HDR preparation

[EOSThree]

I don't think we hijacked his thread, but provided a good discussion that let me learn a thing or two, and hopefully everyone learned from it.

Those examples you posted(boston.com), although cool, move into the realm of clownish to me. I can look at every one of them and say: HDR. UWE's post says meh to me, again I think that with a little post processing the same look can be accomplished. Medley's post shows me more detail without that HDR look. I guess that's what poker has too. But with a little post processing I think I can attain the same look as he got with the merge.

Here's an example of what is possible without any HDR, at least that's what Marc says.
Marc Adamus. Although I suspect his hand blending techniques could probably be labeled HDR. If I can achieve anything coming close to Marc's work I'll be completely happy with or without HDR.

[Medley]

Then let's look at dynamic range. The definition of dynamic range is the number of steps between black and white. Dynamic range is presented as a ratio, and you never count black- you start with the darkest definable step that is not black.

So, you're right. The images I've posted here are NOT HDR images. Every image posted here is, to the best of my knowledge, an 8-bit image. That means it has a dynamic range of 1:255. Therin lies the rub of HDR: you can create a truly HDR image, but to post or print that image, you have to convert it back to a lower dynamic range. Remember when I said that how you discard information will largely define the success or failure of an HDR image? That's why.

Now, a Raw image, captured in 12 bits, will have a dynamic range of 1:4095, but it's presented in a 16 bit, 1:65535, format. So for any Raw image, you could insert 15 tonal values between every 1 value separation in the original image. So I could incorporate 15 images, all with different tonal values in every Raw image, and never break the 16-bit barrier.

So if I have a 6 stop image, and I shoot every stop to the right, I end up with six images that have 50% of the detail in the original image, and can combine those to create one image that has 6x50%, or 300% (3 times) the total information of the original image. But I've still only utilized 20% of the image's 16 bit potential.

EOSThree, you stated that increased definition is a by-product of HDR. I submit that it's a direct result of HDR. The only difference in your theory and mine is where you're adding the information. You advocate adding on the ends of the image, which is a common usage. In my examples, I added it in the middle of the image. The truth is that you can do both, it's not an either/ or choice.

The last, and probably least understood aspect of HDR imaging is that HDR images never remain HDR. You've got to convert it to a lower dynamic range to make use of the image in print or post. The trick in doing so is to discard enoough information to make the image fit into a 1:255 dynamic range, while maintaining the localized contrast between pixels. I'm still working on perfecting that aspect of it.

It's a learning process, and I'm still on the curve myself. If you go back and look at my posted examples closely, you'll notice a decided shift in the saturation of the blue tones. That stumped me for a long time. Turns out that it's caused by not using a UV filter on my lens. Light from the UV spectrum was pushed up into the blue/violet range, causing the shift. Lessons learned.

- Joe U.

[Medley]

By the way, if you're looking to learn the techniques Marc used/ uses, I highly recommend "The HDRI Handbook: High Dynamic Range Imaging for Photographers and CG Artists" by Christian Bloch. link (Amazon): http://www.amazon.com/HDRI-Handbook-...2044838&sr=1-1

- Joe U.

[EOSThree]

Thanks for the link. I love Marc's work, I hope to take a two day hike with him this fall.

[EOSThree]

Here is my attempt using 1 exposure and some tweaking in LR


I apologize for using your image but I feel it provided the best evidence to my pov. I started with your image cropped off the HDR version, then tweaked it in LR to look like your HDR version. I really don't think it's fair to take an unprocessed mid image and compare it to a fully processed HDR image.

I didn't use any layering, stacking, HDR, dodging or burning in creating this photo. I fear photobucket reduced the quality quite a bit, but I hope the attempt comes through. It looks remarkably like your HDR and could have been quite a bit better if I would have had a RAW file to work from. The biggest differences I see here are from the .jpg compression and the further compression from photobucket.

I guess what I am saying is, I am still not convinced of the need for more than 3 exposures yet. But each to his own, I'll keep doing what I do, you keep doing what you do.

[daq7]

Yawn. SInce I need to add more text to satisfy the text limitations. Let me repeat. Yawn.

[OldClicker]

Here is my attempt using 1 exposure and some tweaking in LR

Yours does not have the same detail in the shadows. For a specific spot, look at the rim of the bottom wheel on the right as it goes up into the shadow. Yours disappears, the HDR is visible all the way up. – TF

[OldClicker]

Yawn. SInce I need to add more text to satisfy the text limitations. Let me repeat. Yawn.

Route 66 bores you??? - TF

[EOSThree]

I'll agree with that, I would bet that if I did an HDR with three exposures I would regain that shadow detail there. You have to admit though, it's a reasonable facsimile of Medley's, especially for one done with a compressed jpeg and displayed through photobucket.

It's presented as "look what a bunch of exposures can do for HDR" when in reality it's look what HDR can do with a lot of other PP too. I have tried multiple times to duplicate that look with just using photomatix. It's not just there until I do a bunch of work with PS and LR too. So I still remain unconvinced.

[Medley]

No, I don't mind you using my image. Yes, it IS a reasonable facsimile, especially since you used the smaller image from Imageshack (more options than photobucket).

It's relatively easy to provide a reason facsimile of an 800x1066 size posted image. Provide me a reasonable facsinile of my 20x30 inch print, I'll be more impressed. Even so, it's still relatively easy for me to tell the difference between an image with the illusion of sharpening (provided by increasing the contrast at the high contrast edges), and an image that is visibly sharper due to having more detail- even in these smaller images.

I have tried multiple times to duplicate that look with just using photomatix. It's not just there until I do a bunch of work with PS and LR too. So I still remain unconvinced.

But you see, THAT was the whole point of this particular experiment, because that's exactly what I did. The only edits done in photoshop were to combine the two images, switch to sRGB color, and reduce the image size. Everything else was done in Photomatix. For what does it benefit us to add scads of detail in Photomatix if we're just going to send the image to pp for a bunch of destructive edits?

So, as you say, you keep your methods, I'll keep mine.

- Joe U.

[jeanlee411]

I think these two are excellent candidates for landscape HDR - especially the rocky canyon with the blazing sun and deep shadows. Both still have shadows and highlights with no detail and probably could have used another exposure in both directions. - TF

I believe that more shots probably averages out random noise better, but I cannot prove it. Most of my HDR work now involves taking ONE raw shot and manually blending exposure layers from it. I still often use a photomatix layer as one of the blended layers, but I haven't been doing a lot of shots with HDR software. I agree that for the vast majority of shots -2,-, and +2 are good enough.

[n8]

considering how subjective hdr is, I'm amazed by how much debating there is regarding the technique. With that said, I usually bracket 3 shots 2 ev apart and render in photomatix.