The Vector3 class in threejs has many prototype methods one of which is the Vector3 normalize method. Calling the normalize method of a Vector3 instance will preserve the direction of the vector, but it will reduce the euclidean distance of the vector to a length of one.
A Vector with a euclidean distance of one is often referred to as a unit vector, and what is nice about this kind of vector is that it can quickly be scaled up by just simply multiplying the values of the normalized vector by a desired magnitude that is any value other than one to result in any vector that is along a given line that is the direction of the vector.
Vectors are often described as being a unit of direction, and magnitude, the direction can be thought of as what the normalized vector is in terms of numbers between 0 and 1 for x, y, and z. This direction can then be raised, or lowered actually towards zero, by a magnitude to get any point in space. So then in this post I think I will be going over some basic examples of the normalize method, and while I am at it also end up writing about a few other topics that are closely related to the normalize method.
There is checking out my main post on the Vector3 class where I am going over the Vector3 class in general. Normalizing a Vector is a major part of becoming proficient with the Vector3 class, but there is a great deal more to it when it comes to the basics of Vector3, as well as other various methods of the class.
The thing to keep in mind here is that the normalize method will just set the length of a vector to one, while preserving the direction of the Vector, but that is it. What if I want to set direction of a Vector by a set of given angles in terms of radians or degrees for example? I will be going over some additional methods other than just the normalize method here, but there is a great deal more to be aware of that can be used with the normalize method, and I will not be getting to all of it here.
In this basic example I am creating an instance of THREE.Vector3 that is not normalized, and then just calling the normalized method of the Vector3 instance to get a normalized vector. The value of the Vector before normalizing it was -20, 0, 0, and the value after normalizing it is -1, 0, 0. So in other words the direction of the Vector is preserved but the length of the vector is reduced to a magnitude of 1. Once the vector is normalized I cal call a method like multiply scalar off of the normalized vector to set any desired magnitude, or distance if you prefer while preserving the direction of the Vector.
So then the normalize method will set the length of any vector to a length of 1, and then from there the length can easily be adjusted to any desired length. Also when it comes to the subject of the length of a vector the Vector3.length method can be used to find out what the current length of any vector is. So then the normalize method combined with a method like multiply scalar can be used to set the length of a vector while the length method can be used as a way to get what that length is.
So then one use case example for all of this would be to work out one or more methods that have to do with positioning an object on the surface of a sphere. That is that I can create a method in which I can pass values that will be used to create any point in space, and then normalized that point to a vector with the same direction but with a length of one. I can then set the length of the normalized vector to the radius of the sphere, plus one half the height of the object that I want on the surface of a sphere. That basic method seems to work pretty well, and it is then just a question of making other methods that serve as an abstraction for that kind of method, such as a method where I can just give a lat and long value in terms of values between 0 and 1 for each argument a a way to position something on to a sphere. This will then also serve as a way to take some kind of system that involves positioning things on a grid and make it so that it can also be used to position the same things on a corresponding sphere surface.
This is the sort of thing that I find myself coming back to now and then when it comes to working out new systems for placing objects onto the surface of a sphere. I have a simple project example that I made a little while back in which I was able to work out a solution for doing this sort of thing but it was very different from this kind of example that I like better.
So then there is normalizing a vector to a length of one, and keeping the direction, but what if I want to change the direction while I am at it to. In other words what if I want some kind of helper function that will return a normalized vector, but I can also set the direction of that normalized vector with some angle arguments. In addition I can also set a length as a way to not return a normalized vector but a vector with an interested length, and also adjust what the starting vector is.
One way to make this kind of method would be to make use of the apply Euler method that can be used to change the direction of a vector by way of using some angles to do so. The apply Euler vector3 prototype method accepts radian values, but if I want to use degrees there is a deg to rad convince method in the math utils object. I will then just want a starting vector by which to use the apply Euler method with, and this vector should have a length greater than zero.
That will be it for now when it comes to the normalize method in the Vector3 class, but I am sure that I will come around to expand on this post at some point in the future when I have more to write about on this subject. There are many other methods in the Vector3 class that can be used with a normalized vector that I might get around to writing about sooner or later, but I need to get to working out some demos, and doing some more research first.
Never the less I think I did an okay job covering the basics of what a normalized vector is, now it is just a question of applying this to make some useful or interesting projects. Or improve some ones that I have made all ready, and I can think of a few that I would like to fix up now that I have a better understanding of this sort of thing.