Some photographs — like this one of “Inside Out,” 2006, by Ueno Masao (b. 1949; Kanto region: active in Kamogawa, Chiba Prefecture; madake bamboo, rattan, and gold leaf, Asian Art Museum, gift of Ueno Masao and Tai Gallery, 2006.41, photograph by Kaz Tsuruta) — set an object against a background that subtly blends from darker to lighter tones. (The photo appears in the book Masters of Bamboo.)

In page layout you might want to fit this vertical image into a more horizontal space without losing any of the image. For a long time I struggled with the best way to accomplish this in Photoshop. If you did something like sample a tone near the edges and fill the background with that color you would get an unacceptable result like this one (copying and pasting bits of the background give results that are little if any better):

Currently I use the following simple technique. (I can’t provide many screen shots because for some reason they don’t show selections and cursors very well.)

1. First, I increase the canvas to the desired size.
2. I select a rectangle of background right up to the outside edge of the original image.
3. Next, I paste that rectangle into a new layer.
4. Then I flip the rectangle horizontally. This puts all the pixels on the edge of the selected rectangle adjacent to identical pixels in the original image.
5. With the rectangle still active, I select “free transform” and stretch it to the edge of the canvas.

This image shows the original and the two rectangles as Photoshop layers:

Here you can (sort of) see one of the rectangles that has been flipped, now ready to be stretched. I’ve circled the selection markers where you do the stretching. I unchecked the original image in the layers palette in order to show this better. Of course it fits in the blank space in the middle of this image.

This technique results in something like the image (here a little funky because of saving down for the web) below :

.

More posts related to Photoshop

• Photoshop tutorial: how to extend a graduated background
• Correcting color cast with Photoshop
• Low vibrancy
• Always check your work
• Photo Friday
• Lens blur
• Working with video in Photoshop
• Tilt-shifting the Pioneer Monument
• My Photoshop default workflow

.

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

Photoshop tutorial: how to extend a graduated background

We’ll start with this image of the multimedia center at the main branch of the San Francisco public library.

The first thing we need to do is to duplicate the image.

What we’re going to do with the duplicate image is average all of its color. We will use the resulting value to determine our midtone gray point. If you have a recent version of Photoshop you can use the “average blur” filter to do this. I have an older version, which doesn’t include that filter. What to do?

The answer is to resize the duplicate image to a single pixel. This will give the average value. (You will need to enlarge the view of the one-pixel image a little so you can use an eye dropper on it, as we’ll see.)

Now we’ll open a threshold adjustment layer.

With this we find the black point and white point. We do this by moving the slider all the way to the left and then backing off a little until black areas appear. One of these will be our black point. We mark it by shift-clicking (this marks the spot with the color sampler). Then we move the slider to the right, back off, and mark our white spot.

Once we’ve marked our black and white points the threshold adjustment layer is no longer needed, so we can cancel out of it and open a levels adjustment layer. The levels panel has three eyedroppers that when clicked on the image will set the black, gray, and white points. With the eyedroppers on the left and the right we set the black and white points, using the spots we identified previously with our threshold adjustment. Now we want to set the gray point, and this is where the duplicate image that we used to find the photo’s average color comes in.

For most images this technique really works well. Of course, normally you will still do your usual adjustments and sharpening if you’re really going for the best result. (I haven’t done that here.) The original image is on the left and the image with the color cast removed on the right.

.

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

Correcting color cast with Photoshop

Well, fine, but why go to that trouble? All you need to do to achieve this effect is to duplicate the image, change it to grayscale with whatever technique you prefer, and then apply the image in whatever percentage gives the best result.

Here’s the before and after from the tutorial.

Here’s the result using my simpler technique. I just changed the mode on the duplicate to grayscale but of course there are other ways to desaturate, which might be better. I could have matched the above image exactly but I couldn’t stand how murky and low contrast the original image was, so I couldn’t resist fixing it using my usual Photoshop adjustments technique.

Why complicate matters?

.

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

Low vibrancy

.

via gigglesugar (whatever that is)

.

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

Always check your work

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

Photo Friday

So what’s the difference? In theory, there’s a fundamental difference, which can clearly be seen in the following chart based on a Russell Brown presentation reported on Computer Darkroom.

