Project 52 by Frank Miller

[FrankMi]

This week’s exercise is to see what we can do when we need to clone out something in an image but can’t get good results with just the Clone Tool alone.

When I saw this building I liked the architecture but wasn’t thrilled by the distractions in the scene. The folks I was traveling with were moving on so I didn’t have much time to plan but one thing was obvious, the distractions had to go if the final result was to be what I was looking for.

I noticed that most of the distractions were long and thin – the signs, streetlights, traffic signals, people and shadows. Others were more boxy, like the trash can and the automobile parked behind the building.

The problem was that if I wanted to preserve the detail in the building that was behind these distractions I would need a clear shot of those parts of the building that was behind these objects as well.

I’ll use Image One as the image I want to clean up.

Image One SOOC



After taking the initial image and noting where the immoveable objects were in the scene, I positioned myself just far enough to one side to see what was behind the objects and took a second shot before running off to catch up with my traveling companions.

I’ll use Image Two as my ‘filler’ image.

Image Two SOOC



Normal cloning works well on small objects that have a consistent pattern and colour, particularly if the adjacent area is high contrast and the object to be cloned is thin, like a telephone wire. The larger the object and the more diverse it is, the more difficult it is to avoid producing a repeating ‘clone pattern’ that is a dead giveaway and can look rather unsightly. That would be the issue with this image and the reason I took a second picture to work with.

Back in Photoshop with my two images, how can I proceed?

Well, what if I paint a hole in a layer mask and reveal a small part of the second image where I can see what is behind the distraction? There I go, asking those ‘what if’ questions again! OK, I’ll open both images in Photoshop as separate layers and blend the layers into one final image.

Uh-oh! The images don’t line up! Not only are they not aligned, but worse, they’re not the same size and because the of the different camera shooting angle for the second shot, the perspective doesn’t match either. Bummer!

Well, I could try to get CS5 to do an Edit/Auto-Align Layers but it really doesn’t help because of the perspective differences.

To overcome the perspective differences, let’s take just one object at a time, say the second floor window behind the street light and street signs, for starters.

We'll create a mask for the first image and using a soft black brush with the mask selected, paint out just the streetlight globe and top of the signs that are in front of the window.

Punch-Out Hole Image



Here we can see the layer mask to the right of the layer and the black spot is where I used a soft brush tool to create a ‘hole’ in the mask by painting on the full size image in the main display window.

Anything painted black on the mask will show what is behind it in the layer below. Because I have turned off the little ‘eyeball’ on the left of the layer below, there is nothing to see where I painted but the pattern used by Photoshop to indicate transparency in the image.

Now we can turn ON the visibility ‘eyeball’ of the lower layer and use the Move Tool to position the second image so that the wall and window, what is visible in the filler image, is positioned in the hole.

Fill the Hole Image



We’ll move the lower layer around until the window lines up in the hole. If the building's window is not vertically or horizontally aligned with the hole, we can skew, rotate, compress or stretch the lower image until it matches the original building window. We can change the second image as many times as we like and if we really go too far we can always make another copy of the original second image. Err-ahh, you did make a copy of the images you loaded into Photoshop before you started, didn’t you? Of course you did! I didn’t doubt it for a minute!

Align Building Window Image



We can now we see a complete window image where the hole was.

You might be wondering why the second image layer has a Brightness/Contrast Adjustment Layer just above it? I've found that we sometimes need to change the brightness and/or contrast of the second image ever so slightly to match the first image and this is an easy way to make that change! The third image layer is just the original copy of Image One for reference to see how I'm doing with the changes.

The last step is to preserve this change before we move on. To do that we make copy layer using ‘Stamp Visible’ (CTL+ALT+Shift+E) that merges the two layers we are using into one new layer. This new layer becomes our new Image One.

We can repeat the process with the rest of the lamp and sign poles and other objects in the scene until we have the final result.

What can we do about the objects like the car behind the building where the background is blocked in both images?

We can use the same technique except we may need to position another part of the image that is more appropriate in the hole. We can even position a part of a copy of the first image in the hole if it provides the background we are looking for.

What if neither approach will work, such as the right side of the steps which are blocked in both images?

Well, we can always fall back on using the normal Clone Tool if needed. In fact, there may be a number of situations where the Clone Tool can touch-up small spots that just don’t quite look right for some reason.

So here is the original Image One with just the keystone distortion corrected and cropped to match the final result.

Original Image One – Aligned to Match the Cloned Image



And this is Image One after the Punch-Out Clone process is complete.

