Perspective on HDR

Just a little technical note for those interested in doing this underwater. I have set up an underwater menu mode on my D700 that allows immediate access to the bracketing adjustment when I depress the AF lever on my Subal housing. This allows me to quickly do a bracketed series if I come upon a scene that I think may be interesting for HDR while diving, or even if I wish to bracket for any other reason. In the pic above, I had not planned it, just found a place I thought was worth a try where I could stabilize my housing against something. Ryan (Reef Ryan) helped me work it out in the menu settings.

Acceptable I think.

I would say so. From what I have now seen from you and others in the early attempts at underwater digital 32-bit HDR compositing [... and from what I have done myself, as some folks have been kind enough to send me some raw files to play with...], the technique is without a doubt promising and will ultimately bear fruit. I also think there will likely soon be technology to shoot high quality, high resolution stills at very high frame rates that will make the stitching essentially seamless.

I would also respectfully disagree with Giles on the applicability of this type of underwater situation for multiple exposure, "Wykoffian" photography. We don't claim to have invented the technique of compositing images - although some of the folks here certainly could be pioneers of the technique underwater.

Lighting the foreground with a blast of strobe light is one solution to the problem. Taking multiple exposures - one for the sunball and one for the foreground and one for the DR in between - is a perfectly valid solution to the same problem and also offers the additional benefit of some very interesting local contrast that is not as easy to tease out of a single 14-bit RAW file as it is from a 32-bit composite. Photomatix - just like Photoshop - is just a tool, and just because some people make garish images that just don't work doesn't invalidate the tool itself. Used well, one can produce final images that are closer to what the eye can see than what the sensor can see in a single pass and which fulfill all the other criteria of a "good" and recognizably photographic image.

It certainly demands a lot of post-production work and planning to get right, but so did many of the things we used to do in the darkroom to solve exactly the same problems in B&W landscape imagery.

I happen to enjoy it tremendously in my topside work, think I have produced a few topside keepers and now understand the PP workflow well. I look forward to trying it underwater with my D700 and hope to get the chance in 2009 to put some of this thought and practice on land into production underwater.

I imagine my attempt will involve a tripod, magic filters and some attempts to use some higher ISOs (800-1600). I'd also like to attempt some square HDR panoramas with two frames from the 17-35mm in portrait.

Here's an example of a scene I processed both using HDR compositing as well as "traditionally" and illustrates some of the trade-offs. There is a distinct lack of sharpness in the foreground of the HDR. I think that came from a final layering of the normal over the HDR to try and fix some ghosting in the forest canopy. In hindsight, I probably could have accomplished that last layering while preserving more foreground sharpness, but as it is I think I prefer the "normal" over the HDR. Were the HDR sharper, it might "win".

Here's the "normally" processed image:

Here's the HDR:

Hey Jeremy,

I'm not seeing the normal (first) image.

Hey Jeremy,

I'm not seeing the normal (first) image.

Good Morning, Jeff!

That's odd that you can see one and not the other ... I could see both fine and they were linked to Flickr in the same way ... regardless, I deleted the first image link and re-inserted it ... hopefully you can see it now.

Top' o the morning to you sir.

Yes overall the first is better because of the reasons you mention, but I prefer the shadow detail in the HDR version. Might be worth trying again. Maybe blending in the shadows some.

Yes overall the first is better because of the reasons you mention, but I prefer the shadow detail in the HDR version. Might be worth trying again. Maybe blending in the shadows some.

Exactly. There is just the kind of shadow detail in the HDR that one hopes to pull out using this technique.

I think the best of both can be achieved and will continue to work on this one. When I combined the two layers at the end, I should have masked the "normal" that was behind the "HDR" layer in the bottom-half in the opposite manner from the top-half. I was so concerned with getting the top half to work right that I didn't pay enough attention to the bottom.

One more try: