matching one transform3d's rotation to another while constrained to a maximum angular speed

[maiden mirage]

i have two transform3ds, they are identical except for rotation. i want to rotate the first one to match the second one while not rotating faster than some arbitrary scalar value.

so for example if A is rotated to (0,1,0) and B is (1,0,0) and the scalar is 0.1 then A should rotate towards B by 0.1 radians over one second.

how can i do this?

alternatively i'm also open to a different way of doing this that doesn't involve transform3ds. i'm using them because the rotation function i've implemented modifies the transform with quaternions and then returns a new transform

but to be honest i have no intuition for math more complex than basic vector operations

[manic void]

I think something like this would work:```php

if your forward direction is something other than z, change it

angular difference between the two basi (this doesn't take into

account rotation around the forward axis)

var angle_difference := acos(t0.basis.z.dot(t1.basis.z))
var tween := create_tween()
tween.tween_method(
func(weight: float): transform.basis = t0.slerp(t1, weight),
0.0,
1.0,
angle_difference / radians_per_second,
).set_ease(Tween.EASE_IN_OUT)

If it looks wrong, you probably have to tweak the angle difference calculation to take into account roll as well

[maiden mirage]

thanks! ended up scrapping the tween cos it gave problems, probably related to that thing whit the roll. anyhow i did this:

var angle_difference := acos(transform.basis.z.dot(desired_transform.basis.z))
  quaternion = quaternion.slerp(Quaternion(desired_transform.basis.orthonormalized()), angle_difference / turn_speed)

i'd like to find a way to do this without slerp cos i never know if linear interpolation functions are framerate independent but this works for now

[manic void]

If you've got the time, I highly recommend this talk by Freya Holmer about this very topic https://youtube.com/watch?v=LSNQuFEDOyQ

[maiden mirage]

that's where i've learned it lmao

[manic void]

You should be able to apply the exp(-k*delta) to slerp as well