How to merge multiple videos to make HDR

[Tudor “ted” jelescu]

I see one slight problem with your setup- you will not be able to get identical images on your cameras. To get the result you want you need to shoot using a camera that can record RAW (or almost)like RED or Alexa. Then you can grade the image using all the information available.

Tudor “Ted” Jelescu
Senior VFX Artist

[Jim Arco]

Ted,

Thanks for your thoughts. Getting identical images is certainly one of the problems. The current budget doesn’t allow renting an Alexa or a RED – think 2 DSLRs and a beam splitter.

This is one of those what-would-happen-if projects. We’re getting good results merging matching frames in Photoshop. I just would like to simplify the process and was hoping AE already had a ‘merge to HDR’ command or work-around built into it.

[Jason Brown]

A combination of image sequences and blending modes?

[Tudor “ted” jelescu]

Nothing I know of that does that. The way I would approach this would be like any grading job, where I use animated masks to change the image. In this instance I would mask parts of one clip to reveal the other, using lots of feathering on multiple masks.

Tudor “Ted” Jelescu
Senior VFX Artist

[Walter Soyka]

This may be tricky to shoot, as even a slight misalignment with the beam splitter or any variance in camera sync will soften the final image.

I’d approach this with some heavy luma keying.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog – What I’m thinking when my workstation’s thinking
Creative Cow Forum Host: Live & Stage Events

[Michael Szalapski]

Someone did some work with a beam splitter and 2 cameras posted on Vimeo a year or two ago.
I believe they used luma keying. It’s the only good way I could think to do it. In my opinion it didn’t really look all that great. The image itself was quite nice, but all the edges where it transitioned between the two shots were all kinds of icky. You could see the feathering. It’s very, very tricky to get right.

I wonder if it would be better to import the videos as frames into Photoshop and batch them…

– The Great Szalam
(The ‘Great’ stands for ‘Not So Great, in fact, Extremely Humble’)

No trees were harmed in the creation of this message, but several thousand electrons were mildly inconvenienced.

[Walter Soyka]

The more I think about this problem, the harder (and more interesting) it gets.

[Michael Szalapski] “The image itself was quite nice, but all the edges where it transitioned between the two shots were all kinds of icky. You could see the feathering. It’s very, very tricky to get right.”

I suppose the solution there is to make sure the overexposure’s highlights and the underexposure’s shadows are adjusted to match as well as possible. Easier said than done, I’m sure — and it starts in production, not in post.

Another approach might be to split each exposure into a series of Ansel Adams-style exposure zones or luma ranges (using the footage itself with some Curves or Levels adjustments as a luma matte) and mix in each range, possibly with varying opacity or blend modes.

In all, I’d think that having three exposures (a good hero exposure, plus an underexposure to fill in highlights and an overexposure to fill in shadows) might be the easiest to post, because you will avoid the problem of forcing yourself to mix shadows and highlights to get good midtones.

All of these suggestions are global, meaning they apply uniformly across the image; some tone-mapping algorithms are local, meaning their results are dependent for each pixel on its surrounding pixels. An effect like Fixel Contrastica [link] might help you address local contrast challenges.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog – What I’m thinking when my workstation’s thinking
Creative Cow Forum Host: Live & Stage Events