The compositing can be done in 16 bits per channel (48-bit color) for most applications and operations, but 8 bits per channel (24-bit color) reduces the file size and allows more effects within Photoshop.
In this example, we start with our positioning image. All the images were made from the same location with the same image alignment on the sensor. This means that the files will align in Photoshop. There, you have two distinct choices about how you'll work with the layers. One is an additive process, and the other is subtractive.
The first option is to select a portion of each layer and stack one on top of the other. Each successive layer works toward the final image. However, if you take this approach and you make the selections prior to layering, the corners of the frame should be maintained to assure proper alignment. The second option is to use each frame in its entirety as a layer. When you go this route, you'll need to erase areas after each new layer is in place. While this works just fine and is very useful for some shots, it's usually more time consuming and potentially less satisfactory than the additive process with small corners selected. If the selection is made without including the corners, then moving and nudging the images into alignment is often required. If any nudging needs to be done, be sure to view the image at the pixel level. In either case, no effect should be used on any layer prior to alignment with the master file.
Glenn Rand is a photographer, a writer and an educator. You can find his books at www.amazon.com.