2015 Project 52 Preparation

[FrankMi]

One of the things I'd like to try for the 2015 Project 52 is to be able to capture and effectively display reflections on the surface of a pond or stream and in the same image, have the pond or stream bed in focus.

This would require being able to focus on both the underwater objects and those reflected by the surface of the water. Usually, the reflected image has a much longer focal length so it would require either a very wide DoF, (which might be possible by using a wide angle lens), or by taking two images, one focused on the surface reflection and the other focused on what is under the surface.

So far, so good, but assuming the second method is used so that the DoF can be controlled, what is the best method of combining the two images?

To do some testing, I took images at my desk, which has a glass top. In the first image, you can see the surface of the desktop in focus, more easily identified by dust, smudges, coffee stains... I need to clean that up before my wife notices what a slob I've become!



In the second image, the focus is on the window blinds reflecting off the desk top. This one looks much nicer. Its amazing how easy it is to clean up a messy desktop with just the twist of the focusing ring!



You can click on to open the images in Lightbox, then use the arrow keys to switch between them to examine the differences more easily.

My first feeble attempt to merge the two surface views is to simply put them in layers and set the Opacity of the top layer to 50% so that about half of each image's pixels are blended.



We can see elements of both the surface and the reflection image elements this way, but neither rendition is very clear.

Is there a better way to blend these two images so that both the surface and the reflection are clearer in the combined image?

If you would like to experiment, please feel free to snap a copy of the images, give it a try, and post your steps and results.

C&C most welcomed!

[Shadowman]

You've left out the other portion of your experiment, having the stream bed in focus or did you not mean beneath the surface; such as rocks or fish below the surface?

[FrankMi]

You've left out the other portion of your experiment, having the stream bed in focus or did you not mean beneath the surface; such as rocks or fish below the surface?

You are correct John. Although it is not practical to shoot below the surface in this test, the issue and result is the same, how to find a way to combine images with two different focal lengths. In the case of a water surface, the goal is to get the objects below the surface and those reflected from a much further distance away both in focus at the same time. The reflection might need to be focused at infinity and the scene just below the surface might be only a few feet away.

[Shadowman]

You are correct John. Although it is not practical to shoot below the surface in this test, the issue and result is the same, how to find a way to combine images with two different focal lengths. In the case of a water surface, the goal is to get the objects below the surface and those reflected from a much further distance away both in focus at the same time. The reflection might need to be focused at infinity and the scene just below the surface might be only a few feet away.

The table in your shot looks too thick for the experiment, perhaps you have something else to use.

[bnnrcn]

Hi Frank , why don't you do focus stacking with the images above instead of merging them with % 50 opacity?

[Geoff F]

Yes, but only if you have software capable of working with layers and masks.

Just out of interest, I will download those images and see if the auto combine will also work.

[ajohnw]

I just used an Enfuse gui. A windows version and an LR plugin is available but there is a charge for the LR one.



It stacks HDR and Focus. Never used it for focus before but set both of the exposure weighting settings to 1. That bought up the smudges on the glass more rather than HDRing them. Left the contrast weighting at zero. I suspect that gets more useful when there is a lot of out of focus fuzz. I also set auto align - not needed, optimise centre shift and optimise field of view - that takes care of magnification change between shots. The gui I use generates a reduced size preview. This saves a lot of time when large images are processed.

One other way of doing this with layers is to simply put one above the other and add a layer mask to the top one but scaling can be needed in more extreme cases.

John
-

[Geoff F]

Tried it in CS5. Auto Align followed by Auto Edit and it worked fine but, yes, it did select the table smudges in fine detail. However, I could edit those layers slightly to manually tweak the auto masking.

[ajohnw]

I would be interested to see the results Geoff. I did a bit of whole image mods on the one I posted. Bit of highlight compression/reconstruction and lifted the blacks a bit plus a contrast adjustment that is similar to sharpening.



John
-

[FrankMi]

The table in your shot looks too thick for the experiment, perhaps you have something else to use.

Hi John, I need to get out and find a water surface that will work. As I don't yet know if there might be some shooting technique involved I've been working on the post processing side for now which is where I think there is the most possibility for getting both focal planes in the same part of the image.

[FrankMi]

Hi Frank , why don't you do focus stacking with the images above instead of merging them with % 50 opacity?

Hi Binnur, focus stacking attempts to select the sharpest parts of multiple images based on the highest contrast edges. For most focus stacks, that means starting at the image with the closest focal plane in focus and working forward toward infinity. The problem with reflections is that they can be just as sharp as the object they are reflecting from, or in the case of water, what is just below the surface. The result can be a sharp surface (or just below the water's surface) right beside a sharp part of the reflection, but not both in the same spot.

This image was created from 12 focus stacked shots and to get it to show a consistent surface in one result and a consistent reflection in another result, I had to manually merge the layers.

It is similar to shooting a scene behind a glass and getting unwanted reflections, only here, we want to clearly see what is behind the glass and clearly see the reflection in the same image. Does that make any sense?

