Composite Photos and Digital Photo Processing.
Panoramic Photo and Photo Processing in Adobe Photoshop.

Theory

This work has been done in 2006 with an amateur 5 Megapixel camera. This was the time when I had to compose separate photos together to get a photo of a better quality. In practice, however, it doesn't follow from the fact that I now have a wide-angle lens 10 Megapixel reflex camera that I stopped composing the frames to get the fine result. This doesn't mean to be a rush for extra megapixels; having got a 10 Megapixel camera, I happened to know that it's near to impossible to have a fine copy of a selected part of a frame printed in a format exceeding an A4 size. It's not that the copy will be absolutely ruined or something, it's rather a matter of having a surplus of quality for the sake of your personal confidence.

Next to that, the reason I switched to composed photos is that the optics and the matrix of the camera are also far from being perfect. Not to call the names, I can tell you that every lens has a drawback of degrading detail near to the edges of the frame. Still, sometimes you can avoid the defect by overlapping the edges of two adjacent frames. Personally, I'm not an avid fan of this technique, and I get rid of the defect by means of leaving an extra space at the edges while taking photos, having the excess edges removed after the composition of the adjacent frames.

Composed frames are useful to prevent the smearing of fine detail, the detail being of high importance, for instance, in landscape photography. One thing I'd like to mention - you cannot zoom in the picture substantially and 'slice' it with a help of a 17-40 or 17-50 mm lens, but this is quite enough to increase the detail in small or distant objects in the sky line.

Getting a little ahead of things, I'd like to note that using a 20-30 Megapixel camera with a normal lens, which works within the range of 120-180 °, is a telling argument for frames composition. As of now, there is a variety of panorama-building programs which facilitate this kind of work greatly.

So, leaving the discussion aside, - you'd better come and see how it works.

I use Adobe Photoshop CS3 to process photos. It's also possible to use earlier versions of the program as any changes here affect the interface only, the general principles of image processing remaining the same.



Let me introduce you to the way it works with a fresh example (10 Megapixel matrix, 17-50 mm lens). After loading the pictures from the camera to the PC, we get this:
While taking photos, you get a strip of 8-12 frames, and then take a photo of a general view - this will allow you to receive a rough overall image of what you will get after the composition.

For now, I will not give a detailed account on how to compose single frames into the whole (this goes beyond our topic). Instead, I will show you the meaning of all this steps. Let's take a single normal frame and select the corresponding areas on it and on the composed frame.

By simple calculations we may witness a remarkable effect: the size of the selected segment on the normal frame equals 3339x2218 pixels, while the corresponding segment on the composed frame is that of 5796x3856 pixels.


So, this is the ratio of the two segments:


Here comes the most important thing. Let's take a selection on the normal frame and scale it down to the size of a regular photo. No image sharpness or other image processing operations were applied. You can see the difference:

The sky and river buses

Let us return to the photo entitled "Before the rain". That day there were awfully windy upon the bridge, and it had already begun raining a little. Unfortunately, by the time the sun shone well enough, I was taking photos on the opposite side of the bridge. When I finally made it to the other side, there'd been no sun, and the resulting photos were a little unsatisfactory. I used to take photos with a 5 Megapixel camera without black-and-white mode. Only sepia mode was at my disposal. This fact subsequently led to the necessity of processing every single photo, because the combination of filters with the camera electronics yielded only medium-quality images, and I failed to receive a high-contrast image. Some time had passed before I came to know that it is simply impossible to attain to such an effect with this kind of camera in 'sepia' mode. From that time on I would use gradient filters in order not to let the sky be overexposed.

Here is the result from the camera:



I'd found the first sequence of frames depicting the sky in my archive, and then decided to add it to the composition, changing the whole way of work - this is what made me write this article. It's easy to notice that the sky in the first sequence is different from the frames in the second one, this being the result of a 15-20 minutes delay between shooting the first and the second sequences: my attention was drawn to a spectacular sight of three river buses going up the river at once. I shot a round of film with a film camera also, but the result was awful because the 37 mm M42 lens I had bought for 500 Rubles turned to be a disappointment. So, I shot the main course of the event with a digital camera. Two years later I decided to correct the imperfections by combining the two sequences.



I've chosen the first picture, which shows a river bus speeding away into the blue.

Now it is the time to connect everything together. I'd like to note that I will omit the detail of how to compose the frames (I consider writing a specialized article about this in the future), but here I'll try to tell you the basics. This article is dedicated for the most part to the work with an adjustments layers in Photoshop. This must be of particular interest for the beginners, as far as it is the cornerstone of photo processing; those who are already skillful can download this work in Photoshop CS3 *.psd format right off.

