Amazing optical illusion or "glitch in the Matrix"? | Hacker News Reader

Amazing optical illusion or "glitch in the Matrix"?

cbsnews.com

215 pointsshawndumas60 comments

Strictly speaking, it's not an "optical" illusion at all -- you must film the drops with the camera. You'll see the effect only on the screen.

Still, a clever idea to simply use a sub-woofer to kick the drops. I like it.

The same effect can be achieved in real life by adding a strobe light.

There might be a little bit more to it than that. I think the vibrating subwoofer actually nudges the droplets and shapes them, controls the flow and the way each individual droplet breaks off.

Anyway, I'm definitely going to try this.

I think the idea was to have a vibrating subwoofer and a strobe light, where the strobe light would simulate the synchronized framerate.

It should work. I would try it but strobelights give me a headache.

Oh, I see. So then you wouldn't need a camera.

Even at high frequencies? Because a strobe light gives you a lot more flexibility.

Indeed, when I was at MIT there was a exhibit in the Doc Edgerton hall that did exactly this with florescent food coloring for the water and a strobe. You could turn a dial and change the speed of the strobe, and the 'drops' would change direction into or out of the spouts. There were two spouts with streams hitting each other. A very cool effect.

It's been eons since I've been there, but I believe a reproduction of this is set up at the Boston Science Museum.

There's also one in the Hong Kong Children's Science Museum.

Yes I also think that if you can use strobe light and have the possibility to adjust the strobe frequency you don't even need a subwoofer, but for the proper "matrix-like" effect a lot of "frames per second" are needed and the camera gives certainly nicer result. In my opinion his setup was really quite optimal for what he presented.

See my comment above. You actually need the sound to create vibrations in the stream of water that breaks it up into drops. The coolest thing is that there really are drops here, it's not an optical illusion.

I'm not quite sure I'm following you. You don't need the sound to break the water into drops. You can do it without the sound by just adjusting the flow rate. I've done this.

Are you just saying that by using sound, you can break up an even faster stream, one that would ordinarily be too fast to break up on its own?

With sound, you can break the stream into the same pattern of drop every time

I wonder if there's any way to make this into a decorative fixture? A strobe would be annoying, and a large LCD shutter in front would be expensive and spoil the effect somewhat. Any ideas?

Yes, I was a bit disappointed, I was hoping that the effect was being caused by the AC mains frequency effecting the lighting. Ah well

I spent a few years of my life staring at this exact effect. I was working on building a big industrial inkjet printer. The physics behind it is cool too, but not new. Lord Rayleigh first described it in 1878. Not exactly a glitch in the Matrix :). I have a patent somewhere on a minor tweak to the concept, but can't find it right now. Good times.

More info:

Just to clarify for everyone assuming that this is an optical illusion. It's not. The subwoofer is actually creating vibrations in the water stream that increase exponentially as the stream travels until their magnitude is larger than the diameter of the stream. At this point, the stream breaks up into drops. If you do this with a coloured stream of water (i.e. ink) and run a piece of paper under it, you will see individual drops of ink hit the paper.

There are two types of inkjet printers. The most common is what you see in household inkjet printers, it's called Drop-on-demand. You spit out a drop by doing something to the ink. Ofter you heat up a resistor, converting some of the ink to steam and pushing out a drop of ink.

The other type is called continuous inkjet. The way this works is you have a stream of ink, you stimulate it at some frequency which gives you drops at that frequency, and then you select some drops to hit the paper and others to miss.

The project I was working on used conductive ink and selectively charged the drops before sending them between two oppositely charged plates, sort of like a cathode ray tube with ink drops instead of electrons.

It was ridiculously cool technology, and a lot of fun to work on. The printer we were designing (never got to market in that incarnation) would have had paper going by at ~6 meters/second with very high quality, totally variable output. I thought we were going to disrupt offset printing in a big way. Sigh :).

[edit] Ooo, found the patent. Great bedtime reading if you're interested: http://bit.ly/HXc6o2

If you don't mind me asking, what ended up killing it? It sounds awesome.

Two companies doing the same thing both got bought by the same very large company. The other company kept going. We got shut down. I got to leave and start my own company (Leanpub), and everyone lived happily ever after.

It was a great project, but I'm much happier now doing the startup thing than I was being a physicist.

I used to intern at a company that used printers that had that technology (I assume from your competitors). Massive, massive printers, but they were stupid fast.

