Perspective correction in Photoshop

There are situations when you just cant avoid perspective distortion. Like the one in the photo I will use in today’s guide. I was standing really close to the bridge where the fireworks were. Even with a crazy wide lens, I still would not be able to catch most of them, if I did not tilt my camera up.

So now I have perspective distortion, how to fix it? Lightroom, Photoshop and many other tools offer perspective correction, but I like to do it manually. The reason for that is, that most of them will keep the original aspect of the photo when correcting and so they crop off a lot of the photo. I prefer to keep as much of the photo as I can.

So let’s have a look how I do the perspective correction on one of my recent photos.

Perspective correction workflow

1. Let’s first look at the starting photo. You can easily see on the bridge, how distorted it is. The top is completely falling inwards. So it has be opened in Photoshop, and we can start working on it. Don’t forget to properly level the photo first, I it’s crooked you will not be able to correct the perspective properly.
2. First, I always duplicate the layer, and work on a copy. To duplicate, just right click on the layer and select duplicate. Then, choose the crop tool, and expand the photo, so there is space to work in. You don’t need to expand down, but you can.

3. Now select the layer you created, and select Edit/Free Transform (or just hit Ctrl+T). Now the image will have a border around it. We need to go into perspective edit, so right click anywhere within it, and choose perspective.
4. Now click and hold on the right top corner and start pulling it to the right. Hold down Shift so it stays in line with the original position. You will see how the opposite corner also moves out. Do this until the lines that should be vertical in the photo, are vertical. In this case, this would be the pillars of the bridge.

5. Now you will notice, that while the perspective is corrected, the whole image now looks like squished. We still have to correct that. So again, right click anywhere withing the image, and choose scale. Click on the middle top point and drag it up. If you have the latest version of Photoshop, hold down Shift, so it’s a non-uniform scale. In the older ones, you don’t have to.
6. Pull it up until the proportions look correct. You may have to switch back to perspective and correct the corners again, since the scaling can introduce some distortions again.

7. Once done, just crop the image to the proportions you want. You may leave some of the corners in, and then fix them later. I did a guide on that some time ago and can be found here.

And this is the final photo after all the corrections and edits.

The tilt-shift lens became my favorite lens almost immediately after I bought it, so today I will share with you how to use one. There are two main functions, the tilt and the shift, today we will look at the shift one.

What is shifting?

Shifting is moving the front of the lens up/down or left/right, without moving the camera. Like this you can completely change what the camera sees, without needing to move it all. Lets look at an example here. These four shots have been done from the same spot, with just shifting the lens.

As you can see, shifting a lens moves what you see by about 40% of the photo. There is a bit of distortion, but much less than if you moved the camera.

Correcting perspective distortion

The main use of shifting is to get rid of perspective distortion. You probably had a situation when you were trying to take a photo of something taller and you had to tilt you camera up to get it whole into the frame. The result of this is of course that all the vertical lines in the photo will start to fall towards the center, the more you tilt you camera.

Here you see an illustration of this. First tilting. As you can see, the field of view changes, to what you need, but a lot of perspective distortion is introduced with it.

Now lets looks how it is with the tilt-shift lens. When you shift, the field of view changes, but the camera stays leveled. Like this, no perspective distortion is introduced into the photo, and all verticals stay in a right angle with the horizon.

Since you are shifting instead of tilting, you can take photos of taller structures without any distortions, and also take shots while being much closer to the object.

Here is a photo with a normal lens tilted up, and with the tilt shift lens.


Another situation where perspective distortion can cause problems, is when you are taking panoramas or vertoramas. Each time you rotate or tilt the camera, you are introducing distortions, which makes the combining more difficult. with a tilt shift you can take all the shots without ever moving the camera. Lets looks fist at a panoramaic example:

Here I took two photos, one shifted left, one shifted right. These then perfectly fitted together when aligned. Or here is a vertorama example. Again taking the photos while shifting up to get the whole scene.

Moving the camera location

The last thing that it’s really useful for, is a way of moving the camera position, without moving the camera itself. What I mean by this is, that when you shift the lens, it looks like to moved the whole camera.

