...should only be an issue of using a glass element that serves the opposite purpose of a teleconverter. Right? Does something like this exist? If not, why?
I'm not sure of your question. A tele-converter takes a central portion of the field of view of a lens and magnifies it 2X or 3X to fill the original 'frame'. I have a problem in visualizing the reverse of this process. If it means taking an image larger than the field of view of the lens and compressing it to fill the 'frame', it's an optical impossibility. Additionally, taking the basic field of view 'frame' and compressing it is trivial.
The Golden Rule of Media Editing: You can't add quality or data that's not there. Anything that tried to be the opposite of a teleconverter would have to go outside the lens, causing no end of problems. It's most likely impossible and largely useless anyway. There are extensive benefits to the DX sensor size and it's easier to just get a wide-angle lens and call it good.
Let me see if i can describe this in more detail: image from: http://www.sphoto.com/techinfo/dslrvsfilm.htm Referencing the image above, it is obvious that the crop factor of the canon 10D creates a smaller field of view--effectively a larger zoom--than the standard 35mm SLR does when using the same lens. Now, a teleconverter spreads the image circle so that the field of view captured by the camera covers a smaller field of view of the lens, once again, effectively making the lens a higher-power zoom. Due to the spreading, we run in to problems like chromatic aberation. Now, most lenses are designed for use with 35mm film SLRs. That means that the image circle must be circumscribed on the film. The nikon DX lenses are lenses designed to have the image circle be circumscribed on the digital sensor, rather than on 35mm film. The image projected by the DX lens series is significantly smaller than that produced by a lens designed for a 35mm SLR. Considering that 35mm film is significantly larger than most digital sensors as illustrated above, one could use a biconvex glass element between the lens and the camera (same placement as a teleconverter, opposite purpose) to shrink the image circle on the focal plane to the size of a digital sensor. This would make the lens function on a digital SLR at its true focal length--regardless of crop factor of the DSLR. I hope that makes more sense.
OK. I see where you're going. Thanks for the clarificatiion. As an aside; increased chromatic abberation would probably be due to poor correction in the additional lenses. But to get back to your main premise -- Teleconverters existed mainly in the pre-zoom era as a means of keeping the photog's lens collection within reasonable bounds both physically and dollar-wise. Zooms made teleconverters far less useful, if not yet completely obsolete. I suspect that today folks use them only to keep from popping for really long-focus glass, which remains expensive. Teleconverters do result in a degradation of the image, but it's mainly due to the low-cost construction of the optics in them. A modern high quality unit would be interesting to use. Trouble is, it would be expensive. A good zoom might actually be cheaper [economy of scale]and would probably have greater aperture. The reverse of a teleconverter, as you describe it, would essentially work by reducing the effective fl of the main lens. Instead of a zoom working at an effective fl range of, say, 50 to 150, it would become a 33 to 100. While this would be theoretically possible, I don't see much of an advantage to doing it. Zooms for today's digitals come in a wide variety of surprisingly large ranges. Really wide wide angle lenses are available without having to mortgage the house. There's one more point, and it's a real doozy; the position of the converter lens' rear element. Teleconverters decrease the angle of ray convergence. The 'stronger' they are, the farther from the film plane they must be positioned. No sweat. Makes the rig look cool! Kinda like a 'Hummer' effect. A reverse teleconverter, though, increases the convergence angle and will have to be positioned closer to the film plane. For a RF rig, this might work. However, in SLRs, there's the geometry of the front edge of the mirror's flip-up arc to take into account. Its clearance is usually close to minimal to begin with. I would suspect that a reverse teleconverter of any significant strength would result in an impossible geometry.
And that would be exactly the point. To reduce the effective focal length of the lens. Now that is a good point. I think it would come down to creating some compound optics to get higher "distortion" in the same space. It would be an interesting project for sure.
You might well gain a stop or two. The problem still remains in fitting the reverse teleconverter between the rear element of the existing lens and the film plane. Incidentally, amateur astronomers are very familiar with the teleconverter, but under another name. They call it a Barlow lens. It's usually a double concave cemented achromat. The value of it is twofold; increased magnification, or the ability to use an eyepiece of longer fl and consequently better eye relief. If you've ever worked with a 6mm orthoscopic, you'll know what I mean.
If one existed, I doubt I would use it. It's adding another peice of glass that the image has to go through. It is functioning at it's true focal length. Using a smaller sensor vs. one that matches the size of a 35mm frame is no different than using a 35mm camera vs. a medium format one. It's just a matter of getting used to the difference in numbers (80-90mm is normal for MF, and 50mm is wide angle). It does suck that I can't use my 50mm as I'm used to using a 50mm, but not so much that I'd want to use an adapter. It is an interesting idea, though.
If it was possible to increase lens speed a stop or two (besides creating a larger aperture), I would think Canon would pay you a few million for your idea. I'd love a f/2.8 zoom lens that was actually f/1.4, or a f/1.8 prime that was actually f/1.
KS; In this case, the gain of a stop or two is not a true gain, but simply a re-capture and concentration of the light lost through shearing off the outside of the image. If it were a true gain, it would have perpetual motion machines beat all hollow and I'd be getting a first class ticket to Japan to cash in on it, big time! Unfortunately, simple ray tracing will show the geometrical problem I've noted to be all but insurmountable for SLR rigs. It would be barely possible for a rangefinder or viewfinder design, but the power factor would not be great. Extra lens elements are not necessarily a problem. Just check the number of elements involved in a modern zoom USM vs., say, a Zeiss Tessar or an old Schneider Xenon. Modern coatings pretty much eliminate the old horrors of internal and external reflection losses and contrast degradation.
