Hacker News new | past | comments | ask | show | jobs | submit login
Real Life Model of M.C. Escher’s “Waterfall” (laughingsquid.com)
254 points by mikecane on Feb 18, 2011 | hide | past | web | favorite | 50 comments

Despite how it looks the yellow paddle wheel is not actually at the start of the track. The rectangular box at the start is not connected to the actual start of the track. It's far away from it, and it holds the yellow wheel and a bucket to catch water.

Next, ignore the beams, and focus only on the water track. Despite looking like it's climbing up, it's actually totally flat (i.e. you could make it from one sheet of wood). But tilted slightly so water flows from the start to the end.

Additionally the entire structure is not actually sitting on the garage floor, but rather is high up in the air, and the camera angle obscures this.

Hidden under the structure is a pump and an outlet at the true start of the track.

He pours water into the catch basin at the end of the track, and that same moment someone else starts the pump. If you watch closely you can see how it looks like the water he pours is being "swallowed", and does not actually flow down the track.

The water from the pump flows down the track, and out the spigot over the wheel and into the bucket.

Next you place beams carefully cut at angles to make them look like they are holding the track - but actually they are open on top.

But this is a pretty amazing piece of work.

When he walked towards the camera at the end, I was really hoping he would change the perspective to reveal the illusion.

It pissed me off when they did that in Inception. The whole point was that should be possible in a dream.

But it was; the walking worked, despite how it looked.

Right up until the camera panned. Suspension of disbelief shattered, and to what end?

Are you kidding? If they didn't do that, how would they demonstrate a trap that made any sense? That would have been far worse, I think.

Yeah maybe I totally forgot what actually happened, I just remembered my irritation :)

I was more irritated by the fact that the entire movie was clearly ripped from the chapter "Harmonic Labyrinth" from GEB.


The apparatus is not connected at the red line in this image. The part to the left (the wheel) is on the ground and farther away, the part to the right (the start of the track) is in the air and closer. They just happen to line up due to forced perspective.

The water just seeps down into a hidden reservoir in the left part, and is pumped up at the appropriate time from another reservoir in the right part.

Edit: see below, it looks like the discontinuity is at a different point.


It looks like, given the reflection from the water, that section of the track can't be more than an inch above the floor. If thats true then wouldn't the break have to be somewhere after that point?

If I'm missing some huge detail can someone explain it to me?

I think you might be right. The high-def version makes it easier to see where shadows fall. It looks like it's the back section that's separate from the rest. If you watch the high-def closely it looks very much like the first 3 segments of the track are behind the last segment (notice how his shadow passes over the last segment THEN the first 3 segments):



Ya, if you watch it carefully you can see it burp a few times when the water moves past that post.

EDIT: I made a diff in GIMP of the shadow. It looks like both the top and bottom left facing ledges are connected. Notice the hard edges between the adjacent pillars.


It looks to me like the break is at the first bend - on the HD version, the water coming in doesn't appear to turn a corner there at all, and then on the other side of the tower support it's flowing straight outwards.

I thought that too but there is a very convincing shadow on either side of that pillar.


Excellent. Notice that the same shadow is missing on the last part of the track, at least in the part to the left of the vertical pillar.

Edit: http://i.imgur.com/Kqgtr.jpg

That may be the legs that are holding the whole thing off of the floor.

The legs would have to be mounted in such a way that they were hidden by the structure. So they'd have to be at an angle and those in the reflection appear to be vertical.

I can't believe no one noticed it.

The texture of the upper parts of the wood contraption feels unrealistic; there is almost no grain. The contraption itself is too bright compared to the ambient level of light; the white buckets are darker. There are 2 main sources of light: one on the right casting a shadow of the structure on the ground toward the left, and one light on the left that should cast a shadow on the ground toward the right... but it is missing (contrary to shadows on parts of the structure itself)!

This is obviously CGI. You would be surprised what a person with a few days of work can accomplish with off-the-shelf CGI software.

The poster's background suggests this. He writes "ImD-student" in his profile which may refer to the Interactive Media Division at the USC's School of Cinematic Arts. He also subscribed to the "indymogul" youtube channel which is an online video network and community dedicated to DIY filmmaking.

With this in mind, it's interesting to look at the "set design": miter saw, clamps, drawings, sand paper, wood scraps, sawdust and broom all strategically placed around the scene to add realism.

For giveaways thought, in addition to the strange shadowing and textures the parent mentions, there's the human factor. To me the acting seems just a bit off.

