One powerful advantage of the modern digital darkroom is the ability to expand the camera’s capability by blending multiple photographs into a single image. By blending photographs, you can increase the field of view and/or resolution (stitching), remove unwanted elements, reduce noise, increase the depth of field, and increase dynamic range. Whether you blend using an automated program or by hand, no one method will work best in all situations. Automated blending programs can be simple to use, time efficient, and produce optimal results with some images. Unfortunately though, there are many images where the results are less than optimal. Although blending by hand can be more labor and skill intensive, many images look more natural with this method.
One commonly recommended method of blending by hand involves a broad transition between photos. An example of this method would be editing a layer mask in Photoshop, using the Gradient Tool or by painting with a soft-edged brush. An advantage of this method is that the broad transition allows for a gradual change between photos. This is essential when the photos are different in the area of the blend, such as when blending two images processed differently, one for shadows and one for highlights, from the same raw file.
I use broad transition blends, and find they can work well in smooth areas such as sky, clouds, and water. Unfortunately, on some images this method results in unnatural transitions, loss of detail, or ghosting. Unnatural transitions occur when the difference in dynamic range between the two photos is too strong, resulting in an appearance that appears too dark, too light, or both in the area of the blend. Loss of detail occurs when the photos are even slightly different in luminosity detail; if what is sharp in the photos is not exactly the same, the blended area will be less sharp than either photo. Ghosting occurs when an object in the transition area moves between photos, resulting in a “double-vision” appearance with two faded versions of the object showing in the final image.
Another method of blending is to create a narrow transition between photos. In many images, if the photos are properly prepared beforehand, this method results in a more natural appearing blend. While this method can be used to blend for all the reasons noted above, in this article I will describe my method for blending photographs with a narrow transition using two examples. The first example will create an image with increased depth of field, and the second an image with increased dynamic range. This is an intermediate level article, and assumes a working knowledge of Levels, Curves, Masks, and Tools in Adobe Photoshop. These instructions and accompanying illustrations use my current version of Photoshop, CS3 for Windows. If you use a different version the specific steps may be different, but the general workflow should be similar.
Before beginning, it is important to emphasize that when blending using a narrow transition the goal is to make the areas to be blended as similar as possible in both photos prior to blending. This begins when taking the photos. You should use a solid tripod to minimize movement. Use a similar white balance setting between photos, set either in your camera or in your raw converter. Unless you are blending for depth of field, use the same focus for each photo. Unless you are blending for dynamic range, use the same exposure for each photo.
Okay, let’s get started. In this first example we’ll use two photos that are similar in all respects except for depth of field, one photo with the foreground in focus and a second with the background in focus, with a zone of overlap where both photos are in focus.
1. Create a new Photoshop document with each photo as a separate layer.
2. Align the photos.
3. Note the area where both photos are in focus. In step number four, it is important to know, yet difficult to see, where the photos are in focus. If you are unsure, there are several methods to note the in focus areas.
4. Set the top visible layer to Difference mode. In difference mode, areas in the current layer that are different from the layer below are lighter, while areas that are similar are darker. (Areas that are completely different are white and areas that are exactly the same are solid black.) This allows you to see and avoid areas that are not alike. For example, if a breeze was moving a flower between the two photos, it is very easy to see and, in the next step, avoid this difference.
5. Create a selection.
6. Create a mask from the selection.
7. Blur the mask. Using a very slight blur can help smooth slight differences between the photos along the edge of the mask.
8. As a final step, click the layer’s visibility on and off to be sure the transition is imperceptible. If you can see the transition, you may need to review and repeat the above steps and make changes as necessary. If the transition is visible because one photo is lighter or darker, we’ll figure out how to fix that in the next example.
In our second example we will combine photos similar in all respects except dynamic range, combining a photo exposed correctly for highlights with a second photo correctly exposed for shadows. This is a more challenging situation, because as taken there is no area of overlap; we must first adjust the photos so that there is an area of overlap while protecting the detail we are trying to gain.
1. Create a new Photoshop Document with the photos as separate layers, just as you did in the first step in the above example.
2. Determine the zone of overlap. Where to blend the photos varies from image to image, and takes some experience to determine. As a general rule of thumb, I try to find an area that is not too extreme in either photo. Avoid areas that approach 0 or 255 in any of the channels.
3. Place Color Samplers in the zone of overlap.
4. Approximate the luminosity in the zone of overlap using a Levels Layer. Our goal is to adjust the luminosity of the upper layer so that it matches the luminosity of the lower level. In this step we’ll approximate it, and then refine it in step number five. Again, with rare exceptions, the green value will work best for this step.
