Originally Posted by humbland
Why do you use the A-Lens at all?
From my reading, the Lumagen can do non linear stretch to any aspect ratio you want. Why not just use the Lumagen to get you there and take the A-lens out of the light path?
I think you misunderstand what a lens does. No amount of electronic processing (linear or not) can change the shape of the light the projector puts out. A lens is the only way to change the shape of the projected light from 1.78:1 to 2.37:1. An anamorphic lens is about producing a 2.37:1 projector output using the full resolution/light output of the 1.78:1 projector's panel for displaying wider than 1.78:1 images on a scope screen.
It's not about "filling" the screen or anything of the sort, and it's certainly not the same as doing a non-linear stretch and adding lots of distortion.
But the projector still outputs 1.78:1, without a lens you "waste" a significant (IMO) portion of your projector's output.
Originally Posted by mark haflich
Now as to the question of A lenses and various aspect ratios, I don't use a lens, nor do I zoom. I only have room for an 8ft wide, 1.78 aspect ratio screen. When I watch HD TV etc, I watch on full screen 1.78. When I watch a higher aspect ratio subject, such as a wide screen movie, I gat black bars which I cover up with automated top and bottom masking. After a movie, people ask to see HD TV and when I open the masking for the full light up, they ooh and ah about the size increase...
That's actually precisely the reason why I do
use a lens (a Prismasonic HD5000R in my case). I don't want epic scope films to be a "let down" compared to HDTV, nor do I want to have to fire up HDTV to show off my system in it's most impressive configuration.
Using an A lens, you get to use all the pixels and the full illumination of the projector (with certain caveats) because you electronically stretch the image vertically to light all the pixels. This is a nasty scale to do. All scales except say 2 to one and numbers like that as distinguished from odd number multipliers are nasty and severely scree with the vertical pixels.
There's nothing fundamental about scaling that causes this. There's nothing in sampling theorem that ever says anything about having to scale by exact multiples. The only constraint on resampling is that the original sampling was done "correctly", as in the signal is sampled at at least twice the it's bandwidth.
It seems lots of people look at test patterns, which are not sampled at all, and have a bandwidth far beyond not just half the sampling frequency but many multiples of the sampling frequency (a typical line or 1 by 1 pixel grid is essentially a square wave which has significant harmonics 7 times the fundamental frequency or one cycle per pair of pixels), run them through a scaler, look at the artifacts and claim it as proof that that scaling is bad or damaging. They do this without understanding, or caring that those signals cannot be resampled accurately period, they do not conform to the preconditions/limits that all scaling/resampling requires.
Motion picture content (film, digital, or CG sourced) does not have these same issues, these sources are all sampled data, with appropriate sampling/bandwidth considerations and can be accurately and completely resampled from any sample rate (resolution) to any other (higher) sample rate accurately and without loss or artifact.
Now whether or not all video processors implement accurate resampling algorithms or do other things that may or may not cause issues, well that's another issue.
My point though, the Cliff's notes, is that just because you see artifacts on a single pixel level test pattern doesn't mean those same artifacts are present in real content, you can't look at the results of resampling on a non-band-limited test pattern and make conclusions about the results of resampling on appropriately band-limited data.
The lens deals with the horizontal stretch by doing it optically and what comes into ply here is the optical qualitty of the lens. The 480 is OK but that's it. There are worse and there are many many much better.
My understanding is the biggest issue with the UH480, my HD5000 and all prismatic lenses is that they have a fixed focal length. They have a singular sweet spot and cannot be adjusted. Specifically astigmatism (focus uniformity in both horizontal and vertical directions) cannot be adjusted. Cylindrical lenses like the ISCO have adjustable "focus" (astigmatism) so you can dial them into perfection for whatever throw you have.
On top of that there are some geometric issues and overall quality, but I believe the astigmatism adjustment is the biggest difference.
You end up losing light from the max capoable because anamorphic sets up need long throws for a variety of reasons such as the need to keep the image coming out of the projector lens small and to minimize pin cushioning which all anamorphics cause.
Ah, but you gain contrast
But to the OP, if you want a CIH setup, that's a killer deal on a very nice lens (no not the best, but very good). If you want a CIH setup with a lens, I would grab the lens no question.
If you don't know if you want a CIH setup, it might be worth the cost to experiment, you can always sell it.
Otherwise, the Darblet is the best bang for that buck upgrade you can get, I wouldn't get the lens just because it's cheap unless you plan to resell it for a profit.