6
$\begingroup$

So, I know Blender can render out to .hdr format. And I am also familiar with Blender's compositor. What I don't know much about is the processes or algorithms used to produce HDR photographs, or what sort of node setup would need to be constructed for this - if even possible.

For clarity, here is an example of what this process looks like in Photoshop, compositing three exposures:

Photoshop's Merge to HDR Pro

I would like to do this with Blender's compositor.

What I usually do currently is render a Radiance HDR (.hdr) file out from Blender and then load it into a Mac application called Luminance HDR. There are some limitations with this though, so I am wondering if the same or similar results can be achieved within Blender.

Being able to use .hdr files would be ideal, but working from multiple bracketed exposures would be alright as well. I would like to know if there is a way to do this. Thanks.

| improve this question | |
$\endgroup$
  • $\begingroup$ What are your inputs and what output do you want? Am I guessing correctly that your inputs are bracketed exposures of environment, you want to add 3d element inside and output an hdr? Is this question about a compositing setup to merge hdrs? If yes it's not complicated at all if you know the exposures. If you don't know or they might differ that is a problem and you will have to eyeball stuff. $\endgroup$ – Jaroslav Jerryno Novotny Dec 23 '15 at 0:53
  • $\begingroup$ I hadn't thought as far ahead as a case where CG would be added to an HDR photo, but I'm glad you mentioned it. This question is for two use cases: 1. A 3D scene has been rendered and saved as an .hdr (or .exr) file from Blender, and we are compositing it in Blender to a final format of .png (for example) where the whole dynamic range is visible. 2. We have some photos of a subject at multiple exposures and want to composite them into a single HDR image. In either case the desired result is to get an "HDR look". $\endgroup$ – Mentalist Dec 23 '15 at 17:15
  • $\begingroup$ The basic steps are (1) leverage a proper HDR tool to capture the scene referred values and save as an EXR. (2) load into Blender and leverage the output view transform to map the scene referred ranges to the display referred domain. This can be via either (A) a view transform LUT or (B) the colour management curves. $\endgroup$ – troy_s Dec 29 '15 at 18:15
3
$\begingroup$

I think the term you're looking for is tone mapping.

In the photography world, "HDR" is used loosely to describe the entire process of shooting multiple photos at different exposures, combining them into a 32-bit image with a high dynamic range, and then tone mapping it down to an 8-bit image that you can display nicely on your 8-bit monitor.

In the 3D world, we can render 32-bit images directly and therefore don't need to go through the process of combining exposures to make an HDR image (we already have one), so we can proceed straight to tone mapping.

There is a special node just for tone mapping actually:

Node

And with a bit of tweaking you can do this:

result

If you'd like to play with it, here's the original .hdr file: http://adaptivesamples.com/files/original_untonemapped.hdr - this was made by taking several bracketed photos and merging them to an HDR, but the same method applies to raw 32-bit renders.

| improve this answer | |
$\endgroup$
2
$\begingroup$

I recommend you get Bloch's HRDI handbook, your two assumed questions cannot be answered in short, really; your question mixes topics for whole volumes and even while the answer is "yes", it does come with thw caveat of "why would you want to use Blender for getting HDRI look from photos?" and "why do you want to render bracketed results for compositing if your render output has all the lighting data in it?" :-) There are a lot of free sw alternatives for hrdi photo composition (and the required pre-steps), you get good insight from the book on what to use for what... And I am not associated with the author in any way.

| improve this answer | |
$\endgroup$
  • 1
    $\begingroup$ Thank you. 1. One reason for wanting to use Blender for this is the possibilities for animation. Another is for the sake of pushing Blender to see what other capabilities it has that people may not have thought to explore yet. Another is the convenience of keeping one's project all in one .blend that can be easily shared with anyone without needing to tell them they must also download or buy some other program (possibly only available on some other OS). 2. The rendering bracketed results case is for when you have actual photos you've taken and want to use Blender to do it (for reasons above). $\endgroup$ – Mentalist Dec 29 '15 at 15:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.