Kudos to the guy who made this though; I imagine just modeling that structure (and its shadows and reflections) in CGI was a fun nut to crack!

One more clue supporting my theory: the poster updated his profile, and ImD refers to http://www.intermediales-design.de which is an education establishment specialized in design, offering specifically concept / 3D modelling courses amongst others:


tl;dr 'shopped.

I gave it my best shot in SketchUp, and in the process discovered why I am not a 3d artist:


I think if you look closely you can see the barest edge of a bucket behind the central part of the structure, where they'd drain the actual initial liquid. (You can see that the fluid seems to arrest before that transition in the video, as well.)

and he edited the video so that the fluid flow looks smooth when the water reaches to upper most brigde. look closely at 0:43 there is something strange happening to the fluid flow. but alas great work!!

From what I can figure, it's the usual perspective trickery (flat water track, vertical tower, perspective makes it look like they line up), and then there's a nozzle pouring out water from the tower top.

Note that the water doesn't appear to flow in front of the wooden knob at the ledge it's supposedly falling off.

EDIT: Looking again at the HD version it seems I'm mistaken. ars' comment appears to have it correct, though.

Looking at the sides of the channel you can see that they get lower towards the "top". I can see how the channel is flat on the floor and that the supporting vertical struts are simply cut to the correct height for the camera's perspective. I'm at a complete loss as to how the water drops onto the wheel though.

The shadows the tower seem to be casting on the track are pretty convincing, but the reflection of the tower on the floor might indicate that you're right.

While this is pretty impressive, in the same vein I have to give Dyson (as in the vacuum cleaner guy and inventor of other things) huge respect for re-creating an Escher uphill waterfall as a garden piece, one that I would gladly display.


Better hi-res pic here: http://blog.makezine.com/archive/2009/08/james-dysons-uphill...

> Derek Phillips, the Dyson engineer who spent 12 months building the feature

To be fair - this guy should be a target of your huge respect, not Dyson who merely financed the design work.

Actually, I only have normal respect for him, as he wasn't the one who was putting up the cash to see it done, he was only doing his job. And it DID take him a whole year to pull it off.

All so cynical. Take it at face-value, pay the guy to set up huge versions, and cancel fusion research.

If anyone's having trouble picturing what this will look like from another camera angle, there was a relevant link on HN recently: http://www.npr.org/blogs/krulwich/2011/01/19/133017843/your-...

HN comments: http://news.ycombinator.com/item?id=2137152

Better link (high quality version): http://www.youtube.com/watch?v=0v2xnl6LwJE

OK, I'm stumped.

If you watch the platform near the wheel, you'll see that the water is actually going down since forward it the platform is in a lower level. I'm not an expert but, I haven't ever seen a pump making water flow that way, so maybe the water is always going down and the illusion is made by the camera position over a near flat object. Maybe you could tell the track position by watching how the water flows through corners... a pretty weird way if you watch closer...

I can't figure it out. And I'm sure this is not how he's doing it, but could you potentially use water mixed with very fine metal particles and then use a bunch of small electromagnets turned on and off quickly to pull the water/metal mixture along the path?

Even were that possible, that doesn't explain the impossible geometry.

I am unsure if you could do it that way, but the water is likely mixed with copper, judging by the color.

Probably just blue dye to make it visible to the camera

Agreed, it is likely for the camera. A lot of blue dyes for water fountains and such are copper sulfate. Dyes for tanks or fish ponds are increasingly methylene blue, food dye.

I thought I would just extend the poster's thought experiment, not imply that magnets were indeed used.

This clip is false.

Well obviously, but the question is how did he do it.

I was trying to allude to the "This sentence is false." fiasco :)

Ah, but the beauty of "This sentence is false" is the self reference, which of course breaks when you make the subject anything else. What you could do is say "this is not a waterwheel", and attempt to get at the whole "This is not a pipe" fun-ness.

Guys I know how it is made, but my english is not enough to explain it.

Give it a try

The most amazing thing about this (and other similar illusions) is that even if you know it's fake and know how it's constructed, YOUR BRAIN WILL NOT LET YOU SEE IT.

so confused.

No fair. Gotta move the camera to the left at the end a little bit so it is clear how he built it.

Maybe the structure is actually fairly flat, only a few inches off the floor at the highest point, but laid out, with shadows painted on the floor to make it look like it is vertical.

The shadow cast on the yellow water wheel is not from the pillar you think it came from.

Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact