Bad Astronomy

UPDATE: Fake TV Ads, Pixels, and Parallax

One way to screen applicants.

Photo by LG, from the video.

Yesterday, I ran a post I figured would generate some buzz, and it did. But it turns out I was wrong about why it would do so! To make it more fun, I wasn’t exactly wrong in the article, but I wasn’t exactly right, either. There’s enough to warrant a brief update, because honestly, this stuff is cool.

In the article I dissected a commercial for an ultra-hi-def 4K TV set, where people walk into what they think is a job interview, but is actually an elaborate prank. During the interview what they think is a window on the wall is actually a TV showing a cityscape. During the interview, an asteroid is seen coming in, slamming into the city, and wreaking havoc.

The ad is to show that the TV display is so good, so real, it can fool people into thinking it’s the end of the world (or their local piece of it). I wondered, though, if that claim is legit, and if in reality people would notice the TV wasn’t real due to seeing the pixels in the screen (which in turn would mean the ad was faked, using actors). So I did a little math and concluded that given the sizes of the pixels in the TV, the resolution is indeed high enough that pixelation wouldn’t be a problem. Therefore, I labeled the ad as “plausible”.

That’s where I figured people would quibble; the resolution of the human eye is a fascinating and eminently arguable topic. But in fact, that’s not what happened. I immediately did get a lot of responses in the comments, via email, and through Twitter, but they were not about my math. They were about parallax.

Parallax is the effect where you see a different perspective on an object as you change your viewing angle. For example, hold your extended thumb in front of your face. Now close one eye, note your thumb’s position, and then open that eye and close the other. Your thumb will appear to move as your viewing angle changes.

Parallax
Parallax works by changing your viewing angle on a nearby object. Viewer A sees the star against the blue background, while Viewer B sees it against the red one.

Photo by Wikipedia/Booyabazooka

This effect is what gives us depth perception, and tells you why nearby objects like trees fly by when you drive past them, but distant mountains move more slowly.

After reading some of the comments on my article, I agree that parallax is an issue in the ad; if you paid attention walking into the room you’d notice the view on the TV screen seems flat, with no depth. As I note in the article, though, I don’t think a lot of people would be paying attention to the window; they’d be nervous about the interview, and would be paying more attention to the interviewer. So that by itself might not be enough to claim the ad was faked with actors.

However, there’s more. I chatted with Randall Munroe (of xkcd) about this, and he noted that the viewing angle of the prank victims changes when they sit down versus when they’re standing up. The “window” shows the tall buildings and much of the ground in front, and it’s unlikely they’d see the ground once seated if the window were real. That would definitely look very odd. Also, he noted that as the victims walk into the room the view out the window would change as their distance to the window changed; they’d see more to the sides as they approached. Try this yourself; walk toward a window and you’ll see the overall view left/right changes. This is pretty obvious, and should at least register subconsciously.

This video should help you picture this. It’s actually extremely cool:

I’ll note that I agree with these arguments. Parallax may be the key issue here, and anyone paying even a small amount of attention should notice it. But still, I wonder… my friend (and evil twin) Richard Wiseman is a psychologist who has literally based his career on people not noticing things. This is called change blindness, and he has masterfully made videos demonstrating it. My favorite is “The Color Changing Card Trick”.

Did you catch all the tricks? I’ve seen many videos by Richard where people miss things so incredibly obvious you’d swear the video was faked… but it isn’t. People really can totally miss obvious things if they’re distracted.

Now, this isn’t to say the video is real, that is, not faked with actors. It very well may be. In fact, after reading the comments I think I’m leaning toward it being faked. I’d bet that way, but I can’t prove it using these parallax arguments because I know how important these psychological effects are. It’s possible they filmed 50 people, 47 of whom weren’t fooled, and three of whom were.

I’ll note that other commenters pointed out various other factors as well. For example, some folks claimed a TV doesn’t have the contrast to realistically fake an actual window, but I wouldn’t agree with that necessarily. The room could be darkened a bit to brighten the TV, and modern TVs have astonishing contrast between light and dark. That kind of thing can be controlled by an accomplished prankster, as can other factors. Again, I think this falls into the “maybe, but I’m not convinced by this particular argument” category.

Also, a mea culpa, kinda: After reading some of the parallax arguments (but before chatting with Randall) I updated the article acknowledging I should’ve discussed them more, but felt they weren’t critical. Clearly I shouldn’t have dismissed them so casually, but I still don’t think they clinch fakery. What we really need is to do this same prank under more controlled circumstances and see how people react. I think that would be fascinating (though maybe a less intense view out the window might be useful; some folks can be shown just a static scene, while others might see an airplane or flock of birds moving across, or something else in the foreground).

In the end, I think my resolution argument stands, and the pixels won’t give away the trick—and that was in fact the narrow point of my post, and I should’ve made that more clear in the end. However, parallax is certainly a different issue, and clearly provides a different angle on the argument.

We’re gonna need a bigger Adam Savage.

Photo by Phil Plait. Rings by Ryan Consell.