Unfortunately, no, it doesn't work that way with this kind of rendering. There's a huge amount of re-sampling that goes on converting a texture on a surface into a rendered image. Even with a 1000x1000 pixel texture on a square polygon that is perfectly positioned in front of a camera set to a 1:1 aspect ratio and an output res of 1000x1000, there will be degradation in the final image compared to the source texture map.mitchino wrote:I'm not sure I get the logic, but I'm willing to be persuaded.
Just say my output resolution is 1000 x 1000 px, and I'm rendering a square plane straight on with graphics on it that fills the frame - whether my textures are 1000 x 1000 px or 10,000 x 10,000 px they will surely look the same on the final render.
If I was to use textures at 500 x 500px then they would look rough of course.
Might do an experiment and test this out.
Every ray bouncing off that textured surface involves several levels of sub-sampling, approximation, and rounding errors. Every pixel in the final image is the result of hundreds or even several thousands of those rays, with all the cumulative rounding/approximation you'd expect. There simply is no 1:1 relationship between the pixels in your image texture and the pixels on the final rendering, no matter how you set up your scene. Instead, there's a whole lot of messy mathematical approximations and assumptions about 3D space and the way light interacts with surfaces that transform your texture into a rendered image.
Even starting with a higher res source texture, it will never be perfectly reproduced at a lower res, as with any down-sampling, although starting with more info to sample from helps. There is a point of diminishing returns, though, and more than 2X is generally not going to make any noticeable difference.
Definitely do some experiments. In some situations, 1:1 will be fine, but for the kind of exacting detail you need, I think upping your texture source res will make a significant difference.