You can see that Gaussian blur, in the middle column, erodes edges and grays white areas, while lens blur, in the right column, applies a geometric (somewhat hexagonal) effect similar to that of a camera lens without dulling highlights.

In the real world, however, the results may not be as noticeable. Here is an image with a Gaussian blur (top) and a lens blur (bottom). Because they use different metrics, it was difficult to get the amount of blur identical; nonetheless, the results look fairly similar — maybe the lens blur is a little crisper. That could partially be because I didn’t blur it quite as much, but I think its range of tones is also a little different.

I’m getting to like the lens blur effect. It feels like it gives you a little better control. Right now I’m working on a book about the Chinese artist Zhan Wang. For this book I’ve applied some selective lens blurring to some of the images. Following is an example — original photo on top and modified photo beneath. For this image I applied a quick mask to select the figures and foreground rocks, then lens blurred the rest. (Of course I also did my usual adjustments.) Do you like this effect?

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

Lens blur

• In part one he demonstrates combining two video sequences using a layer mask
• In part two he shows how to blend objects over time
• In part three he explores spot colorization and looks further at video blending and painting over time

These techniques require CS3 — guess I’m going to have to upgrade my software.

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

Working with video in Photoshop

Wasn’t that fun? Many people see this effect as creating the illusion of a miniaturized landscape. You can do it, or something very like it, with an extremely expensive camera called a “tilt-shift” (whence the name of the effect). Or you approximate the effect in five or ten minutes of Photoshop.

The technique is described, with a few variations, in many places around the web. You can check it out on your search engine. Or, you can just read on.

First, you have to go into quick mask mode.

Then, with your colors selected as black/white (you might want to invert them so that white is foreground and black background) you select the reflected gradient tool (the fourth gradient over).

Now just stretch a line from the focus point in whatever direction you like (experiment). I find a short stretch is better than a long one. Your image will look something like this:

Switch back out of quick mask mode and apply a blur (you might have to invert your selection). You will read that you should use a lens blur, but a Gaussian blur works just about as well (I don’t have lens blur as an option on my home version of Photoshop, where I still haven’t upgraded from 7.0).

Then I just adjust the levels, curves, and saturation, and apply a high-pass sharpening, as I’ve explained before.

Let’s recap. Some tutorials are frustrating because they leave out some small but critical step along the way. So there can be a value to detailed instructions. On the hand, it can be difficult remembering eighteen or nineteen discreet steps to an operation. It’s worth understanding its basic principles. What we’re doing here is applying a gradient mask and blurring the result. Remember that and you can play around until you get the effect you want.

Here’s a before and after:

.

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

Tilt-shifting the Pioneer Monument

I’ve chosen an image more or less at random (except that it is one that I like, from this photoset). My vantage point was looking down at a river from overhead, with colorful leaves on the right. For the purpose of this demonstration the image has been resized to fit this space (435 pixels wide).

The first thing I do is to open an action I’ve saved under the name “open adjustments.” This opens three adjustment layers: levels, curves, and hue/saturation in that order, which is the order I make the adjustments.

First I look at levels. If they look well balanced I might leave them alone. Often they are weighted to either darks or lights, and I slide the midtone triangle to get a better balance. That often makes the image look worse but it puts it in position for the next adjustment, curves. Usually I find a midpoint that looks good and then generally make an ess-shaped curve in order to get a good range of darks and lights. Finally, I adjust hue/saturation. With my current camera this usually means just increasing the saturation a little bit.

Next I open an action I’ve saved under the name “hi-pass sharpen.” This sharpens the image using the duplicate layer – invert – blur – overlay – adjust transparency workflow that I have described previously. I don’t like oversharpening, so my default transparency is a modest 40 percent. It’s important to remember to select the background layer first or you will just be sharpening your hue/saturation adjustment. One nice thing about this way of sharpening is that it is size independent, so I can resize my image and do a save for web to reduce the file size without having to resharpen.

The entire process is done with adjustment layers and is completely nondestructive — no changes are made to the original image. Below the image as it came from the camera is on the left and the adjusted image on the right.

Post from Right Reading, Tom Christensen's guide to print and electronic book publishing.
Follow me on twitter.

My Photoshop default workflow