julian-fietkau.de

Today's again something else entirely. I want to show you an idea with which I experimented some time ago, that has to do with the perception of partially obscured images. The idea comes from a video I've seen long time ago. There, someone creates this effect with black silhouettes. At this point I am brave and try it directly with a color image, but a simple. Here is a spiral:

For this experiment, we need an easily detectable, not too detailed picture. But this spiral is well suited. However we need a movement in the picture, so we took the spiral also rotate around its origin:

This animation consists of eight frames to which we return below.

Now the human brain over millions of years of evolution has developed the ability to recognize partially hidden forms and objects. How this works is not exactly clarified. As a reason, one might conjecture that the brain not to this internal support needs the full input of the eye and essential recognition events as well work in the dark with only some existing image information. The effect can also produce a specific form of pretending that there is not so, as in the famous triangle illusion. Even if we cover 7/8 of the (stationary) spiral, we continue to believe to see a spiral.

Our brain "sees" the subject continues, connecting the seemingly hidden lines. This works, although the information content of the image (i.e. the number of contributing to the image pixels) was reduced by 87.5%!

Of course, the works not with any motive - in fact, in most of the works of all kinds of pictures are not good enough to be presentable. The "compression" comes from the fact that the viewer fill in the gaps in the picture with information needs, which he has previously. (Anyone know how to look a spiral, and does so unconsciously.) Case of unknown images or those with many small details that fails principle.

However, we can do with our spiral is this: we can put the mask on a blank white image and apply only through the thin gaps in the mask a few strips of our spiral image. Then we can move the mask after a gap width on the left or right, so that the previously placed strip just hidden. Then we can apply through the gaps in the mask to the (now white again) strip another picture. In this example we are going, where each matching strip insert the next frame of the animation. The principle can we blithely continue and add six more sub-images, until we meet again (moved to a "beam" of the mask) arrived at the original image are.

When we're done, you see the image under the mask almost exactly like this:

And now the kicker: If we now move the mask slowly to the left or right, it looks like it would turn the spiral, even though only a single rigid image below the mask.

There are no tricks or ambiguity - the only thing moving in the animation above, is the black bars. The underlying image, which is assembled from the individual images of spiral animation, thereby completely rigid.

Who will give me the basis of the pre-animation still does not believe the round may also like to play itself out with. Here is a little HTML/CSS code that specifies the makeup of the composite image, so that they can be moved by scroll bar.

What can you do with this idea? I have not had a breakthrough idea for applications. It works fine but after all, even when printed on paper, unlike GIF animations. Perhaps there are such subjects as postcards at some point... As I said, it's a nice little gimmick, I think.