I created this dog / cat test image to experiment a bit on my Instagram account. It’s a gain map which is designed to make it very easy to tell if you are viewing with HDR support. When you view this image and have HDR support, you will see a dog. But if you have an SDR display (or almost no HDR headroom at all), you will see a cat.
This test image is taking advantage of the fact that a “gain map” offers two different views of an image, based on the level of HDR headroom you have. The intent is that you would encode two variations of the same image: a basic SDR which is safe for viewing on any display, and an enhanced HDR which looks much better on displays which supports both HDR and gain maps.
But I’ve hacked the gain map format so that it is rendering two completely different images. If you lack HDR support, you see the base image (the SDR image of a cat). If you have at least 0.5 stops of HDR headroom, then the gain map is applied to generate the alternative version of the image. The intention would normally be to encode HDR content, but I’m not really using HDR pixel values in the dog image. So the effect is that you see one of two regular images, depending on your display’s capabilities.
I have posted a higher-resolution version of this image on Instagram (IG), where the experience is also a bit different due to how IG uses it.