You might be able to get a similar effect to the naked eye, if you used a sound wave at/near the same frequency as your local mains electricity, under incandescent mains lighting. A harmonic frequency might even work, too.

I recall seeing a tool for fine-tuning vinyl turntable speed that used that form of the trick - it was basically a cardboard "record" with spokes printed on it every (* grabs a calculator *) 4.02deg, so as to appear to be stationary when rotating at exactly 33.5rpm under light flickering at 50Hz. Thought that was ingenious.

I still have a turn table and it has stroboscopic marks on the rim for just that purpose. :-) Sadly I only have one more cartridge for it and after that, no more replacement needles.

I prefer the water and corn starch on a speaker [1] to get standing structures of weirdness.

[1] https://www.youtube.com/watch?v=SbTHExJzJXs

Here's a slightly more fascinating video of Oobleck (water + cron starch) on a speaker: https://www.youtube.com/watch?v=Yp1wUodQgqQ

Technics turntables have strobe dots on the rim for calibration: http://www.youtube.com/watch?v=XLtlELevKdk#t=0m12s

Damn, so that's what that's for!

s/incandescent/fluorescent/

Incandescent (filament) bulbs glow by black-body radiation as the filament heats. Mains cycles are too short for the bulb to flicker (though you can get interesting effects on slower, ~2Hz or thereabouts, voltage fluctuations).

Fluorescent and LED lighting actually flicker multiple times per second, though this can be tuned somewhat.

If you remember CRT displays and monitor flicker, the problem wasn't so much the flicker rate of the monitor, but the interference in rates and timing between your monitor and the overhead office fluorescent lighting. Setting refresh to anything other than (and preferably above) 60Hz (US) would resolve this.

I stand corrected.

I'm sure I remember my parents calibrating their turntable under incandescent lights, though... maybe it was their halogen up-lighter instead of their standard incandescent ceiling light? It definitely worked with one light in their living room but not the other.

Aren't LEDs fed with DC? If so they would obviously not flicker but I'll admit not cracking my LED bulbs open just to see, as they're expensive enough.

LEDs are DC. What I've observed in certain applications, particularly LED taillights/brake lights, is a flicker pattern which I suspect is either an AC/DC conversion artifact, or a brightness regulation. I find it quite distracting.

It's also fairly well known that LED's response to voltage variation is fast enough to be of concern in highly secure data environments. Demonstrations have been made of reading line signal from modems, and even Ethernet and other high-bandwidth networking equipment status/indicator lights.

LEDs are generally fed with DC, as current only flows in one direction, but they flicker as the power coming into them fluctuates -- the more ripple, the more flicker.

Cheap LED holiday lights don't even bother rectifying the AC voltage across the lights, so they are only on for half the cycle, giving them a very noticeable flicker. It's especially apparent as you scan your eyes past them.

In contrast, incandescents allow current to flow both directions, and will glow for a little while after current stops flowing, so they don't appear to flicker on 60/50hz mains current.

The turntable from A Clockwork Orange (Mitchell "Transcriptor" - my dad owns one) has a strobe and a marked wheel for adjusting its speed.

See the second image - http://www.1stdibs.com/furniture_item_detail.php?id=333935

Sure to cause humorous results for us Americans with our 60 Hz mains, if we were to import such a tuner.

They often have a 50/60hz switch on them.

Since water droplets look alike, you get the illusion, at the proper frequency setting, of water dripping up.

I wonder how it would look if a different dye was injected into the flow every few seconds, just above the faucet. Presumably, the colors would appear to flow down as the water appeared to drop up.

Here's a crazier version: https://vimeo.com/4041788 The camera takes frames at 30 fps. But each frame is only exposed for 1/4000 of a second, which lets you see the string as if it is holding still during the frame. And finally, the camera has a "rolling shutter" so the bottom of the frame is captured about 1/50th of a second after the top. This creates the visible wiggle in each frame, as the string vibrates several times as the frame is recorded from top to bottom.

I think the relevant wikipedia terms here are "Aliasing" and "Nyquist Frequency"

I would have said "Poincaré map", as that's basically what this is, in some respect.

Whatever you want to call it, it's a very cool effect.. (edit: but thank you for providing some educational context! Off to Wikipedia I go..)

Also [wagon-wheel effect](http://en.wikipedia.org/wiki/Wagon-wheel_effect).

I mocked someone when he said he could see the effect in normal daylight with his eyes, but in fact I checked and could also see it sometimes! http://en.wikipedia.org/wiki/Wagon-wheel_effect#Truly_contin...

