I agree. I was interested enough to download the gif, and get an image editor to seperate it into individual images. They all look like perfectly normal celebrity shots to me. In fact I think they got bored trawling the web for pictures :- most of the images are used as both left and right components, and some of them are used more than once on the same side.
And it’s not related to stereoscopic vision :- I’m blind in one eye, and see the distortions.
Individual frames here, if you’re interested:
They don’t look photoshopped to me. I don’t think the effect is anything to do with the animation. That just makes it easier to keep looking at the cross and not directly at the images.
Yes, I have done the same thing now (I also had to download a GIF editor suitable to the purpose, so it took a while). They seem to be random celebrity photos, pretty much randomly shuffled together, and, as you say, with most of them used a few times over.
I can’t even see the damn gif on my phone, but from your descriptions of it, the following may be what you are looking for.
Just Google: flashed face distortion effect.
Here’s some background - A new and bizarre illusion – Why Evolution Is True
There’s a link to a pdf on there, but my phone doesn’t do pdf, so I don’t know what’s in it.
Here’s a link to the research abstract. perceptionweb.com
I can’t access further than that front page though. Google might have more links, but I’ve got some other stuff to do.
I am not sure what you mean by this. I think it is quite clear that the effect is produced by the animation. Do you know of a way to get the effect without it?
Anyway, I an assure you that the animation in peripheral vision does not make it easier to keep the cross fixated, it makes it considerably harder. It is well known that moving, changing stimuli in peripheral vision attract attention and tend to draw the gaze toward them. Furthermore, it seems to me that you do still get a weak version of the effect even if you look straight at one of the sets of changing pictures. (I am not sure if anyone else in the thread had yet confirmed or denied that they see this.)
Anyway, I am pretty sure I know in outline how this is working. As I have said, the visual analysis of different aspects of faces (e.g., eye position, mouth shape) will depend on different visual routines, some of which will take longer than others. The pictures are changing too quickly for some, but not all, of the routines to complete, so we are seeing some aspects of a new face before we have finished analyzing certain other aspects of the last one. This causes the features belonging to successive faces to get mixed up, so that we may momentarily experience (for instance) something like Sean Penn’s general face shape, and nose, with Mila Kunis’ eyes and mouth.
(Those re just hypothetical examples. I do not know, and I do not think much is known about, which particular features of faces have distinct visual routines dedicated to them, or how long the different routines might take to complete their analysis. This demo, though, suggests that the time required for at least some of the routines to complete is a bit over half a second, which is the frame rate of the gif. Other routines must be completing more quickly, but it is not necessarily all that much more quickly.)
The effect is present for me in the static images, although stronger in some than in others. Halle Berry + red lipstick lady is a good one. Generally I find that if I cue up an image without looking at it, cover the faces, look at the cross and then uncover the faces, the effect is there. It also helps to stare at the cross for a while.
I do agree that the effect is more obvious with the animated images.
Please have her look at it and then report back what your Guinea kid perceived.
Is it me or do some of the images elicit repulsion/fear in others? If I caught a glimpse of some of those faces while in confined/dark places, I’d have a fear response.
Thanks. It looks like that is it.
From your info I found the actual paper [PDF] via Google Scholar (although it is unillustrated text, and so not nearly as impressive as seeing the actual illusion “in the flesh”). I don’t think I entirely agree with their hypothetical interpretation in terms of contrast effects, however (although those may be playing some role too - many illusions involve multiple mechanisms). I am sticking by the theory I gave in my previous post.
Hmm, yes the eyes do seem to stand out abnormally when those pictures are kept in peripheral vision for a while. That is probably making a contribution to the overall effect, but I do not think it is the whole of it, and probably not the main part. The discoverers certainly think that the animation is what is important, and, as I have said, I see the effect even when I do not look at the cross, but straight on at one of the sets of changing images. It seems to take a little longer for the effect to kick-in that way (it takes a second or two even the regular way), and it may not be so strong, but it is definitely happening.
They do look like celebrity photos taken without posing, so that most of them are at odd angles and the faces have funny grimaces, squinting, unusual cropping, and as mentioned by others, variations in lighting. These are not posed shots. I don’t know if that’s responsible for the effect, or not. I think peripheral vision has a lot to do with it.
This propensity for facial recognition explains something I find myself doing all the time. When I’m standing at a urinal or toilet peeing and the wall in front of me is textured or irregular, I pick out faces in the bumps, spots and rough areas I see. It might be a goofy smiling face or an evil sinister face. Does anyone else do that?
No?
Never mind.
Bumber,
I’ve had that with some shower curtains or blankets. I think pretty much everyone has seen easily recognizable objects, especially faces, in clouds. When I was in the reserve, there was a ver well camouflaged guy I spotted and the only thing I saw of him was his face. I didn’t even have to make an effort and he certainly had to hide himself.
So, what was crucial for survival when it comes to faces?
Recognizing that it is a face. This allows us spot other humans, especially those waiting in ambush.
Recognizing what the human is looking at. To see whether he has seen us, try to guess what he’s thinking and planning to do.
Facial expression. To recognize his mood. Knowing whether he’s friendly, neutral or hostile can be critical.
Are there other elements in what I’ve named?
Are there other elements besides those I’ve named?
The pictures are normalized to put the eyes in the same location (there’s one exception where a man’s face is too small). So people with smaller faces or narrow eye spacing get enlarged relative to larger faces or eye spacing.
Another point that may or may not be relevant: Whoever put this together seems to have made a particular effort to keep the eyes of the faces lined up.
EDIT: Yeah, what ZenBeam said.
Since the eyes are lined up with the cross, the other facial features (forehead, nose, chin) keep growing or shrinking around the eyes as the faces flash by. Our brain tries to make sense of the rapidly changing images in our peripheral vision, producing the distorted features we “see”. It’s sort of like watching a movie, where the film moves at 24 frames/second, producing the illusion of motion.
As I have now repeaedly pointed out, and as scientific article (to which I linked) whic describes teh discovery of this effect also makes clear, teh effect does not depend on peripheral vision, and also occurs when you are are looking directly at teh rapidly changing pictures.
Ximenean does seem to be correct that a subjectively somewhat similar effect can be produced with a static face (or perhaps it has to be a pair of faces) in peripheral vision for a while. The originally posted demonstration may involve more than one illusive effect. However, you can definitely achieve the effect in central vision. Try it!
Whatever. (Presumably you learned that from the link I posted.) My point was that the effect does not depend on the faces being distorted, or in any way weird, extreme, or unusual.
I admit that it’s only my perception but while I perceive the same effect when looking directly at the pictures, the effect seems several times stronger when looking at the center cross and keeping the face in peripheral vision. Is anypony else getting that too?
Yes. That’s what I see. Covering one side reduces the distortion, as does viewing from a distance. But there may be multiple effects at play. There does seem to be something that happens just as a result of the pictures flashing.
I just had another go with Ximenean’s static images, and I think that* with the static images in peripheral vision the full effect only appears when the two different images are there on either side. When I covered up one, so that only one face was in peripheral vision, the precept did seem to simplify and lose some of its features after a moment or two, but I do not think it took on the weird, distorted cartoonish appearance that we are otherwise seeing. I only say this tentatively, however, because I am very tired now, and cannot hold fixation for very long.
Yup. Patterns & textures in fabrics, carpet, etc. We find patterns in things. It’s part of who we are.
