Protohaven Member Forum

CMU MHCI students need help/ideas


I am an advisor to some students in the MHCI program at CMU. They are in the final 7 weeks of their Masters degree. As part of their Capstone (read 8 month long project) they are working on a physical prototype and are looking for some construction advice.

The basic idea is that they need to project (via rear projection) onto a curved surface. A flexible projection film will be placed onto a curved, rigid surface. The challenge is that there needs to be empty space behind surface (between the back of the film surface and the projector) as to enable proper focus from the projector.

Once constructed, people will interact with the surface by pressing, touching, and moving their hand/fingers along the surface. It needs to be resilient enough to withstand that pressure.

To build this, the initial thought was to have a frame or skeleton supporting the (clear) rigid surface along its edges. The surface itself will likely be acrylic and will need to stand on its own.

Does anyone have experience with a similar project such that you could offer some guidance on construction methods or otherwise?

I can connect you with the students directly if so.

Thank you.

Hi Derek,

As I understand what you are trying to do, you may need to rethink your idea.

The premise of either a front or back projection screen is that the lense is focused at a fixed distance from the projector. If the screen were to move closer or further from the projector the lense would need to be refocused. If the screen were to be flexible and portions were to move closer and further from the lense, this would require a distance measurement to each area of the screen as it moved, and a dynamic refocusing for each small area, or ideally each pixel of the screen.

I am not one to say that anything is impossible, but it would be a major research project just to solve this problem.

Separately, I would want to know how flexible you want the screen to be. Do you want it to flex like a fun house mirror, or do you imagine that it would be even more flexible than that? If you allow participants to flex the mirror there is always a chance that they will exceed flexibility specification and kink or fracture your projection surface. This flexible surface needs to be chosen carefully and the supporting framework constructed based on the material specification.

In this case I might recommend a mechanism to flex the screen that would be calibrated to prevent over flexing, maybe with participants able to operate the mechanism.

If you have further questions, please feel free to contact me directly at

Best, Phil

From the drawing, it looks like there’s a clear RIGID piece behind the projection screen, so I’m going to assume Phil’s legitimate concerns are answered with that. I guess it will be thin enough to flex into place, and rigid enough to withstand touches once it’s held there. You may have trouble with cracking after a while if you can’t heat it enough to let the strain relax in the curved position.

I agree with Phil that the optics may be a challenge. Even a singly curved surface (a cylinder) might be difficult to focus on with a projector designed to focus on a flat surface. The S-shape of your drawing will be even more of a problem. If the projector is far enough away, you may be OK.

For fabrication, I’d suggest cutting out plywood in the curved shape you want. You can put a groove in it with a router. An attachment that keeps the bit a fixed distance from an edge could be used to follow a fairly arbitrary curve.

Chris M

Phil, Chris — thanks for the input. I’ve passed it along to the team.

To address the specific concern about the focusing across surfaces of different focal distances, there is a solution in mind for that. There is software, which you can use to warp the image being projected, such that once it arrives on the membrane, it will be un-warped. I don’t know the name offhand.


There are laser projectors now that do not have a focal length, so if the image distortion is corrected pre-display, there aren’t any optical inhibitors to this project. A traditional projector would certainly cause the display to have large portions be out-of-focus regardless of distortion correction.

After bending the rigid plastic to the desired shape (using heating elements spanning the width of the piece and using a wooden form to guarantee the shape), creating a plywood or MDF frame with a groove routed out as suggested by @csmullin sounds like a great way to present the surface to the user.