Image One - Cloned



Have some fun! You can open these images in the LyteBox and use the arrows to switch back and forth between them for comparison. See how many changes you can find!

One last thing to do; crop a bit tighter to eliminate extraneous objects such as cars and power poles.

Image One – Cropped



Now that I think about it, maybe I should have left the antique streetlight in the image after all?

I’ll just use the Punch-Out Clone technique to merge the streetlight back in!

Image One – Final with Streetlight



Now that's more like what I had in mind when I first saw this building!

[Myk]

Frank, you continue to outdo yourself in this project.
Your attention to detail may only be surpassed by your ability to clearly articulate, and illustrate the steps you have taken to get from A to B.

[Dave Humphries]

Whoa Frank,

How are the rest of us supposed to keep up if you keep 'raising the bar' in your P52 posts?

Only kidding - thanks for another brilliantly described process, I'll have to give this a go one day when a suitable subject presents itself.

I do recall having shot for this process a couple of times, but never felt the captures were worth the effort when viewed on the PC, so I'll have to be more careful with my captures.

Cheers,

[Donald]

This week’s exercise is to see what we can do when we need to clone out something in an image but can’t get good results with just the Clone Tool alone.

Remarkable. Just remarkable. Knowledge and skill way beyond anything I possess.

[FrankMi]

Frank, you continue to outdo yourself in this project.
Your attention to detail may only be surpassed by your ability to clearly articulate, and illustrate the steps you have taken to get from A to B.

Thanks for the feedback, Mike. My hope is that folks will come to realize that fixing these kinds of issues in their images isn't all that difficult and that they'll lose some of their fear of playing with post processing.

One of the great hidden benefits of carefully examining your images in the PP activity is that it can really help you to improve your compositional ‘vision’ during the picture taking process.

[steve ward]

Thanks Frank, just did a quick glance through your thread....you've opened my eyes, thankyou.

[FrankMi]

How are the rest of us supposed to keep up if you keep 'raising the bar' in your P52 posts?

Oops! Sorry Dave. I thought we were only competing with our own past P52 posts!

In that regard, I keep running out of new things to try to learn from. Guess I'll need to get on the stick and start learning about artificial lighting or macros or any one of a multitude of other photographic skills that I know almost nothing about!

Oh well. I can't think of a better place to learn than right here at CiC!

[FrankMi]

Remarkable. Just remarkable. Knowledge and skill way beyond anything I possess.

Ah, but there's the rub. I may have learned something about layers but you are a Master of B&W images far beyond my ability to comprehend!

[FrankMi]

Thanks Frank, just did a quick glance through your thread....you've opened my eyes, thankyou.

I'm glad that you are enjoying the thread Steve. Let me know if you have any questions that I might be able to help you with.

[Wendy Stanford]

Thanks for the very comprehensive tutorial Frank, I am one of those that doen't have a very good grip on PP, perhaps as winter approaches here I will have more time to start experimenting with using layers etc, my list of things to try and experiment with is getting longer.

[FrankMi]

Thanks for the very comprehensive tutorial Frank, I am one of those that doen't have a very good grip on PP, perhaps as winter approaches here I will have more time to start experimenting with using layers etc, my list of things to try and experiment with is getting longer.

LOL! I certainly hope you do. There is so much more that most images have to offer that a tad bit of post processing can bring out for you!

[FrankMi]

This week I have been shooting images of local artwork and trying to display them in a pleasing view.

This picture was taken in uptown Charlotte, NC and is one of four bronze statues at Independence Square. The statues were a gift to the City of Charlotte by the Queens Table, a private philanthropic group. The statues were created by Washington sculptor Raymond Kaskey.

The statues are figures of Commerce, Industry and Transportation represent elements of the past which have contributed to Charlotte’s growth as a major commercial center. The three figures are all looking toward the fourth figure, Future.

As the focal point of the four-cornered composition, the future is represented by a mother and child. The mother figure emerges from a bower of branches, leaves and flowers of the dogwood, which is the state flower of North Carolina. Stretched between the branches is a hornets nest to make it more specific to Charlotte.

Each of the statues weighs about 5,000 pounds, is attached to a granite base, and is about 25 feet tall, including the base.

This is the SOOC image:



And this is the post-processing version:

[Donald]

A masterful piece of work on your part, Frank (again). And a beautiful sculpture as well.

Your cloning work is, it seems to me anyway, of the highest quality.

[Kris V]

This week I have been shooting images of local artwork and trying to display them in a pleasing view.

