13.5 Merging Poses

The nice thing about digital photography is that with a copious stash of memory cards you can take far more photos in the field that you’d be likely to do if you were shooting film.  This allows you to be pickier when choosing which of your images to share with the world.  If you’re a real perfectionist, however, you may find that for a particular subject you simply have no photos that satisfy all of your aesthetic preferences.  Sometimes, however, you’ll notice that one of your images satisfies some of your requirements, while other photos of the same bird satisfy other requirements.  In these cases you might consider merging these images, keeping the best parts of each image for inclusion in the final, composite image.  We’ve already been doing this in the preceding chapters: first as we merged the background of one image with the foreground of another image, and then as we merged portions of differently exposed photos in the case of HDR.  In this section we’ll consider yet another application: merging different poses of a bird into a single, frankenstein creature.

Fig. 13.5.1: Two images of the same bird in different poses.
The left image features some nice wing action, but the angle
of the head isn’t ideal.  The right image features a better profile
of the head, but lacks any wing action.  These photos can be
merged to combine the better aspects of the two poses.

The figure above shows two photos of the same bird.  The left photo features some nice wing action, though angle of the head isn’t ideal—if it was turned a tiny bit to the right, a better profile would be presented.  The photo on the right features the same bird presenting a slightly better profile, though unfortunately it’s facing left instead of right, and it’s slightly larger in frame.  We’ll show how to very easily overcome both of these issues, resulting in an image that can be merged into the first, thereby combining elements from both poses.
    The first thing to do is to mirror the second image (the right image in the figure above) so that the bird is facing to the right.  This is as simple as invoking Image > Rotate Canvas > Flip Canvas Horizontal.  Once this is done, we can begin thinking about how to resize the image so that the bird’s proportions match those in the left photo.  The first thing to do is to ensure that both images are being viewed at the same zoom level.  The two photos can then be positioned side-by-side on the screen, with the scroll bars adjusted so that the bird’s head is clearly visible within the window frame.   Now we can invoke the Image > Canvas Size tool and try different sizing proportions (see section 12.7) until we find one that equalizes the sizes of the two birds.   If you wish to be more exacting, you can use the Ruler tool to measure various dimensions of the bird (e.g., its beak length) in the two images.  Just keep in mind that some measurements can be deceptive, due to different angles (e.g., foreshortening of features due to camera perspective).
    The next step is to select a portion of one image to be merged into the other image.  For this example we’ve selected the bird’s head from the rightmost image in the above figure, and copied it via Cmd-C / Ctrl-C.  We then pasted this selection as a new layer in the other image (the leftmost photo in the above figure).  The figure below shows the Layers panel resulting from this operation.  The top layer contains only the bird’s head from the other photo.

Fig. 13.5.2: The Layers panel after copying
the bird’s head from one photo into another.
Though not shown here, a layer mask can
be used to trim the new head and aid in
blending it with the lower layer.

    What remains now is to adjust the new layer so that it blends in seamlessly with the lower layer.  First we’ll need to invoke the Move tool (press V) and use the mouse cursor to drag the new head into place in front of the old head in the lower layer.  In this case, some parts of the old head ended up still being visible behind the new head (not shown here).  There are several techniques that could be used to rectify this, including the use of the Clone tool and/or the use of layer masks.  We opted to use the Clone tool, since it’s generally very fast and convenient.  The figure below shows just the top layer after the Clone tool was applied to cover up parts of the old head showing through.  The source for the Clone tool was taken to be a region of the background layer close to the drawing (target) location. 

Fig. 13.5.3: The new layer after some touch-up with the Clone tool.
Pixels from nearby areas of the lower layer were cloned into this
layer to cover up parts of the old head that were showing through.
The Eraser tool was also used to trim away some of the darker
background pixels from around the new head.

    In addition to the Clone tool, the Eraser tool was used to trim away some unwanted pixels from the new head.  Recall from previous sections in this chapter that it’s sometimes useful to enlarge a selection before copying pixels, to ensure that nothing is inadvertently omitted.  In this case, some of the pixels around the new head were slightly too dark, and needed to be trimmed away using the Eraser tool; a layer mask could equally well have served this purpose.
    The figure below shows the result.  The top image is the original; the bottom is the composite image featuring the better head profile. 

Fig. 13.5.4: The result of merging the poses.  Top: one of the two original
images.  Bottom: the final
frankenstein image, after grafting the head from
another image onto this body.  Merging poses takes time and effort, and is
perhaps best left for the birds that are most special to you, or most rarely seen.

     Like replacing the background of an image, merging different poses of a bird can be time-consuming; I rarely ever do it, because of the time and effort involved.  These types of highly manipulative techniques are perhaps best left for those rare cases where you’ve got a photo of a very special bird (e.g., a species rarely seen in your area), but the photo isn’t ideal and you don’t have the option of simply taking more photos of the bird in hopes of getting the ideal image. 
    A related application is that of adding canvas when part of the bird (typically the wingtips) have been clipped.  Personally, I think some photographers put too much emphasis on the issue of clipped wingtips; when proportions are fixed by publication constraints, the logical (if not artistic) desire to retain all extremities often needs to take a back seat to the more important consideration of subject size and position in the frame, and it’s doubtful in many cases that general viewers would find the clipping of a bit of wingtip to be that terribly offensive, given current trends in artistic photography.  Nevertheless, it’s sometimes the case that you’ll wish you had captured more of the scene around the periphery of the photo.  The same technique described above can be used to clone in extremities such as clipped wings or feet from other photos of the same subject.  You’d first have to expand the image via Image > Canvas Size, to make room for the addition of grafted elements.  Once this is done, you can copy a selection from a second image and past it into the (enlarged) first image, or use the Clone tool with the cloning source set to the other image.  When choosing a new canvas size, it’s usually best to over-estimate how much additional
canvas you’ll need, since you can always crop the image later using the Crop tool. 
    Note that if you’re desperate to re-create a clipped wingtip or other extremity but don’t have any other photos of the same bird that you can use, you can sometimes get away with mirroring the symmetric feature from, e.g., the bird’s other wing or other foot.  This needn’t be as difficult as it may seem.  Though the source for the clone operation will likely be at the wrong angle and orientation, the rotation and mirroring tools available under Edit > Transform will often prove sufficient to achieve the desired effect.
    Another possible use of this technique is when you’re shooting with an extremely shallow depth-of-field (DOF), so that only part of the bird can be in-focus in any one image.  You might conceivably want to try merging the different in-focus regions from different shots, to produce a composite image having all or most of the bird in focus.  There are surely other applications that one can think of.  The wonderful thing about digital processing technologies is the great flexibility that they afford the creative mind.  And creativity is a good thing.