Clarify texture mapping in Texel glossary

[KirtiRamchandani]

Problem

The Texel glossary entry describes a mismatch between a "source texel grid" and a "destination pixel grid" without first explaining the texture-coordinate step that connects a 2D texture image to a 3D surface.

Root cause

The paragraph compresses texture mapping, sampling, and filtering into one sentence, which makes the grid terminology harder to follow.

Solution

Clarify that texture mapping assigns positions in a 2D texture image to points on a polygon, then describe filtering in terms of sampling and the display pixel grid.

Tests run

• git diff --check
• corepack npm exec --yes -- prettier@3.8.3 --check files/en-us/glossary/texel/index.md
• corepack npm exec --yes --package markdownlint-cli2@0.22.1 --package markdownlint-rule-search-replace@1.2.0 -- markdownlint-cli2 files/en-us/glossary/texel/index.md

Fixes #44290

[jdejonge206]

I don't think the committed change is really a very accurate representation of the subject matter in which it's attempting to elucidate, what is the point of removing most of the information that explains why and how it does what it does? You're not really 'assigning positions in a 2D texture image’ a 2-D texture image is something that is self-referential and that in order for its completed formed to have visually recognizable coherency it must already have self assigned positions of pixels. You're assigning faces of a polygon to a fixed grid in 2-D space, that is the whole process of unwrapping and UV mapping.

'Texture mapping is a stage of the process of rendering a 3D image for display’ confuses what the process is; object unwrapping and texture mapping, are stages of creating finished models whether you want to create a single static image or use them in real time or in visual effects is process-dependent.

'The texture image's texel grid and the output display's pixel grid usually do not line up one-to-one.’ The texture image doesn't itself possess a texel grid the object does, I think you may have reversed these things in your head? There's nothing about UV mapping in and of itself that makes impossible a one-to-one lineup of pixels to texels, texture filtering is more an aspect of the difficulties of rendering in real time such as Anisotropic filtering which is used to visually clarify the rendering of textures on objects at sharp angle, perspectives which alter things like tangent bases, and object normal. These things have drastically less application in purely rendering images like in ArchVis.

'The final output of the rendering process is a flattened 2D projection of the 3D model, where the texture has been 'wrapped' around the model.’ It's not a 2-D projection of the model it's a projection onto the model.

Just saying if the object is actually clarification, these things are important.

[KirtiRamchandani]

Thanks for taking the time to lay this out. I agree the previous paragraph mixed texture mapping with final rendering language and made the mapping step sound less precise than it should.

I pushed a follow-up that reframes this around UV/model-surface coordinates, texel sampling, and filtering for magnification/minification. I also removed the flattened-projection sentence so the glossary stays focused on what a texel is and how texture mapping uses it.