Composite Image Example

On my recent Diamond Grove Sunset post a couple of you mentioned that the prairie at the horizon line seemed uncharacteristically light. In the image above I've posted three images. From left to right they represent the original unprocessed image of the prairie (f/g), the original unprocessed sky image and the final composite image.

I've included the exposure info under each image but to summarize there is a 3.5 stop difference between the exposures for the f/g & sky images. Both shot at the same focal length, aperture and ISO.

In the final image I used the following CS4 layers settings:

Brightness/Contrast: Bright = 24 Contrast = 20
Vibrance: Vibrance = 21 Saturation = 0
Curves: Small midtone darkening adj
Selective Color: Added + 10 to each of the Red/Yel/Gn/Cy/Bl/Ma/Wh channels (a custom setting I encorporate to all images)

Based on what you see here and the information provided would you still consider this image to represent an inaccurate representation of the scene before us that night? How should I have processed this differently to achieve the darkness on the distant horizon that you feel this is missing?

When I view these images it appears the final image is an accurate representation of the scene that night without overly optimized adjustments.

I thought this would be a good opportunity for people to chime in and express their thoughts and ideas using the example above. I look forward to hearing your opinions. Thanks!

Thank you for showing us how you make some of your magic. The results speak for themselves! Simply beautiful.

Great example.

I have not had much luck with CS3 but have gotten better results from Photomatix and doing a bracket of 5 images all one stop apart (so -2 through +2) gives me 5 stops to work with. Pretty easy to do in Ap Priority and using exposure compensation.

Wayne, was this a photomerge (or does cs4 call it something else?) If not, perhaps 3.5 stops is too much contrast for a straight overlay. What I often do is a gradient mask. It usually works pretty well where the horizon is straight, much like a graduated nd filter

Here is a quick and dirty attempt with your sample.

Wayne, I don't think the distant horizon lighting is off by much. Regardless, if that's the way you like it, then I would go with it. Could be my eyesight is getting worse in low light conditions, but here's a quick and dirty composite of one of my comps from DP last week. I used a painted mask - more time consuming, but I like the added control...

Thanks for responding Ken. I did not use Photomerge or Merge to HDR in CS4. I probably should hae but for this image I simply cut the sky in the prairie scene and drug it onto the sky image to create the composite.

I don't believe that I have clearly stated what my question is here. Your example is fine and I understand why our minds would tell us that your version is the correct representation of the scene. But if you look at the even distribution of light in my prairie image there is no shadowing at the horizon line. While I did expose to the right, the prairie was evenly illuminated and not clipped. Why introduce shadowing/darkness where none existed?

I also forgot to mention that these two images were exposed just 31 seconds apart with the sky being shot second. I appreciate your input. I'm not trying to be argumentative here, just trying to better understand others perceptions to give myself a broader point of view.

Thanks!

It looks like you and Bill must have been at the same sunset.

I think I understand your question Wayne. Specifically, the simple answer for me is that I think that much e.v. difference in a composite doesn't look quite natural without some sort of gradient to help it.

In general, you ask quite rightly why introduce darkness when none exists. I might ask why introduce light where none exists? We have these great computers with lenses on them that allow us to take pictures in the dark and see details that often we couldn't see with our eyes. It might be amazing but is it natural, especially in landscapes?

That's my opinion, what are some others?

I have CS3 still, I haven't messed with CS4 yet but I can't hardly stand the Merge to HDR in CS3. The double exposure technique seems to work better and I agree with Bill the painted mask is my preferred method when I am being serious. I will use a layer mask technique when I am being a bit lazy.

Wayne the shot looked fine to me. I think we all tend to get a bit caught up in technical stuff when we really should be just looking at a beautiful image and appreciating it. I say good job.

Wayne, first let me say I like your original perspective. I have used a combination of all the techniques that have been discussed by you, Ken and Bill. The bottom line is the decision on which method really depends on the photo you are working with and your sense of what you saw when you took the photo. Most of my blends now start with a gradient mask that Ken mentioned. The great thing about this technique is you can control the amount of gradient from a hard gradient to a very long smooth gradient, based on each situation. After applying the gradient mask, you can then refine the photo by painting on the mask like Bill suggests. A great discussion, thanks for your effort.

Too much Photoshop!!! At the next gathering I am forcing everyone to sit through a two hour Velvia 50 slide show.

Tom I agree that different situations require different methods. If we all did everything the same our photos would all be the same. Who wants that? Not me. I love looking at Tom Kennon's shots. I know they are Tom's. Which is a good thing. Which goes for Tim, Bill, RevCurmudgeon and just about everyone else that posts here as well.

I have actually been on a jpeg kick lately. It keeps my processing time down and is just more fun to shoot for now anyway. It allows me to stay up late and rant like this instead of playing around in photoshop.

Thanks everyone for expressing your opinions, they are much appreciated.

Ken your question of "why introduce light where none exists" really hit the nail on the head for me. The reason I have come to add light to a "dark" sunset image is because my workflow is geared toward a printed output. I found through seeing other peoples reaction that if I printed an image with the darkness of say Bill's example here the image would only get a cursory glance. A brighter image with vivid color almost always makes them stop and look. Maybe if I had better printing equipment and improved techniques I'd get different results.

So I guess the thing that was confusing me was rooted in the fact that from my experience people don't usually respond as positively to a dark, accurate representation of a sunset as they do one that has been brightened for printing.

This has been informative, thanks for your input. Have a great week everyone!