Multiplicity: The five of me
Jul 1st, 2007 by Intermanaut
Following on from my “Multiplicity: Pouring yourself a beer” tutorial that used the eraser tool to merge the images, this slightly more advanced tutorial explains how to use layer masks to achieve the result. Layer masks are better suited to this because you can selectively “undo” changes to the layer long after you’ve made other changes.
This example uses Photoshop CS2, but the method is similar for CS3 and Corel’s PaintShop Pro X.
Start with a new image with a white background. I leave the initial background as it is so that I can change the order of the important layers if necessary. Open your base image and paste it in as a new layer. The start image in this example is the scene with no subjects, tripods, cameras, etc. My start image was taken at the 10mm end of a Sigma 10-20mm lens, so there’s some distortion at the edges.
Now load the next image in to its own layer. Be careful of the order in which you stack the images to ensure that you don’t end up with fiddle masking, body parts incorrectly stacked or passing through body parts in other images, and so on. You need to consider this while composing your shots as well, though.
Once all of your images are loaded, turn off all but the clear background, the base image, and your first “subject” image. Select the first subject layer and add a mask layer. With the mask layer selected (the white rectangle next to the layer’s thumbnail) select the “Brush” tool and set the foreground colour to black, and the background colour to white.
Start painting over the parts of the layer that you don’t want in the finished image. In this example I’ve painted out part of the bottom-right corner because there was a colour cast or shadow created by one of the other images. The area that will be removed is shown in the layer mask’s thumbnail as the black area.
When you’re happy with the masking on this layer, make the second subject image visible. Create a layer mask and repeat the process. In this example I’ve masked out (that is, painted over) everything but the Teapot Beer Server (”I’m a little teapot, short and stout…”). Repeat this process until all of your images are visible, and masked correctly.
The glass in this image is stuck to a tripod using Sellotape’s “StickyFixers” - other doubled-side sticky foam pads are available. When positioning the tripod, be careful to place it so that you don’t need to be a gymnast or an acrobat to get around it, and so it will be relatively easy to remove from the scene later. This tripod is quite sturdy, which was important for this scene because I’m pouring the beer in “real-time” so that it looks authentic. Don’t forget to make sure that it’s in arm’s reach, too.
The final example image here shows the state of each layer once they’d all been masked. You can see that I decided to mask out each layer so that only the subject (me) was left. This is where the clean base layer comes in. Each of the subject layers has a slightly different colour tone which is caused by reflected and bounced light. The position of the subject alters that tone, so the base layer gives consistent colour and tone throughout the image.
A note about layer masks
If you make a mistake with your layer mask, simply swap the background and foreground colours around and start painting. Painting in white over the mistake restores the layer. This is what makes layer masks so much better for this purpose than the eraser tool I used in my previous tutorial.
The finished article
You can see the result in my Flickr photostream.