Running water + 24Hz Sine wave =

[GaijinPunch]

This.

Okay, science dorks. Enlighten me!

[moh]

*edit* it would help if i watched the video

[KAI]

Water, so insipid and fascinating at the same time.

[BIL]

Water, man's greatest invention.

[null1024]

That was way cool, but I'm wondering if that's an artifact of the recording [and how it would look if I was standing there], because the 23hz adjustment had it look like it was flowing upwards [which was way way way cool].

[GaijinPunch]

There is a 25hz version where it looks like it's moving downwards a little. So does the frequency have something to do with the recording rate of 24fps?

[Moniker]

There is a 25hz version where it looks like it's moving downwards a little. So does the frequency have something to do with the recording rate of 24fps?

My guess would be that you just need to match up the frequency with the framerate. I remember a high school physics demonstration where someone was swinging a jump rope in front of a strobe light. Depending on the speed of the strobe, it would look like the rope was frozen in a sine curve, or undulating really slowly up, or down, respectively. In other words, a series of snapshots of a continuous flowing movement can give the illusion the movement is doing weird things.

It might be that water molecules sympathetically vibrate at 24hz - that's basically how microwaves work at 2.5Ghz - although I don't get why it would happen on such a macro scale. *shrug* NB: Couldn't watch the clip as my computer hates youtube, apparently.

[GaijinPunch]

Couldn't watch the clip as my computer hates youtube, apparently.

Wow... how do you make it through a day of work?

[jepjepjep]

There is a 25hz version where it looks like it's moving downwards a little. So does the frequency have something to do with the recording rate of 24fps?

That's a really clever experiment. Yes, the recording frequency is what is causing the optical illusion. The water hose is oscillating at 24Hz, so when the camera is set to 24 fps, it looks like there is no motion. Take a look at the clear hose: at 24Hz the hose looks like it's perfectly still. When the camera is at 23 or 25, you can see the hose oscillate. That's why the water looks like it's going forward or backwards depending on whether it's 23 or 25 fps, it's an artifact of the sampling frequency. The water is good for this experiment because once it leaves the hose, it's velocity is fixed; you can imagine that the velocity (the direction) of each "patch" of fluid changes because the hose is moving back and forth.

[Ed Oscuro]

The forward effect is easy to explain using old Hollywood movies. Imagine an old Western where some hick is fleeing from the Indians on his trusty Conestoga wagon (very realistic of course). We know that the wheels are turning forward the whole time, but gosh! On the screen, the wheels turn backwards! The way to think about this is that the wheels are being captured at an instant (I am simplifying away shutter speed and angle here, although for digital cameras it's closer to the truth, especially if the shutter exposure time is quite fast, as is the norm on many digital cameras' movie modes), and at the next frame capture, it captures where the wheels are aligned at that moment, instead of any actual representation of the wheels moving. Arrange for the right speed of the wheel, and at the moments the new starts, wheel spokes (not likely the same ones, of course) will appear slightly behind the position of those in the last frame. Change the speed of the wheels or the number of frames per second of the camera slightly, and the wheels may appear to turn forward again or even to stand still. A few weeks ago I thought the front fan on my new case was doing just this (which suggests that somehow the LED lightbulb overhead was pulsing at a rate to illuminate the fan blades more strongly at certain moments).

Now the acoustic waves bit I'm a bit more rusty on. By superimposing two waves you can produce a (nonexistent, in the original signals, but quite obviously existent in the actual hearing) "beat" noise, which is merely the coincidences of the two (or more) waveforms. But here there is really only one wave to worry about - the speaker's output. Right? Oh, actually, the frame rate of the camera can in effect act as another tone, able to record inputs only along one point in the sine wave.

Imagine if there were another speaker brought in front of the one currently seen. If the wave being sent out the speaker was matched by another tone generator driving the new moving speaker, the wave should be amplified (magnification of waves - simple loudness increase, although I'm not sure how much). if the second wave was temporally shifted so that the peaks of the one tone coincided with the valleys of the other, you get what's called destructive interference and the water would run free again. And, if we forget about the camera for a moment longer, if you produce a 24Hz tone from one speaker, and a 25Hz tone from the other, the peaks and valleys are continuously lining up and separating, moving (slowly) through the entire range of constructive to destructive interference. This would probably appear as the water "jumping" out in a line roughly once a second. If the second wave is 26Hz I believe the wave will appear twice a second (I could be way off base with this math though).

Notice where the speaker is oriented, and where the hose is. The hose runs over the speaker and so does the water run past the speaker on its way down and out. The sine wave is simply a pressure pattern - from high pressure to low, and back again. In the high pressure portion, the overpressure wave (I wonder how much force this takes? Not enough to be dangerous to the person, apparently) is quite literally pushing the water out further (I'd be interested to know if this pressure difference is built up inside the hose, which I think, or immediately after it exits the hose). In the low pressure portions, the water is able to flow along the course it would take if there was no pressure wave.

