I'm trying to achieve keystoning after placing my projector on the floor. I am using a doublet fresnel, so keystoning using the collector fresnel is out of the question.

here's my idea:

the following is having a overhead projector style setup-

[mirror]----light path-----[wall]
|
[lens]
|
light path
|
|
[object]
[light source]

1. if we leave the system such that the light hits the mirror at 45 degrees (mirror is at 45 degrees to principle plane), no keystoning will be required, and the projection will be perpendicular to the principle plane.

2. if we were to tilt just the mirror such that the light hits it at 50 degrees (mirror is at 40 degrees to principle plane), keystoning is required, and the projection will be raised up the wall.

3. if we were to leave the mirror and wall as they are, but tilt the rest of the system (lens, object, light source) such that the light hits the mirror at 40 degrees (mirror is at 50 degrees to principle plane), keystoning required, and the projection will be lowered on the wall, by the same amount that it was raised in case 2.

Am I correct about case 3? especially about the last part ("by the same amount"). also, would the horizontal distortion be the same, just opposite (ie. same amount of keystoning required)?

If I'm correct, then my idea is to use case 3, then tilt the whole system (light source, object, lens, mirror) by 5 degrees to raise the projection. would this work?

Take a look at puni5her's first PJ, the projection lens is at an angle, his uncomplete second uses a method similar to what you have stated. Either way, the field fresnel needs to be tilted relative to the collimator to achieve keystoning so I don't think you'll be able to achieve any with the 2 fresnels stuck together. Best you can hope for is a bit of lens-shift.

wouldn't the projection lens at an angle result in the image (or one side of the image) not being as focused?

anyway I just realized a flaw in my thinking.. if it's lowered by X distance, then I tilt the whole system to raise it by X distance.. I'm back to square one

additional post:

Refer to this thread: http://www.lumenlab.com/forums/index.php?showtopic=2751 [Keystone correction with mirrors, How is this done?]

brainchild, damage and pjgibbs seem sure that it's possible, but I'm having trouble deciphering what they're trying to say.

I still stand by what I said in that thread.

DJ

I'm not surprised, as at this moment I agree with you

However, I just did a ray diagram... it may be a little inaccurate but here's what I found:

note- all positions used are position of the center of the image.

I placed the mirror at 50 degrees to the principal plane. The resulting image was focused at an angle of about 12 degrees to the principal plane. (just for reference: a mirror at 45 degrees to the p.p. would result in an image parallel to the p.p.)

The change in perpendicular distance between the original (mirror at 45 degrees) to the new (when mirror is at 50) is about 9 mm (of course, this number depends on the scale and other values I used)

If I were to then rotate the whole system by 12 degrees to have the image parallel to the original principal plane, the change in perpendicular distance between the second position (at 50 degrees) and the final position (entire system rotated by 12 deg) is 11.5mm! meaning a net movement of 2.5mm while keeping the image parallel to the p.p.

hmm silly me, I could have simply measured the distance between the original (at 45 deg) and the final (rotate whole system) positions

One mistake I may have made was when rotating: I didn't actually rotate the whole system, but just rotated the 2nd position about the point where that particular ray touches the mirror. I figured this is ok because you can always rotate the whole system about that point and should get the same result. If anything, this will result in the "height" of the system changing, which may defeat the purpose (depending on how much the change is), but I may be wrong .

edit: Turns out I am wrong. My findings of the center of projection changing appears to be due to the height changing. Have removed my next two posts to avoid confusion.
A rotation about the mirror was what I should have done (as opposed to about the bulb), and when I did so the center of projection was approximately back at where I started.

<post removed due to terrible mistake I made>

They are all fine drawings but take note of the image plane. To get a square image the screen will need to be parallel with the image plane. If the image plain is not parallel with the screen, keystone distortion will be present.

DJ

<post removed due to terrible mistake I made>

I don't believe members should plan on using the mirror as a way to correct for keystoning if the mirror sits behind the triplet. In this position the mirror is used to reflect the image into the CENTER of the triplet. If the mirror is rotated, then the image may easily miss the triplet. If the mirror is rotated even slightly, the image no longer hits the center of the triplet, and the image becomes blurred. To correct, members would then also have to rotate their triplet to be in line with the projected light... which would be a time consuming feat.

Click to view attachment

Fulcrum

Instead, the best location would be with the FS mirror in front of the triplet, much like an overhead projector. Here the image could be corrected for keystoning using the mirror. However, as the image is moved up or down on the screen, portions of it picture would be out of focus (I remember this from the lecture halls in college). The triplet would then need to be moved up or down to get the center of the image back in focus, but the edges would still always be out of focus. In order for the image to be perfectly focused everywhere, the screen would have to be perpendicular to the line of light. But if this happens, keystoning is removed.

Click to view attachment

Fulcrum

Sorry about the last picture, as the wording is blurred. I tried to correct, but could not. Here is what the text says:

Green Lettering:

Key Stoning Occurs:
> Top Edge Shrinks
> Bottom Edge Grows
> Triplet is Moved Vertically
To Refocus Center of Picture
> However Top & Bottom Edges
Remain Unfocused
- Top is Too Close
- Bottom is Too Far Away


Orange Lettering:

Key Stoning Occurs:
> Top Edge Grows
> Bottom Edge Shrinks
> Triplet is Moved Vertically
To Refocus Center of Picture
> However Top & Bottom Edges
Remain Unfocused
- Top is Too Far Away
- Bottom is Too Close

yes, the mirror-after-triplet setup is what i'm trying to do here.

I stumbled upon this pj plog today. This pj was built like an overhead projector, and to help remove keystoning, the screen was tilted.

Cheers!

Fulcrum

thanks! that's a good idea.. which I had already thought of, but unfortunately I do not have the luxury of doing that as I'm using the wall as my screen. but it's good to know that other people actually did that with good results, maybe I'll work in a screen too.