This picture was taken in uptown Charlotte, NC and is one of four bronze statues at Independence Square. The statues were a gift to the City of Charlotte by the Queens Table, a private philanthropic group. The statues were created by Washington sculptor Raymond Kaskey.

The statues are figures of Commerce, Industry and Transportation represent elements of the past which have contributed to Charlotte’s growth as a major commercial center. The three figures are all looking toward the fourth figure, Future.

As the focal point of the four-cornered composition, the future is represented by a mother and child. The mother figure emerges from a bower of branches, leaves and flowers of the dogwood, which is the state flower of North Carolina. Stretched between the branches is a hornets nest to make it more specific to Charlotte.

Each of the statues weighs about 5,000 pounds, is attached to a granite base, and is about 25 feet tall, including the base.

This is the SOOC image:



And this is the post-processing version:

Wish I had your patience (and cloning skill). Awesome!

[Teton Chuck]

Frank, very impressive. - Chuck

[Geoff F]

Yes, clearly a lot of work was involved there.

[FrankMi]

A masterful piece of work on your part, Frank (again). And a beautiful sculpture as well.

Your cloning work is, it seems to me anyway, of the highest quality.

Thank you Donald! This image is a tag-on to last week's work where I take several images and simply use a mask to reveal what wasn't hidden in another image so there was very little use of the Clone Tool per-se.

[FrankMi]

Thanks for the kind words Chuck.

Yes, clearly a lot of work was involved there.

Hi Geoff, in some respects it wasn't as difficult as the previous week's effort as only the background had to be modified. The subject, which is what we tend to concentrate on, didn't need any cloning work at all and the busy background is easier to manipulate.

[FrankMi]

Wish I had your patience (and cloning skill). Awesome!

Thanks Kris! At one point not very long ago I was where you are now and only because it is difficult for me to get out to shoot, I started playing with Layers and Masks.

It doesn't take very long at all to get the basics down and from there it's just playing with it from time to time as the images pull you in that direction. My hope is that you and others will be encouraged to play with it a bit, which is why I try to include some 'how to' in my Project 52 thread.

[FrankMi]

For this week I thought I’d pick up on an issue that others had posted concerning shooting through a fence.

There are a lot of different types of fence and the scenes behind them that we want to get. For wire or mesh fences, the finer the wire/mesh, the easier it is to make the fence disappear. In all cases, the closer you can get the lens to the fence, the further the subject is behind the fence, and the more shallow the DoF, the easier it is to get the fence out of focus to the point where it isn’t seen in the final image.

In this experiment I’m dealing with a more difficult to blur, chain link fence. As long as you don’t have a subject that moves (like wildlife), you can take multiple images with the fence in different locations verses the position of the camera and combine them into a fence-free image.

Up in the Tennessee Mountains there is a series of 28 dams built to control flooding, improve navagation, and provide electric power. This image is of the Chedah Dam on the Little Tennesee River.

I took four images of the left side of the scene and four images of the right side of the scene. For each of the four images I tried to center the same focus point and simply shift the camera left, right, up, and down to position the chain link fence in different locations to ensure that across all four images I could uncover the entire scene.

Image Left



Image Right



The first thing I did with each set the four images I shot was to align the subject. You may need to do this manually with the Move Tool as Auto-Align Layers will likely attempt to align on the fence rather than the subject! If you need to manually get the images in alignment you can use the Move Tool.

Select the first image and switch back and forth between the first and second images by alternately clicking on the Layer Visibility ‘Eyeball’ at the left of the first layer. Unless you used a tripod to take the pictures, you may find that you need to rotate the second image just a bit as well as moving it up/down and left/right to get them in alignment. You can align each of the image layers to the first one of the set using this technique.

Once the images are aligned, add a Layer Mask to the first image and with a soft black brush, paint out some of the wire fence. If the alignment is correct, the missing part of the image that was behind the fence should appear.

If you didn’t get an exact match on the exposure of the two images, you can add an Adjustment Layer for Brightness/Contrast to the second layer and adjust the sliders until you can’t see any transition difference between the original and what is showing of the second layer through the Mask in the first layer.

Once you are satisfied with the result, save the changes before you move on to another layer by using Stamp Visible (Ctl+Alt+Shift+E). This will create a new top layer so I usually move the original top layer and mask to the bottom of the layer stack in case I want to return to it later.

If you have finished merging the second layer with the first, I would move the second layer to the bottom of the stack as well. Repeat the process with the new top layer and the next layer down until none of the fence is visible in the final image and you’re done!

In this image, I had to do this process for both the left and right sides (for a total of 8 images). In my case, the one additional step is to merge the two final images into a panorama.