Tweet
Struggled to think of a good title for this, but i'm sure many people who work in rendering for events and projection mapping have dealt with this.
We have a curved LED screen - there's an ideal viewing position that will be indicated on the floor, and from that position, we want it to be perspectively correct, so it feels like you're looking through a window.
The typical workflow for this is to set up a camera from the ideal viewing position, have a guide for the screen/projection surface in the viewport, render from that camera - then in a second scene/post file, project from that camera onto the screen surface and render to texture. All well and good, the workflow is nice and simple.
The rough setup with the distorted image applied as a texture looks like this:
And the final screen output after render to texture and upscaling to native resolution, with the distortion baked in:
My question is - Is there some way to set up a vray lens file that allows us to render directly to the distorted version? We're losing a ton of resolution in the center of it during the distortion phase while the sides are staying sharp, and if we could set up a camera to render native 1:1 pixels straight to the LED screen, we'd save a ton on over-rendering for the sake of a consistent level of sharpness. We've worked with screens way more complex than this in the past too, this has been something i've been wondering for a while but finally had the time to start looking into.
It feels like the lens analysis tool should be able to do this but from what I can see it's going in the wrong direction and wouldnt be able to produce something accurately using the photography grid system.
I am imagining something a bit like what I've got set up - where I can project a shader onto the geometry and do my render to texture, and then vray can load that output into the camera lens distortion file and my raw renders come out of the box pre-distorted.
We have a curved LED screen - there's an ideal viewing position that will be indicated on the floor, and from that position, we want it to be perspectively correct, so it feels like you're looking through a window.
The typical workflow for this is to set up a camera from the ideal viewing position, have a guide for the screen/projection surface in the viewport, render from that camera - then in a second scene/post file, project from that camera onto the screen surface and render to texture. All well and good, the workflow is nice and simple.
The rough setup with the distorted image applied as a texture looks like this:
And the final screen output after render to texture and upscaling to native resolution, with the distortion baked in:
My question is - Is there some way to set up a vray lens file that allows us to render directly to the distorted version? We're losing a ton of resolution in the center of it during the distortion phase while the sides are staying sharp, and if we could set up a camera to render native 1:1 pixels straight to the LED screen, we'd save a ton on over-rendering for the sake of a consistent level of sharpness. We've worked with screens way more complex than this in the past too, this has been something i've been wondering for a while but finally had the time to start looking into.
It feels like the lens analysis tool should be able to do this but from what I can see it's going in the wrong direction and wouldnt be able to produce something accurately using the photography grid system.
I am imagining something a bit like what I've got set up - where I can project a shader onto the geometry and do my render to texture, and then vray can load that output into the camera lens distortion file and my raw renders come out of the box pre-distorted.