[FrankMi]

I just used an Enfuse gui. A windows version and an LR plugin is available but there is a charge for the LR one.



It stacks HDR and Focus. Never used it for focus before but set both of the exposure weighting settings to 1. That bought up the smudges on the glass more rather than HDRing them. Left the contrast weighting at zero. I suspect that gets more useful when there is a lot of out of focus fuzz. I also set auto align - not needed, optimise centre shift and optimise field of view - that takes care of magnification change between shots. The gui I use generates a reduced size preview. This saves a lot of time when large images are processed.

One other way of doing this with layers is to simply put one above the other and add a layer mask to the top one but scaling can be needed in more extreme cases.

John
-

Thank you for looking into this John. I may be chasing a white rabbit down a hole with this attempt. When I closely compare our two images they are almost identical and it may be the best we can do under these circumstances. I've tried starting with a 50% blend, then enhancing just the parts of each focal plane that have the most detail. That seems to help but still leaves the inability to get any more that a ghosted rendition, be it a little cleared one.

Here is how it came out with a selective blending of the two focal planes.



It may actually be easier to blend using a pond/stream as what is under the surface is already blurred by the water, and the reflection (unless the water surface is mirror smooth), will likely be distorted and at a much higher brightness and contrast.

I think I'll need to get into the field to continue this endeavor.

[FrankMi]

Yes, but only if you have software capable of working with layers and masks.

Just out of interest, I will download those images and see if the auto combine will also work.

Thanks Geoff. I'm looking forward to see how you make out.

[FrankMi]

Tried it in CS5. Auto Align followed by Auto Edit and it worked fine but, yes, it did select the table smudges in fine detail. However, I could edit those layers slightly to manually tweak the auto masking.

I'm using CS5 as well, Geoff. Are you using the term Auto Edit to refer to the 'auto' functions in general?

[FrankMi]

I would be interested to see the results Geoff. I did a bit of whole image mods on the one I posted. Bit of highlight compression/reconstruction and lifted the blacks a bit plus a contrast adjustment that is similar to sharpening.



John
-

Good point John. I hadn't taken it to the next step and played with selective sharpening as yet.

[ajohnw]

I suspect you will get better results under more extreme situations with specialised software Frank. Kathy who is sometimes on here reckons Enfuse did have problems then. A lot of people use Zerene Stacker. Some use Helicon focus. Arguements rage about the merits of each. Nobody I am aware of who is doing this sort of thing seriously uses PS etc. That statement is probably heavily biased to photomicrography where dof can be unbelievably small. I suspect Zerene may have a bit more of a learning curve than Helicon. Last time I looked Helicon was more expensive. That might add to it's appeal.

John
-

[Shadowman]

Hi John, I need to get out and find a water surface that will work. As I don't yet know if there might be some shooting technique involved I've been working on the post processing side for now which is where I think there is the most possibility for getting both focal planes in the same part of the image.

Frank,

For in the field work the cpl should get you the focus through the surface of the water, you might have to remove between shots so that it doesn't influence your surface reflections.

[Geoff F]

I'm using CS5 as well, Geoff. Are you using the term Auto Edit to refer to the 'auto' functions in general?

To be strictly correct, Frank, Auto Align option followed by Auto Merge; under the Edit Menu.

I don't think I kept the resulting image but can easily do it again, only a few clicks needed.

[FrankMi]

I suspect you will get better results under more extreme situations with specialised software Frank. Kathy who is sometimes on here reckons Enfuse did have problems then. A lot of people use Zerene Stacker. Some use Helicon focus. Arguements rage about the merits of each. Nobody I am aware of who is doing this sort of thing seriously uses PS etc. That statement is probably heavily biased to photomicrography where dof can be unbelievably small. I suspect Zerene may have a bit more of a learning curve than Helicon. Last time I looked Helicon was more expensive. That might add to it's appeal.

John
-

I'm certain that you are correct John. So far, my stacks have not exceeded 12 images, as was the case here. That can easily be managed in CS5, including the manual cleaning up of the artifacts.

However, in preparation for doing low-level shooting for P52, such as the view that would be seen through the eyes of a cat, small dog, or squirrel, I have been working with Helicon Remote on a cell phone to drive the camera for focus stacking. My first test captured 751 stacked images, WAY to many to be practical and there is no effective way to merge that many in CS5. I've played with the sample version of Zerene Stacker and read the comparisons between Zerene Stacker and Helicon Focus. On the cost basis, Zerene may be the better bet for me as I'm not planning to dedicate a lot of TLC (Time, Labor, and Cash) to macro imagery.

[FrankMi]

Frank,

For in the field work the cpl should get you the focus through the surface of the water, you might have to remove between shots so that it doesn't influence your surface reflections.

An excellent point of using the CPL John! It could mean running two sets of stacked images, one with the CPL and one without but it may be just what is needed, particularly to try to achieve the blending to the two focal planes. Thanks for the reminder!