It's not a difficult task to connect the frames in this particular case - there're no substantial perspective distortions.

Getting started

I won't bug you with such basics as how to create a new layer, or change the brush size, or where to look for a specific button - if you don't know this, you'd better get something else to read. I'll try to put down a step-by-step guide to photo processing, which knowledge you can easily apply to your personal needs having just some basic skills in using Photoshop.

1. Open the central frame in Photoshop CS3.

2. Use a Crop Tool on the frame to add a free space around the image

3. Create a new layer and place it under the layer with the frame, painting it black after that with a Paint Bucket Tool

Open the rest of the frames in Photoshop. Let's now begin to transfer and connect them inside the first file.

1. Transfer a single frame.

2. Turn down the opacity of this layer to 50% with an 'Opacity' option.



3. Move the layer to find a similar place on the first frame.

4. Turn the opacity back to 100%



Add the rest of the frames the same way (Note: the place where the images would match might not be exactly on the edges of these images).

Having composed all the frames, we got the following as a result:



I intentionally compose all the frames at hand, despite the clear fact that some of them are redundant. It's necessary to leave a reserve, and not to have the frames butted as if they were wallpapers. Otherwise, you won't be able to make use of a smooth transition from frame to frame.

I also use 'Transform' and 'Rotate' options to get the frames joined correctly. It is a must to align the frames relative to each other and to the horizon line. I usually draw a line and align all the frames according to it.



Let's start to connect the frames by means of a mask. This is the easiest and most convenient way of doing it.

1. Pick the desired layer and activate the 'Quick Mask' mode A white area will appear on the 'Layers' palette, to the right of the frame.

2. The white area must be selected; if not, click the left button on it.

3. Click 'Brush Tool' button pick the black color, switch the opacity to 100%, and begin painting over the surface of the frame.

The picture will disappear there, where you have painted it with a brush. If this didn't happen, and if the surface of the frame became painted with black, green or any other color, then roll back the action and re-activate the white area to the right; then repeat the action. I'll try to show on the diagram how the mask works.



Let's proceed with our work. To make it more comfortable, I'll raise the image contrast by means of the adjustment layer. Adjustment layers are the most important features of image processing. They allow switching off the changes at any time.

1. Click on button (it stands to the right of the 'Quick Mask' option. Pick 'Brightness/Contrast' option.

2. Raise the contrast in the pop-up menu minding the overexposition of any areas, and then press OK.



This way makes possible introducing the changes to the image, which can be rolled back easily. The white area makes it possible to selectively apply the changes, and correct or add them at any time. Here's an example:



Let's return to the photo (I also added an adjustment layer to turn the photo into black and white) and remove extra frames.



Switch off extra frames; create masks for the frames needed, and begin to connect them.



It's necessary to note that I apply brightness correction while connecting the frames. There's no need to apply the correction at once, i.e. to the unconnected frames, because the edges of the frames are darker (due to the imperfection of the optics), and the frames will need additional correction after the connection anyway. Generally, I apply this correction to the frame at once, before the master layers have been used.

Image > Abjustments > Brightness/Contrast (Don't forget to select the frame, not its mask)

It's necessary to control the overexposition while applying such a correction. It would be a pity to find at the end of the work that some part of a cloud had lost its half-tones, and there will be nothing to help this, because of the presence of master layers.

I'll divide the connected frames with colors to make them more obvious (the same thing I've been doing with the first strip of the frames).



Now, let's join the results and apply brightness correction to them.



The first part of work is done - the photo looks homogeneous, although dark. Let's proceed to the main part of our work of image processing.

The first thing to do now is to choose the color gamma - as for me, I choose sepia here. The secret here is that of how the color space is being edited: I add one more layer over the layer with sepia toning. This additional layer replenishes the shadows with a blue hue, and the bright zones with the yellow one. Such a color division deepens the image on the photo. If you pay your attention to the live colors on a bright sunny day, you'll see the shadows dip into the color blue, and the well-lit areas filled with yellow hues. Having taken a group of experiments with the colors, I now often apply the blue-and-yellow combination even to color photos, because this livens up natural lighting. It should be noted that such a color edition of the photo is equivalent to the raise of contrast - that's why you should mind the overexposed areas.

Thus, let's remove the redundant black background, and add three master layers.



The result is shown on the following picture:



The sun had been gone down while I was taking the photos, the lower part of the image turned out to be way too dark. Let's correct the lower part, paying no attention to overexposed areas in the sky.

Remove the layer which was used to raise the contrast when the frames were being composed. In the future, every newly created layer should be place under these three layers, which set the color gamma of the image.