These kind of computer vision illusions are always really fascinating to me. It's kind of like a natural zoetrope effect.

More: Bass String Video: https://vimeo.com/4041788 The Phonotrope: http://www.jimlefevre.com/?page_id=200

Bill Viola did this using strobe lights in an installation format. You can recreate this yourself easily with a programmable strobe. Thanks to Edgerton for the inventions. http://en.wikipedia.org/wiki/Harold_Eugene_Edgerton

My question is would this be possible with a strobe light? Or even better, a florescent light? That would make some very impressive shows, if you could have a whole hallway with different streams, all synchronized to the frequency that the lights strobe at.

(http://hackaday.com/2011/03/11/water-droplet-sculpture-using...)

There are plenty on YouTube. It is an amazing illusion.

I find it hard to understand how "a drop" seems to stay the same shape. I need to think a bit more about that.

think about an old tire rolling through paint. you'll see the same artifacts every revolution of the tire. the key idea is every cycle creates the same pattern.

so, you could have a mechanical system that hit the tube every /24th of a second, and you'd get some effect, but i think it would be hard to get the impact and recovery to be EXACTLY the same, or close enough you don't notice.

If you use sound to squeeze the tube, rather than pushing on it, the distortion and recovery are much more consistent, so you get the same pattern. I'd guess you don't have much control over where the drops actually appear. you'd have to wiggle the tube around on the speaker to get the right distortion.

It's a 'standing wave'.

I've done it in an almos empty disco, with some strobes and a cord with a key(or any other weight) at the end. I turned around the key (like using a sling). Playing a bit is pretty easy to make the key stop in the air like pointing in the direction you choose. Also you can see the effect in some videos of flying helicopters, they seem to be flying with the rotor freezed. It is Pretty cool!

That's an interesting idea, and it should be possible with a strobe. A florescent light oscillates too quickly though; the gap between the water droplets is going to be inversely proportionate to the frequency of the sine wave.

I've seen it at a children's science museum, forget which one. They had a dark box you could look into with some kind of strobe which produced this effect in the box.

There's an exhibit at the Boston Aquarium that is exactly this with a strobe light. It's really cool to see in person.

Does this signal our perception of the universe is structured by the energy's surrounding us? We did evolve along with all of these energies.

The human mind is malleable and dynamic, not fixed like a 24 FPS camera lens (I imagine most humans don't have a fixed framerate for incoming sensory) - I do wonder if other critters with less dynamic minds would fail to some kind of trick.

I wonder too if it was this kind of dynamic ability that allowed us to disconnect from a fixed reality, thereby giving us contrast, the ability to analyze our surroundings, and therefore consciousness.

zoetropes on bicycle wheels: http://www.designboom.com/weblog/cat/8/view/17851/bike-wheel...

really an optical illusion? I heard two different flows of water splashing in the bucket which I would not expect if it was just an optical trick.

Just another trick like the "hovering" russian helicopter...

The matrix didn't go BSOD when Crysis launched, it wont go because of a reverse stream of water...

my dog started acting up when I turned this on, any relation?

Camera's are not the same as reality. They take images at a certain number of frames per second. 24, 30 or 60 are common frame rates for digital cameras. Thus, the camera is effectively its own strobe.

Since the frequency in the speaker is 24 hertz and camcorders can be set at 24 frames per second, it is likely that this is what is happening here-- the speaker frequency is causing pressure differentials in the water, the water's surface tension causes the succession of drops to form similar shapes, and the camera taking an image at the same frequency of as the speaker makes things look still because the drops are in exactly the same spot each frame.

You can notice on one of the lower bubbles further away from the spout the effect is less precise and the bubble seems to be changing shape or moving around.... this is because the effect of he speaker on the water decreases over time as the water gets further away and air turbulence causes the drops shape to pick up more randomness.

We can read the description on the video, you know.

Another trick is to take a rope and wiggle it at say 2 to 20 hz and take a strobe light to light up at 1 to 15 hz. In the dark the rope will show a perfectly formed sin wave which is spooky to look at.

Another one is putting clear plastic wrap over a subwoofer, making it face upward, and sprinkling salt on the surface. The patterns it creates look like alien symbols as you explore the spectrum.

This is your universe at 24 Hz.

It usually looks like the universe at any other reasonable frequency, unless you're doing weird things at a multiple of 24 Hz.

Cool.

Reply on news.ycombinator.com