Now add what you know about how digital cameras record images to explain why it is that only at those moments where the water flow is configured a certain way, visually, that the camera records the images, rather than those moments where the water has fallen below. Of course, as with the wagon wheel spokes, when the camera moves around the apparently frozen water scuplture you're not actually looking at the same droplets from frame to frame. They might be forming the next "step" of the zig-zag pattern in the next frame, and then the next "step" in the frame after that, and so on.

also agh that tone hurts my ears

[neorichieb1971]

Its the wagon wheel effect that makes it look like its going upwards and downwards.

Whats more interesting is that its the pipe causing the patterns, as you can see it sway with the frequency level.

[jepjepjep]

The water hose itself doesn't naturally oscillate though, the flow is being driven by the difference in pressure between the source (faucet) and the atmosphere. There's no interference because there is only one frequency at play, which is the speaker and it moves the hose (the point of exit of the water) back and forth.

[Ed Oscuro]

I don't think the hose is responsible for much of the movement, since the water is spraying rather significantly past what you'd expect from the pipe shaking back (I need to look at it again). I believe rather the pressure from the speaker is pressurizing the water slightly within the hose itself (it would be neat to get a very scientific backpressure reading from the water source). I could be wrong though. Edit: On another look, I did notice the hose jumps a good bit when they turn the speaker on. Most of the way to explaining the distance of the water quite possibly.

In person I'd expect this to look like a messy fountain spraying everywhere like somebody spitting water between their teeth.

[jepjepjep]

This is why the water looks the way it does.
http://en.wikipedia.org/wiki/Nyquist_frequency

http://www.youtube.com/watch?v=Fy9dJgGCWZI

In person I'd expect this to look like a messy fountain spraying everywhere like somebody spitting water between their teeth.

I think so too.

[Ed Oscuro]

Nyquist is the boss. That video is interesting but I wonder what's going on with the comments...

[GaijinPunch]

...or a video I just saw from the bowels of the internet.

[Ed Oscuro]

Bowels can be fun.

[MOSQUITO FIGHTER]

neato

[casualcoder]

I'm quite sure this is related to a well known and common effect in audio that we audio engineers call phase cancellation. Basically, when you take a simple oscillating waveform and duplicate it but invert one of them they will cancel each other out. Meaning you won't hear anything. It's sort of like telling your speaker to vibrate inwards and outwards at the same time. Kind of like equal opposing forces canceling eachother out.

24hz is a very low sound pressure wave. Its so low that you can easily see its effect on water (think of that scene in jurassic park where that dinosaur plants it's foot down far away but they see the ripple in the water). Anyways, since 24hz is 24 cycles per second, and 24 frames per second is essentially the same thing you get cancellation. Likely the cancellation is only occurring at peak amplitude and that's why you still see some of the water and then a gap (gap being the cancellation point). This is purely a video effect.

Funny this was a topic today. I was just now playing pink sweets and had to run a mono cable through my audio card for sound. Of course i forgot that summing the signals would cause cancellation so tons of effects and music were completely silent. Cancellation is such a pain in the ass when dealing with audio but it's experiments like this that remind you how you can learn a lot from it.

[GaijinPunch]

Funny this was a topic today. I was just now playing pink sweets and had to run a mono cable through my audio card for sound..

You sure do have balls posting that around these parts.

[mesh control]

fucking miracles man

[jonny5]

Weird, my buddy texted me a link for this article this morning. Neat stuff

[guigui]

Some car wheels look like this



Those are the same phenoma :

Forward flow : you see the car at a certain speed and wheels seem to go slower, same direction.
Fixed flow : you see the car at a certain speed but wheels seem still.
Reverse flow : you see the car at a certain speed and wheel seem to go slower, reverse direction.

All of this depends on the frequency at which your eye (i.e. the camera) takes pictures.

[Ed Oscuro]

http://www.geek.com/articles/geek-ceter ... -20130313/

[rapoon]

in addition to Ed's post:

http://en.wikipedia.org/wiki/Rolling_shutter

[Ed Oscuro]

Looks like good info, but it seems somewhat ironic to title the video "rolling shutter" (as in...shutters that roll) and then have a bunch of example images taken with no physical shutter.

[neorichieb1971]

In relation to the point that people seem to think it looks different in real life. Wouldn't that mean the shadow would be different on film? The shadow reflects the exact motions of the water.

[Ed Oscuro]

Shadows aren't painted on. The speed of light is high enough that you can assume that the shadow represents light that's being blocked at that moment by the water stream. If we catch the stream at a different moment, the shadow will move, and you can see this when he plays with the speaker's sine wave.

Thinking about it further, though, the water might just be waving back and forth, I think saying it's like a messy fountain is wrong. It wouldn't be misty for example, and I'm not sure why I implied that. Basically it probably looks like a rope of water moving up and down quickly, just like you've shaken a hose back and forth.

[rapoon]

Looks like good info, but it seems somewhat ironic to title the video "rolling shutter" (as in...shutters that roll) and then have a bunch of example images taken with no physical shutter.

check this out: oh sweet satan!

detailed information + additional vids showing the effect as demonstrated in GP's post

[Ed Oscuro]

I like to try and forget about CCD devices.

I've seen the information in that article before, many times. Rolling shutter sounds very close to rotary shutter, don't you think?