Now, let's create a 'Brightness/Contrast' master layer. The brightness value should be about +61, so that the lower part of the photo be lightened. Here comes the first one of my favorite tricks. Let's take the black brush in a mask mode and remove the obvious overexpositions manually. I call it - whether it is right or wrong, I have no idea :) - a manual enhancing of a dynamical diapason of a photo image. This is what is to come of it:



Now that the lower part of the image was lightened, I'm going to 'shut down' the sky - I use 'Multiply' and 'Overlay' options for this task. I've never known what the correct way of calling these layers is, but I can tell you that the first one is for darkening, and the second one is similar to the raise of contrast and brightness. The easiest way to demonstrate what 'to shut down' means is to present a diagram of three layers, which follow one another. Here it is:



Let's add a 'Brightness/Contrast' layer above these layers, and raise the contrast up to '+90' value. A mask should be used for the action to affect only the sky area. In our case, I keep the upper part of the photo image intact as long as it's dark enough as it is.



Now it's clear to see that a white stripe has appeared between the earth and the sky. I'll tell you why I have left it, and how it can be removed.

The white stripe should be left in order to connect the earth and the sky accurately. If you begin to cut out the buildings from the sky and correct every single one of them by itself, you probably won't succeed, but even if you do, it'll take many hours. The main difficulty is that of detaching the buildings from the sky thoroughly. Take my word for it - the black-and-white fringe persists, and it'll take hell of a work for you to remove it. It's much easier to bring it to one color.

Let's make the ground more contrast by raising the contrast level up to +30. Then take normal brush and pick an average color in the sky area. With 'Opacity' level set at 20%, paint over the whole image at the places where the buildings and the sky join.



Add brightness and contrast to the whole image minding the overexposed areas. By means of a mask neutralize the effect of the adjustment layer in the overexposed areas, should they appear.



The two pictures are about to be the same, but there is a feeling of the earth and the sky existing independently, with the photographed buildings having been 'glued' to the sky in the background. This is the effect of a bit too high contrast at the outlines of the buildings.



We should part the distant buildings and those in the foreground to make the photo look natural. Spotlight lighting of the buildings might be used to achieve this. Let's lighten the bridge with a spotlight. Lighten the necessary points by means of creating 'Brightness/Contrast' master layers and using a mask.



The same way can be used to lighten the ground and water, four more master layers being added.



The picture has become more homogeneous, what brightness and lighting are concerned, but the outlines of the buildings are still too sharp. Let's remove this effect via the already familiar way of coloring the image with a neutral color.



Now only a few details left to finish the image processing, and it is the addition of missing parts to the water and sky areas. I use the frames I’ve already used when I composed the image. Let’s add the same frame above the image, shifting it a little higher than the original, and embed it in the image via mask. I used a different frame to add necessary parts to the water area which I’ve found in my archive.



Once again I convert the picture into black and white mode in order to avoid the color influence the master layers above. This is going to raise the contrast a little; there where the contrast is sufficient, the effect of the adjustment layer can be neutralized by means of a mask.

The article must have come to an end at this point, but in the course of writing I had decided to add the frames from the film camera. The frames are not very good in quality, but they fit the screen size. Let's use a mask and 'Brightness/Contrast' master layers as a means of image correction..

It's time to add the river bus. I'm going to do this in parts in order not to become confused with the great quantity of layers.

Let's save the photo under the new name. Then, for the sake of ease, merge all the layers except for the three upper ones and the black background, and add the frame with the river bus. Using the 'Transform' option, find the junction points of the two frames.



When the junction points have been found, let's make the frame non-transparent (fill: 100%) and add a mask. Now, let's join the water areas of the two frames.



Now, the omission of a frame to the left has become obvious. To handle this, I'll make the task easier (perhaps, some say this way is not the best one). Let's merge all the layers in this file. Open the saved file without the river bus, and merge all the layers in it.

Do not forget to save a copy of the file without the river bus.

Now, we have two flat images with and without the river bus. Add the photo with the river bus over the photo without it.
Add a mask to the layer with the river bus.

There's nothing of a difficulty here - here is a diagram to show you the way of doing it. I usually call such files '*_last_fix.psd' and apply little corrections to them. In this case, I had to lighten a little the area with the buildings, and added the river bus.

РHere is the result of combining the two files, and some corrections applied to the photo.



The file structure will be more obvious, if the main layer with the photo is turned is turned off.



Now, let's add my favorite framing :), and… Here you are!

Attached files (Photoshop CS3)

v_book_001_cs3_in_front_step3.psd - Final cut of the river bus and the whole photo.
v_book_001_cs3_in_front_step2.psd - Adding the frame with the river bus.
v_book_001_cs3_in_front_step1.psd - Frames composition and the main part of the photo image processing.


Translated by Euristem

Tutorial index