Let’s imagine a situation, where you stand on the edge of something and you just can’t move the camera further out. What to do now, when you want the view to be from there? You shit the lens, so getting the view from further out that you can move to. Same when you standing on top of something, and you would want the view to be even from higher up and similar.

This lens give you views that normally would not be possible at all.

That’s all about the shift function of this lens, next time I will take a look at the tilt function.

Multiple exposures

I almost always take multiple exposures for all my photos, using the AEB function of my camera. Even if I don’t have to. There are reasons to do it. But there are reasons also for not doing it. And today, I will try to take a look at reasons for doing so, and also for the ones against it. So here goes.

Why and when to take multiple exposures

High dynamic range
Of course the first and the biggest reason to do multiple exposures is to catch the whole dynamic range of a scene. Even the best camera can’t capture everything in one photo. You will need to blend images if you want to avoid overexposed or underexposed areas.
Better safe than sorry
Even if you think that the dynamic range is not too big and you will get everything from one photos, you can be mistaken. It’s better to have the extra exposures and don’t need them, that not having them and needing them.
Automatic backup
Anything can happen when you are taking photos. Even a tiny movement of the camera can destroy the image. When you ave multiple exposures, you easily can get around this just by using other exposures. If you took 3 and one is bad, you still have two that you can work with and so on.
Removing people from photos
Since you easily can under or over-expose a photo by one to two stops, you can use the extra exposures to remove moving objects or people from the final photos (more on this here). The same goes for other moving things. For instance, if there was wind and the leaves are blurred in the shot. You can take a faster exposure, like -2EV one, overexpose it to match the exposure you want, and then just blend it in. Since for that one the exposure time was shorter, the leaves will be more stable in it.

Why and when to take multiple exposures

Why and when not to take multiple exposures

Files take more space
The most obvious reason, you will get more photos. The file don’t take that much space at first, but over time it stacks up. I know this best from personal experience :)
Takes more time
It takes more time to take multiple exposures. Especially in the evening, when you photos get into the 10-30s range, you will get to the point when taking just one photo takes up to few minutes. If you light is changing fast, and you have only few minutes to take you photos, this can be a problem.
You are taking photos handheld
Some do multiple exposures when taking photos handheld. I don’t bother anymore. Even in the brightest of sunlight you are just not able to hold the camera steady enough. You will not be able to blend the shots perfectly. Rather than this, underexpose the photo a bit. You will always be able to better correct underexposed areas than overexposed ares afterwards.
Taking photos of moving subjects
Again, if you are taking photos of moving subjects, just take one shots. You will not be able to blend the shots anyway. A good example here is fireworks. They change so quickly, you can get only one photo. But remember, if you did not move you camera, you can still take a multiple exposures of the scene afterwards, and just blend the moving subject into them later.

Why and when to take multiple exposures

Adding watermark

On almost every one of my photo you can see a small text watermark in the bottom of the photo. Over the years I have added it many many times. But I did it very easily. I made myself a quick simple action in Photoshop, so it take me one click to add the watermark to the photo. And today I will show you how to create one for yourself.

How to create simple action to add watermark

As I mentioned, this will be a text based watermark, how I use. I will do a separate guide for a graphic on later, but it’s quite similar.

1. open any image, does not matter which one
2. choose Windows/Actions (or Alt+F9) to turn on the Actions window
3. either select a group to put your new action in, or create a new group by clicking the folder icon, name it and select it

Create simple action to add watermark
Create simple action to add watermark

4. create the new action, by clicking the new action button, name it and click record
5. choose the Text tool, select your font, font style and size, click on the image and write the text of your watermark
6. once done, select the move tool to deselect the text

Create simple action to add watermark
Create simple action to add watermark

7. select the whole image with Select/All (or Ctrl+A)
8. now align the text with the image, I will do with the bottom right corner
9. choose Layer/Align Layer to Selection/Bottom Edges to align with the bottom edge
10. choose Layer/Align Layer to Selection/Right Edges to align with the right edge
11. deselect the selection with Select/Deselect (or Ctrl+D)

Create simple action to add watermark
Create simple action to add watermark

12. now we move the text a bit away from the edge. Just hit the UP arrow key twice and the LEFT arrow key twice to nudge the text a bit (the move tool should be still selected)
13. click on stop to finish the action recording
14. you should have the following action recorded: make text layer, set selection, align current layer, align current layer, set selection and move current layer

Now on any image you can run this action you just created, to add the text you chosen to the bottom right of the image. If you want a different position, just choose different alignments in 9 and 10th step.

