Red Beard: The Well (a technical question)

Just a quick question to those in the know about film techniques. How was this shot, and the camera movement that precedes it, done?

This is at the end of the scene where the women shout into the well to bring back Chobo. Approximately 2:58:10 to 2:58:20.

While I am usually able to tell how this kind of “mirror shots” are constructed without revealing the camera, I cannot for the life of me think how it was done here. Is the camera somehow hidden in the woodwork above the well? But if so, the whole woodwork must somehow have been lifted for the camera movement?

The way the scene ends is marvellous, with that tear falling and making ripples on the well water’s surface. A brilliant ending to a brilliant scene.

Curiously enough, it is also one of the very few scenes (if not the only one) where I as the person watching the movie become aware of Red Beard as a movie. Elsewhere, the movie seems very careful not to pull out tricks like this. Perhaps my follow-up question should therefore be — why here? Could it be because the movie is almost finished by now (only 7 or so more minutes to go), and it is therefore a calculated trick to start pulling us out of the movie?

I looked at this many times forward and backwards. Basically, I have no idea how it was done. But, that didnt stop me from make a rather complex diagram that makes little sense. I did my best, I suck at drawing, and I dont know how to put anything in words…..

Giving the production level on this movie, I dont think we should looking at the well as a real well, but rather a structure above ground, built for this one shot. It’s a rather important scene, and it marks a closing with a fade out, so I dont think some large efforts done for this, is asking for too much.

To hide the camera, a series of tilts, pulls, crane ups and downs would be needed. Most important the camera is needed to be below the subjects, hence the above ground well.

1. Starting with a close up of the girl-the crane height is set low, but the boom arm is extended high and at a extreme angle. (the weights opposite of the camera on the arm, would be touching ground).

This would put the camera up high, but with a angle matching up with the shoot, by using a slight zoom.

2. The camera tilts down to it hits 0*, as the dolly pulls out a bit, along with a crane up. This would put the boom arm level and the camera still at 0*.

3. With the camera at 0* tilt and the boom arm level. A crane down, allows for a tracking shot of sorts down the wall.

4. Once point is crossed, the camera tilts down, as the boom arm tilts down(weights up in the air) and the crane ups, as the dolly pushes in back to original location.

On that, I present you “Hell in da Well”

Using some slight zooms, you can maintain perceived distance from the wall, making the camera appear to be riding it. A bit of lighting should hide any jerks in movement, and the fact a cut out is in the fake well.

The mirror part is just a guess, I would do that just for ease, only having a inch or two of clear water above it.

This is all a bit complex, but nothing impossible and nothing that couldnt be done in a few tries.

I’ll properly think of something better later on, and I also have a editing idea that might work.

Anyways, maybe this will at least get some ideas going and someone might be able to offer a better solution.

I find it hard to think it out, without having something physical to work with and test, but this more or less how I would do it.

I can’t resist screwing around with MS Paint once it’s open, so here’s my Mifune pic.

$350,000 buys the original –it’s 1 of 1!!!!! 😉

Awesome!!!!!

I was hoping someone else had something to offer.

Issues like the camera size, focusing, most importantly hiding the camera cant be solved within the confines of a well for me.

My editing solution I mentioned, doesnt appear to work after I work it out, if it did, it would require so much complexity, I could see a whole day being spent on just getting it to work, and another in editing out hiccups and matching footage.

I’m not going to detail it, as it has a flaw in being unable to explain where the camera is, when the tear drops anyways.

The only new thing I discovered is a better idea of the angle of the final shot. It’s extremity would support my crane idea, as well go with common practice of shooting a mirror scene-a camera at extreme angle, outside the lens field of view.

There also appears to be a twisting effect as the camera tilts, and tracks down the well. This I think matches up to what happens when the you have height, tilt, crane down/up and pull all happening at once. It might be used to hide a hole in the wall for the camera too.

I finding this rather interesting, so somebody else offer something, I hate being the one guy answering the camera questions.

Thanks for “Hell in da Well”, as well as the “Hell in da MS Paint” that follows! 😛

It took me a few read-throughs to understand the technicalities of your drawing (the first one!), but I now think that I get the set-up. It seems like a plausible explanation to me, although also I was wondering whether it accounts for the angle of the camera’s final position. But I think that it could.

In any case, I had not realised that you could actually build a well above-ground for these purposes. You live, you learn…

I’ll keep thinking about this.

The drawing sucks-both of them, but the first is certainly poor, and hard to understand, plus I left out a few things I just realized out.

I might try another shot tonight, maybe with a step by step.

The best I can tell, everything I mention would sync up to whats on screen.

I know the above ground well, isn’t the most attractive idea, but it’s the only one that I can find, to explain how the camera could operate in such tight spaces. Cameras were big and heavy back then, and there was no such thing as wireless or remote wired focusing and tilting, so you need a guy or two to be with the camera at all times as well.

Let me just state again-I have no idea how this was shot. 😥

But here is my best idea, the same as the last, just with more confusing detail.

I’m not happy with it, but it appears to work out fairly decent, even with some possible angling issues-that I’ll guess could be solved with a push of the dolly or some zooming techniques or I dont know-that being the more likely.

There is just too many unknown factors to make it all 100%, this reverse engineering is really hard.

The drawing is poor, and it hard to show the motions, just keep in mind there are 3 camera position in the picture.

There is nothing more I can offer.

I give you “Hell’s hell in da well” -damn I’m clever 🙄