Bringing a Fairy Tale to Life: Creating a Convincing Composite Image

Alternate title: How to find legally usable images through Google image search.

In my mind, the process of creating a compelling composite image is as close to painting as a photographer can get: it allows you to create almost purely from imagination, not being constricted by physical locations as with traditional photography.

Last summer, I had a shoot with a few friends while we were staying at a cabin over the weekend. This was a low-key, just-for-fun shoot in some creepy woods near the cabin. Here’s another shot from that day, which I uploaded to Instagram:

For this shoot, the model was lit with a Nikon SB700 through a semi-transparent reflector on her right (this is a handy trick if you happen to find yourself without a softbox but with an assistant who can hold the flash and reflector). Here’s the image before any compositing:

Beauty and the beast - before

As you can see, this version looks quite a bit different from the finished composite. While editing photos from a shoot, I have the tendency to ask myself: would this work as a composite? Or, would adding some other elements make this a more interesting picture?

For this picture, the answer was clear (in my mind) right away. I needed a bear.

When you have an idea for an image and need a specific picture to make it work, there are several options. If you have a budget, you can search on traditional stock photo sites (such as iStockPhoto or Dreamstime); if not, you can check out some royalty-free options (such as Unsplash or Pexels). Since this was a personal project and my budget was precisely 0€, I skipped the paid options. And since I had a very specific image in mind (a bear standing on its hind legs photographed from the side), I couldn’t find an appropriate image in the limited catalogues of the royalty-free galleries. In this case, I delved into the third choice: Google image search. After a few tries, searching for “grizzly bear hind legs” came up with some promising choices (the highlighted photo was the one chosen).


Now, it wouldn’t be good to use a photo found from Google without permission (even if it is for a personal project). Thankfully, the photo I found had a watermark, which I used to track down the photographer, Kirsten Phillips, through a combination of Google and Facebook sleuthing. She responded through a Facebook message and was kind enough to send me the full-resolution file in exchange for giving her credit when the photo was published. Here is the full-resolution photo Kirsten sent me:


Now, the bear’s posture was perfect for the picture, but the lighting and colors didn’t quite match. Nothing a bit of Photoshop can’t fix. First, I flipped the bear horizontally, then used a layer mask to get rid of the background and cleaned up the grass strands on the bear’s legs. I also changed the angle of the bear’s head so that it would be looking down at a bit more of an angle. Here is the bear added to the original image:

Beauty and the Beast before 2

I still had to get rid of my handy assistants and match the bear to the background. Instead of attempting to clone them away, I copied the area to the left of the model, mirrored it, and placed it behind the bear. After a bit of touch up, the image looked like this:

Beauty and the Beast before 3

The only thing left was to color- and light-match the bear. After numerous adjustment layers (mainly curves and solid color overlays), I had a finished image:


Overall, the process took several (aka way too many) hours. But I’m happy with how the final, out-of-the-ordinary image turned out. In case you want a bit more of a taste, here is a short time-lapse style video showing all the different layers involved:

Model: Ruut Uusitalo.

Comments (2)