#3d graphing

I fucked up. None of the previous links on live weave are going to work because I deleted them all. I’m sorry for anyone going through this in the future.

On the bright side I’ll start saving them now, and also I’m making progress. Graphs now go through each other instead of sitting on top of each other based on which was drawn first. all I did was draw each row for each function individually. It’s not really perfect but it absolutely does the trick.

Here is the latest version. It should actually work. All I have to do now is fix the second overlapping problem, get a better input system and I’ll have a prototype.

Yea so added multi graph support to my 3D graphing thing.

All I did was add an array to hold multiple functions and then a loop that draws all functions while a variable is less than the length of the array. So

The problem still being that the rendering of every part of the graph is independent of where each point on the graph is in 3D space. So even if triangle A is behind triangle B, but triangle A is drawn first, triangle A will be in front of triangle B. I don’t know how to establish a distance from the camera type deal so I can’t tell it to draw certain things first.

yea so all the online 3d graphers suck huge dick so I decided to make my own.

Game plan was this I think:

Make the 3d axis

make points plotable on said axis

create a way for a two independent variable function to be graphed and drawn on the axis

Make the 3d axis:

This was actually super fucking hard for no reason.

Yea so I used Goegebra as my template for my 3d axis. Just because (in my opinion) that the lesser of all the evils that are the 3d graphing tools online.

So it looked like there were two angles that the rotation of the space depended on.  The rotation around the z axis, and rotation around an imaginary axis that was perpendicular to the z axis and parallel to the screen. Looks like everything is dependent on the z-axis so it makes sense to start there. It also looked like the z-axis didn’t actually rotate “side ways”. As in it was always the vertical axis. Which made sense. What good does it have for your vertical axis to rotate around anyway?

So I went into desmos and got to work. I made an angle for the rotation around the z axis, A_z, and an angle for rotation around that imaginary axis i was talking about, A_p. I mentally flipped a coin and decided that the initially the x-axis should come out of the screen, and the y-axis should be parallel to the screen, perpendicular to the z-axis.

The way that I have this setup is that I have three 2D points that represent each axis. Those points can be used to create a line through (0,0) to create an axis. Each one is just a function based off of A_z and A_y but that’s all it is.

I won’t detail how I got there because it was a couple days ago that I did this and also a lot of it was guess and check as I went, but here are the properties for each axis with certain conditions for each angle.

the syntax for each point value is:

x_a = the x value for the axis point, where a is the axis

y_a = the y value for the axis point, where a is the axis

I’ll have a link to the actual desmos page so you can take a look at the math and this will all probably make a lot more sense. Anyway so here is the logic behind getting to the formulas for each point.

When A_z = 0 and A_p = 0

y-axis should be fully extended on the 2d x-axis, fully retracted on the 2d y-axis

x-axis should be fully retracted on both 2d axis

z-axis should be fully extended on the 2d y-axis (always fully retracted on the 2d x-axis)

When A_z = 0 and A_p = 90

y-axis should be fully retracted on both 2d axis

x-axis should be fully extended on the 2d x-axis, fully retracted on the 2d y-axis

z-axis should be fully extended on the 2d y-axis(always fully retracted on the 2d x-axis)

when A_z = 90 and A_p = 0

y-axis should be fully extended on the 2d x-axis, fully retracted on the 2-d y-axis

x-axis should be fully extended on the 2d y-axis, fully retracted on the 2-d x-axis

z-axis should be fully retracted on both 2d x-axis

and finally when A_z = 90 and A_p = 90

y-axis should be fully extended on the 2d y-axis, fully retracted on the 2-d x-axis

x-axis should be fully extended on the 2d x-axis, fully retracted on the 2-d y-axis

z-axis should be fully retracted on both 2d x-axis

So when you put the pieces together the formulas for each x and y value for each point for each axis are as follows:

x_z = 0

x_y = cos(A_p)

x_x = sin(A_p)

y_z = cos(A_z)

y_x = sin(A_z)cos(A_p)

y_y = sin(-A_z)sin(A_p)

COOL all done with the axis itself. All I did was assign the angles to a move-able point and boom. rotatable and stable 3-d axis.

Make points plottable on the axis

This part was similar in structure to making the axis itself. Essentially the points that I was talking about for each of the 3d axis are points in the 3d space itself. I just needed to find a formula that would take point (x,y,z) and make it point(x,y) based on the same parameters.

I’m gonna be totally honest with you, I don’t remember how I reasoned this out. I think I took the formulas above for each axis, multiplied them by the input value for the corresponding axis, then added them all together.

So for 3d points (a,b,c) I think I took:

x_abc = a*sin(A_p) + b*cos(A_p) + c*0

y_abc = a*sin(A_z)cos(A_p) + b*sin(-A_z)sin(A_p) + c*cos(A_z)

Then did some trial and error and got

actually the above is the right answer I think

So yea the point (x_abc,y_abc) is the corresponding 2d point for the 3d point (a,b,c) with the corresponding angles. Now all I have to do is

create a way for a two independent variable function to be graphed and drawn on the axis