Also, don’t forget to save the action set you created. Select it, click on the 4 lines icon in the top right of the window and choose Save Actions…

Create simple action to add watermark

Correcting missing parts in photos

You don’t see it in finished photos, but quite often, especially with panoramas, you may have missing parts in a photo. It’s mostly in the corners. When two photos are merged into a panorama, the software has to distort them to match them, and the result is not a rectangular shape. Here is an example of that in a photo from Hallstatt:

Correcting missing parts in photos

You can see that the corners are missing and also part of the sky and water as a result of the panorama merge. I could crop it all off, but then I would loose a lot from it, as seen here:

Correcting missing parts in photos

And that’s just not acceptable. So instead, I would crop it to the desired compositon I want, and then fix the missing parts afterwards. So let’s say I crop it like this:

Correcting missing parts in photos

And lets fix the problems.

Using content aware

Content aware tool is great for areas of single color or areas with smaller repeating details. So for a sky, wall, water, forest in the distance and similar, it creates some great results. It’s really not great for very detailed objects, where the details change.

Here it works best on the sky, water, and the mountain in the distance. Just choose any selection tool, select the are a little bigger than you need and then choose Edit/Fill (Shift + Backspace) and choose Content Aware from the drop-down selection and confirm.

You usually get god results, but sometime one need to select the transition area again and repeat.

Correcting missing parts in photos
Correcting missing parts in photos

Using clone stamp tool

The clone stamp tool is the best on repeating patterns. You can also use it on single color areas and similar, but that’s usually simpler to do with content aware. Again, it’s not really that great for detailed object like bushes for instance, where you will never match the objects exactly.

To use it, just choose the clone stamp tool from the toolbar on the left, hold down alt and click on the source location that you want to copy, and then paint in the missing ares. I personally prefer to do this on a separate layer, as that makes it easier to go back or blend only parts of the correction.

Be careful when using it not to create visible repetition (if there should not be one).

Correcting missing parts in photos
Correcting missing parts in photos

Warping parts of the image

If nothing works, you can try warping the image. By this I mean selecting a part of the photo, next to the missing area, and stretching it around to fill the missing spot.

You have to be careful not to stretch any straight lines or objects, as that is very noticeable. This works best on organic things, like bushes, water, clouds and similar.

What you do, is to create a rectangular selection around the area. You go a bit wider than you need. Than you choose Edit/Transform/Warp (or hit Ctrl+T, then right click on the selection and choose Warp). Now choose the corner or the middle point (base on where the selection is) and drag it outside the photo.

Correcting missing parts in photos
Correcting missing parts in photos

Don’t over do it. Once a hole start’s appearing between the selection and the rest of the image, it’s too much. In this case, undo what you did, and do it again but in parts. Drag the point a bit. Confirm. Create new selection and repeat. Piece by piece you will fill in the spot.

Correcting missing parts in photos
Correcting missing parts in photos

Once you are done, sometimes you get a very noticeable distortion inside. In that case, reselects it, go into warp, choose a spot inside and drag it towards the distortion. What you want to do, is to get the distortion over a bigger area, so it’s not so visible.

Correcting missing parts in photos

How this photo ended

And that’s it for this guide. Hope it’s understandable and if not, feel free to ask :) I wanted to show you this photo finished with all the edits, but I never did this one, I did one taken right after it, so I will share that one here :)

Hallstatt, Austria
Page 1 of 2012341020... Last »
Subscribe to my newsletter and get a free Capturing fireworks ebook.