Exposure Blending or HDR

[ArmySoldier777]

Hey everybody, this is more of an opinion topic than anything else I guess. Or maybe there are real differences that I don't know about. That's what I'm here to find out.

I've been reading posts on here as well as other sites regarding HDR images. The feel I've gotten is that most people, when they see "HDR", they automatically think "Another overdone, cartoonish image", before they even see it. And I'm not here to do get into that whole discussion about what real HDR is, I believe that both the subtle, realistic, as well as the artistic have their places, and there have been images in both categories that I have enjoyed looking at.

But here's where the thread title comes in. I've seen a lot of people that refer to realistic HDR as "Exposure Blending", which is what HDR is, they just don't like using the term "HDR" because of the reasons I stated before, that people automatically start thinking "over-processed image". I just want to make sure that I'm not mistaken and they really are the same thing, just two different terminologies being used for it.

And no, this thread isn't really an important issue lol, just looking to satisfy a curiosity.

[Krusty79]

I've always thought Exposure Blending was taking images made for the dark and light part of a scene and using a layer mask to blend them. HDR is usually done by the software with your bracketed shots. I then divide HDR into "realistic" and "non-realistic" HDR. Am I wrong?

[kcoppock]

That's the way I understand it. All* "HDR" images you see currently are not actually HDR (32-bit), but have been tonemapped back down to a standard 8 or 16-bit image by compressing the tones. This compression process is where the "realistic" and "unrealistic" types come into play. Just like any image modification, it can be taken too far, or it can be done judiciously to get a more realistic result.

(*unless you have an INCREDIBLY expensive HDR-capable display)

[inkista]

To me, exposure blending/fusion is a different mathematical technique vs. HDR.

In HDR, bracketed shots are taken, and then the values are remapped along a larger dynamic scale, and an HDR-file is created with the entire dynamic range numerically represented. The image from the HDR file is then tone-mapped down into the smaller dynamic range of regular graphic file formats/displays, and how the tone mapping is achieved gives the widely varying looks.

Exposure fusion, otoh, with software tools like enfuse, operates a completely different way. It goes through a stack of bracketed images, but does not "remap" things numerically, but rather selects a specific pixel from one of the member images, based on selection criteria (in the case of enfuse, there are three criteria: exposure, contrast, saturation, which a lot of people mistake for tonemapping controls), and then uses the "best" value from the members in the resultant image: there's no numerical manipulation to a different range, but rather a filtering down to the "best" pixels. There's no HDR file, and no tone-mapping, and the resulting image is directly created in a graphical file format that can be displayed.

The reason exposure fusion tends to look more natural by default is that you can't create a pixel value that doesn't actually exist in one of the member images. With HDR, because of all the numerical sliding around, you can easily create a value that wasn't there in the original member images.

Two different techniques: two different names. This is why "exposure fusion" is also called out separately as something different from HDR in Photomatix.

Here's an example of exposure fusion with enfuse, using the LR/Enfuse plugin for Lightroom.



No HDR; no tone-mapping. Just feed it the three images, and the result pops out. I only applied a little bit of curves adjustments to get the final result, and that was it.