has not fully embraces 32 bit workflow within Photoshop, but you could use After Effects, or any other pro app such, Fusion.

 

Having said that while working on photoshop I transform the beauty pass in to smart object, then I change the bit depth to 16 in Photoshop, this will enable all the tools that didn't work before on 32bits. If I need to do any major lighting adjustment to my beauty pass, then I just jump in to that smart object and voila, you are in 32 bits again.

 

There is a paid plugin (Pro exr) that will let you load multipasses inside of photoshop, very easy to use.

If you render with V Ray or Corona, you only need to put the saving output for the main image (Beauty), all the passes will carry on the same file no need to save them separate. Witch save you time and eliminate errors of missing or wrong saved files

 

 

The files will get larger as you put more passes and use 32 bits, but you also have the option of saving 16bits, that most of the time will be more than enough, if you set your main rendering close of what you need it.

If you want to save to 16 bits or 32 bits, you should not burn the gamma in to the image. You could burn the color mapping, but this cut the lighting information. Of course you could mix and match all this options but that's the flexibility that EXR give you compared to TGA, PNG or Tiff

[Scott Dombrowski]

Turning the 32bit pass into a smart object is a good idea, I'll have to try that.

 

Vray has the VrayLightSelect render element that sets up the lighting passes more automagically. Corona will have similar functionality in the next version in a few months. I gave up on mental ray so I don't know what they're doing these days. But for mental ray and Corona, here's how you can can manually set up lighting passes for ultimate lighting control in Photoshop:

 

For an interior, say you have some ceiling can lights, some 2x4 fluorescents, some wall washers, and daylight. You'd turn your can lights on, and all the other lights off (including daylight), render, save as 32bit .exr. Then, turn the can lights off, turn the fluorescents on, render again. Rinse and repeat for each group of lights you want to control. I don't do this often, but if it was, I'd set up the render passes in a render pass manager or in State Sets to make the rendering part more automatic.

 

Once you render out your light passes, then you can do a File/Scripts/Load Files into Stack... in Photoshop to load all the different light pass .exrs as layers in one file. Change the blending mode on all of them to Linear Dodge (Add). Now your image should look just like the normal beauty pass you get if you rendered all the lights on at once. Add an Exposure adjustment layer to each light pass, and you'll be able to increase/decrease the brightness of those lights just like they were on a dimmer switch. It's pretty neat.

[Francisco Penaloza]

not much time to write right now but pretty much the same workflow that Scott mentioned, will work fine for Mental Ray.

Mental Ray does not have the light select pass like V-Ray, but you could render just the lighting pass and work with it.

In a nut shell it is a little more complex because you'll need to work with RAW passes to get correct reflections/refraction affected by the lighting but at the end you'll get the same result.

I write a little more about it hopefully soon.

[MarcellusW]

Thomas,

 

The corona tutorials on youtube explain step by step the process in photoshop.

 

 

FYI: I use Photoshop CS4 and the EXR format was not working with the blend modes in the tutorial. I followed the tutorials exactly but changed the format to 32-bit TIFF for the render elements, and the blending modes worked just fine. Not sure if this is fixed in newer versions of photoshop.

 

Cheers.

Edited May 28, 2016 by MarcellusW

[TomD_Arch]

@Marcellus- Thanks for posting link, even though I am working in mental ray the '2 of 3' was great and useful to see.

 

@Scott D- Played with this technique you described. Works, but the Gamma adjustment seems to have a limited range within which it affects the dimming ability. Seems like a lot of extra work and render time. I supposed if you really needed to be able to take control of light dimming for a shot this would be helpful, but once again if that's all I'm able to do with .exr, then why does everyone love it? Don't get me wrong, I'm picking-up what your putting down, but I'm still not convinced it's going to give me a winning hand.

 

@Francisco- Yes, the reflections on glass are hard to deal with. Not sure how you would go about using RAW render element outputs to deal with it. Can you advise more on this? I am now wondering if investing in the $50 ProEXR photoshop plugin would be a good investment.

[Francisco Penaloza]

Well no need to convince you to buy a software really or use a file format, each one of us have a different workflow and type of work.

For me it help me in my workflow as explained early. Several of my project the clients are a pain in the neck and that extra lighting info really come useful to adjust colors, lights and so on.

For reflections you can just put it in top of your beauty and adjust transparencies, You can do color correction on it too, now if it is too reflective, then you can use the translucency instead, this will show more of the inside. Also you may need to adjust colors and brightness.

 

Or you can start from scratch and use Diffuse, GI, Shadows, reflections, refraction and such. a pass of AO for good measure too

[MarcellusW]

Thomas,

 

You're probably more experienced than myself, but I experimented in Mental ray for a good year of free time to learn how to render. Does it work? Yes. Does it work without troubleshooting a greater portion of the time? For me, No.

 

My reason for using it was financial, it was free with 3ds Max. And people can make the final images look really good if they know what they are doing. However, I can say that when I switched to corona I saved a lot of time and I had a lot less headaches, and get better looking raw renders from the start.

[ismailrebbane]

So I've recently started to experiment with OPENexr format. I've done the research, got the alpha thing working in photoshop, I can even get render elements applied to the exr before hitting render. I open the file and have a really nice 32-bit image. What next? Do I just go straight to the HDR Toning adjustments? How do the render elements play into the file I just saved? Am I supposed to still save those and use them in post production for compositing, or were they just for the rendering calculation and now they are all influencing the .exr file I saved? I can't apply them with filters unless I save the image down to 16-bit. I keep reading about how everyone is using OPENexr and all the benefits it presents, so I feel like if I'm not using it...I'm an idiot! I get it, they store a ton more information. I am just not sure how to leverage it. Can someone please help me out of idiotness of the first degree when it comes to OPENexr's?

 

Hi, Thomas,

 

Check this page Out: http://cgalter.com/resources/

 

You'll find all you need to know about LWF, Multipass compositing etc..

 

best regards,

 

Ismail