Due to the 1.9 native aspect and difference in zoom when using a 1.33 lens, the brightness change is pretty much a wash between stretch mode and 2.35 mode.
I had a UH480 lens but sold it.
Here's what I saw, aside from the neutral brightness change:
1) Significantly increased on/off contrast, due to the difference in zoom reducing the aperture.
2) Significantly decreased ANSI contrast, particularly when using a closer throw and more of the lens, but this was improved when using a longer throw.
3) Very slightly reduced pixel sharpness. This could be improved by tuning the color registration with the lens in place (at the expense of color registration with the lens removed). I'm talking about a fraction of a pixel, not something you could see at normal viewing distances, but you could see it when you walk up to the screen.
4) Significantly reduced effectiveness of upscaling sharpening, when used with the internal stretch. This disappeared when I switched the JRiver/MadVR as my source and used it for vertical stretch. This also improved the 2.35 scaling mode without a lens. This was all with 1920x1080 output. I don't have a 4k video card.
5) Typical a-lens geometric distortions and other headaches.
So, assuming you are using a good external source and scaler like JRiver/MadVR, trading ANSI contrast for on/off contrast was the biggest difference I saw. In my judgement, it wasn't worth the a-lens cost or headaches.
A 1.25 lens would give a significant brightness boost, but good ones are really expensive, and the brightness boost would be only half of what you get with a 1.78 native aspect projector.
I'm curious how one of the variable aspect Prismasonics would work out when set to 1.25, but not curious enough to buy one. It might be a good combination for a long throw, with the projector in the back of the room.