5. Fine-tune using curves layers.
6. Now that the photos have a similar zone of overlap, create a selection and mask using the difference mode as described in steps four through eight above.
These two examples illustrate blending with narrow transitions. As you can see, this method relies on both images being similar at the blended area. In that regard, there are several situations that are important to discuss.
In the depth of field illustration, you should note that the method requires an area of overlapping focus to be successful. This works best when there is a steady transition from near to far, similar to the type of scene where a tilt-shift lens works well. Sometimes this is impossible when you have a sudden transition from near to far. An example of a close object superimposed onto a far object would be a flower in the foreground in front of, and in a different plane of focus than, a distant mountain directly behind. In this situation, there is no area where both photos are in focus because the out of focus margin of the flower in one photo is wider than the sharply focused margin of the flower in the second photo. (If there is no movement between the photos, automated blending software such as TuFuse or Helicon can sometimes combine a series of photos variably focused from near to far to produce an acceptable blend.)
In the dynamic range illustration I sometimes obtain better results if I use a levels layer for each photo, one to brighten the photo optimized for the highlights and one to darken the photo optimized for the shadows, so that the dynamic range of the two meets in the middle. In other words, I adjust both photos a smaller amount as opposed to adjusting only one photo a large amount.
In an image with extreme variation in dynamic range, you may need more than two photos blended to keep detail in both the shadows and highlights. For example, a photo exposed for the shadows, a second for the mid-tones, and a third for the highlights. In this situation I would use Curves and Levels layers to darken the photo optimized for the shadows, and another set to brighten the photo exposed for the highlights, so that both can be seamlessly blended with the mid-tone photo. I typically have good success if my photos are bracketed 1 ½ to 2 stops apart, and then blend enough photos to both avoid blowing the highlights and to minimize noise in the shadows.
I often find that I can combine depth of field blending with dynamic range blending. As an illustration, if you have a bright sky in the distance and a dark foreground you may be able to take three photos that when combined give you good detail throughout, one optimized for shadows with the foreground in focus, a second optimized for shadows with the background in focus, and a third optimized for highlights with the background in focus.
Feel free to experiment with other ways to create a selection. Do you have a scene where you just need to blend some detail back into small areas of blown highlights and nothing moved in the scene between your bracketed photos? Save some time by selecting the highlights (Select:Color Range… and choose Highlights from the drop-down menu) and then expand the selection (Select:Modify:Expand…) a little. You can then use the difference mode to see if you need to add or subtract from the selection before proceeding.
And finally, you may find some photos need an additional Levels or Curves layer restricted by a mask to a certain portion of the blend. An example of this would be if between photos a cloud made a portion of one photo darker than it was in a prior photo.
If you are interested in blending or have been blending with broad transitions, hopefully these two examples will give you ideas to begin experimenting with narrow transition blends. Experiment with this technique to fit your own needs. If you are trying to photograph a waterfall and you just can’t get a shot without a water drop on the lens, take two and blend the drop free areas together. If you are trying to photograph an iconic location but just can’t seem to keep people from wandering through the scene, take two or more photos that when blended together leave the scene people free.
Since no one method of blending works well in all situations, automated, broad transition hand blending, and narrow transition hand blending all have a place in the photographer’s list of tools. With experience you will find situations where you will favor one over the other, or you may even use multiple types of blending in the same image. Whether you have experimented with blending before or not, hopefully you will find narrow transition blending to be one more potent tool you can use to improve your images.
Comments on NPN digital nature photography articles? Send them to the editor. NPN members may also log in and leave their comments below.
John Williams is a small-animal veterinarian practicing in Battle Ground, Washington, USA. Right about the time he moved to the Pacific Northwest at the close of 2002, he became interested in expanding his bird watching hobby by digiscoping birds. Photographing birds quickly led to photographing the surrounding scenery, leading to a passion for landscape photography. Whenever possible he continues to escape, with his (occasionally) trusty dog Jep, to the beauty of nature, hoping to record those magical moments of light that capture him far more then he captures them.
Comment posted by Ed Cooley on 02/01/10 at 7:00 pm
Comment posted by Jim Erhardt on 02/03/10 at 10:02 am
Comment posted by John Williams on 02/03/10 at 11:14 am
Comment posted by John Williams on 02/08/10 at 6:59 pm
Comment posted by Bill Brennan on 02/25/10 at 9:35 pm