Thank you very much for this package! And please forgive any misunderstandings I have; I'm not terribly familiar with the package internals.
I've found that for certain types of 3D diagrams, a perspective projection rather than an orthographic one sometimes conveys the idea more clearly. I've found this is particularly the case when dealing with round things, e.g.
#import "@preview/cetz:0.4.2": canvas, draw
#set page(width: auto, height: auto, margin: 0.6pt)
#let sph-to-cart(r, theta, phi) = (
x: r * calc.sin(theta) * calc.cos(phi),
y: r * calc.sin(theta) * calc.sin(phi),
z: r * calc.cos(theta)
)
#set text(size: 20pt)
#canvas(length: 6cm, {
import draw: *
circle((0,0,0), radius:0.0125, fill: black)
ortho(x: -60deg, y:0deg, z:-100deg, {
line((0,0,0), (1,0,0), stroke: red, mark: (end: "stealth"))
line((0,0,0), (0,1,0), stroke: green, mark: (end: "stealth"))
line((0,0,0), (0,0,1), stroke: blue, mark: (end: "stealth"))
content((1.1,0,0), $x$)
content((0,1.1,0), $y$)
content((0, 0, 1.1), $z$)
for phi in range(0, 360, step: 45) {
let points = ()
for theta in range(0, 181, step: 1) {
points.push(sph-to-cart(1, theta * calc.pi / 180, phi * calc.pi / 180))
}
line(..points, stroke: (paint: gray, thickness: 0.5pt))
}
for theta in (30, 60, 90, 120, 150) {
let points = ()
for phi in range(0, 361, step: 1) {
points.push(sph-to-cart(1, theta * calc.pi / 180, phi * calc.pi / 180))
}
line(..points, stroke: (paint: gray, thickness: 0.5pt))
}
})
})My understanding is that Cetz is already using homogeneous coordinates for everything internally, and indeed it's already possible to override the 4x4 view matrix. I think all that would be necessary to enable a true perspective projection instead of orthographic ones would be to add an option to do the final (x/w, y/w, z/w) perspective division step.
Thank you very much for this package! And please forgive any misunderstandings I have; I'm not terribly familiar with the package internals.
I've found that for certain types of 3D diagrams, a perspective projection rather than an orthographic one sometimes conveys the idea more clearly. I've found this is particularly the case when dealing with round things, e.g.
My understanding is that Cetz is already using homogeneous coordinates for everything internally, and indeed it's already possible to override the 4x4 view matrix. I think all that would be necessary to enable a true perspective projection instead of orthographic ones would be to add an option to do the final
(x/w, y/w, z/w)perspective division step.