Digital Darkroom

We use a flat-bed UMAX PowerLook III scanner for scanning reflective material. It has a maximum scan area of about 8.5x11.7". While that size handles a lot of what we need to scan, every now and then a client will bring in a much larger old picture that they want restored.

Recently, a client brought in a 16x20" picture and wanted it restored. Our problem was how to perform the scan. Here's how we did it.

First, we set the scanner in the middle of the table so we could work all around it and have plenty of room. Next, we removed the lid of the scanner. The lid is only necessary to scanning transparent material or for holding things flat on the scanner surface. Removing the lid would then give us unrestricted access to the flat-bed surface of the machine.

After carefully removing the picture from the frame and determining that it was in good enough condition to be gently handled, we proceeded to make a total of six different scans: top left, top middle, top right, bottom left, bottom middle, and bottom right. Next, we would have to "stitch together" the pieces. And, that is what this article is all about.

Higher Res Than Required
When we scan things we prefer to capture a little more raw data than what we will need, and to then use Photoshop to "throw away" the un-needed data by interpolating down to the actual size that we need. We think that this technique assures us that we will have plenty of raw data. In this case, the original scans were made at 100 percent and 400ppi. So, the first thing we had to do was interpolate each of the scans down to 300ppi, which is what we will use when we finally send the finished picture to the ink jet printer.

Now, it was time to start the stitch-together process. We would need a 16x20" finished picture. So, we created a blank canvas that was a little larger than what we needed to give us plenty of room to work. We created a blank canvas that was 18x22"x300ppi, and copied the top-left scan onto it.

Next, we copied the top-middle scan onto the blank canvas.

In order to align the two images, one slightly overlapping the other, it is necessary to magnify the blank canvas so you can more easily see what you are doing. Then, go to the Layers Palette and adjust the opacity of the top layer to about 50 percent. This will allow you to see through the top layer and aid in the alignment process. In this case, it was easy to use the sharp edges of the tear just above the right eye to perform the alignment.

In order to move the top layer in very tiny increments while performing the alignment, you can use the arrow keys on the keyboard as long as you have selected the Photoshop Move Tool first. The arrow keys will move the top layer in one-pixel increments.

Once the alignment has been performed, you can go back to the Layers Palette and reset the opacity back to 100 percent. Then, in the Layers Palette, perform a Merge Down to combine the top layer with the layer under it.

Next, we copied the top-right scan onto the blank canvas.

The top-right scan is then made 50 percent opaque in the Layers Palette and dragged into position for final alignment. We used the sharp edges of the tear over the subject's left eye to judge the quality of the alignment. Here, the image before the alignment was corrected and after the alignment is shown. Again, we used the arrow keys to perform the critical movements of the final alignment.

After this second alignment was performed, we again restored the opacity to 100 percent, and merged down the layer.

The remaining three scans were aligned and merged together in the same way. When we had completed the picture, we used the cropping tool to crop the picture to exactly 16x20".