I notice that in the overall sequence of the course, we first modeled, then unwrapped UVs, then did texturing/lighting/rendering. This makes sense as there seems to be a linear dependency structure, i.e. can't texture until UVs are unwrapped, can't unwrap until the model is done. So I'm surprised that the last step includes more model deformation, but the results still look fine. How much mesh modification is okay after UVs have been done? Does it matter that the modification in this step is by lattice modifier, rather than directly pushing verts in edit mode?
Not such an easy topic, but let me try to show to you what basically happens when deforming a UV unwrapped mesh.
(Disclaimer: this is a very incomplete answer!)
Suppose you have carefully unwrapped and textured your model, like this:
But the you decide to change the shape a bit, by moving vertices (or using a Lattice). This will happen to your texture:
It depends on your texture if that is a problem or not and how much deformation is acceptable.
great demo (as you say very incomplete...) to add to it maybe juust a bit, a way to think of it is where the mesh gets deformed the texture or uv's get deformed along with it
Or look at it this way:
Give every vertex a number. (Gets messy real soon, but this is what Blender does)
The UV-map is like a lookup to see what color each vertex has. It doesn't matter where each vertex in in 3D space, the color stays the same and all points between vertices are interpolated,
How much mesh modification is okay after UVs have been done?
ttarrosion Great question. I understand how adding the lattice in this last step can be confusing and your question cuts right to the heart of it.
One of the benefits of Modifers, like the Lattice Mod, is their non-destructive nature. Meaning that you can add modifiers and not worry that the UVs will be negatively affected; they're always intact at the "base level" of the mesh. This is why we can add the lattice and the UVs behave well as one would hope.
Does it matter that the modification in this step is by lattice modifier, rather than directly pushing verts in edit mode?
Again the non-destructive nature is the key significance in using a modifier instead of changing base mesh. Consider how cube-ish we constructed the chest throughout the course. We only had to worry about our shapes being perfectly oriented toward the absolute axes: X, Y, Z (aside from the curve of the lid). This means our mind can operate in the comfort of 90 degree angles / increments and true orthographical perspective.
As soon as we adjust the proportions to be slanted and "wonky" for the sake of style, we lose that comfortable cube-ish context. UNLESS we wonkify the shape with the modifier because we can simply turn that modifier off at any point and make changes in the cube context again, including UVs.
I hope some of that makes sense...it's a hard thing to explain textually.