xyz


(klepow) #1

In most applications, and in the real world, x is width y is hight and z is depth. In Blender z is hight and y is depth. x is width in both. Is this strictly a bug in the Mac version? Is there any way to change this ether in Blender or in reality so that they match?


(theeth) #2

I don’t know where you saw your real life reference, but in math and science, Z is up and X, Y are parallel to the ground.

Martin


(Enzoblue) #3

hehe

[Edit] Sorry :slight_smile:


(S68) #4

God used a spherical reference when creating the word, not a Cartesian, so real world has a Rho, Theta, and Phi coordinates.

For Blender, z up matches common practice, as theeth said.

On the other hand, also for Blender x is width y is height and Z is depth, this applies when you are using alpha (and se Ztransp :o where Z is the z axis) and when you do DoF with the Zblur plugin, where blurresness is dictated by the distance from the camera, which again is calle Z.

In any case, references are a convention, so where’s the problem?

Stefano


(klepow) #5

Are you sure about that? If you have a two dimensional image, you have a X axis and a Y axis. You do not have a Z axis and the dimension that you don’t have is depth.


(Enzoblue) #6

yeah… but you’re using two 2 dimensional images and putting them perpendicular to fake 3d. The reference in Blender is the ground plane. Above and below the groundplane is the Z axis…

Depth is the Z axis on a computer if you lay the monitor flat on the floor. Depth can’t be constrained to an axis though, it can be any axis you want depending on which way your facing. I can look at a computer sideways and say it has no X axis if I wanted.

A computer screen only has two dimensions, so we had to do away with one. We chose Z arbitrarily because we wanted it to be the same as looking -->down<-- on a map or a piece of paper because traditionally you put a map or a piece of paper on a desk so you can write on it. If you think about it, computer screens should be on X and Z axis because it’s in front of you.

Ok, now in 3d modelling we have to make it so the X is left to right and the Z is up and down as you look at the screen. Because the screen is in front of us and we are pretending to see it as if we were standing on the ground and looking forward. We’re not looking -->down<-- as if on a map.

Alles Klar?


(Alltaken) #7

i really couldn’t care what axis is what because it doesn’t matter.

but to point it out.

x width
y height
z depth

is the standard that i ahev been taught in maths and science at school it is also a recognised format for data display (depends on what is being displayed)

x width
y depth
z height

is another format which after doing a google search i found is rather more popular than the first with mathematical formula.

http://math.ucsd.edu/~gptesler/10c/plotxy2.jpg
http://www.sitraka.com/software/images/prodxrt3dpic2.jpg
http://dlmf.nist.gov/about/publications/nistir6297/img15.gif

however i did find a few graphs using the first convention (but most were of a poor quality)


(S68) #8

Mmmm…

interesting.

I have always been taugth the second, x,y,z with z vertical, never seen y vertical, unless you are duing some 2D but, in this case, the fact tat y is vertica is merely a special effect.

In any case I suggest that new Blender version should hyperbolical coordinates.

Stefano


(hairu) #9

In all my calculus books xy have been widht/depth and z is height. And i’m not sure where the real world statement comes from because it makes no sense to me. The coordinates in the real world are whatever you call them.

But it’s really easy to convert between any cartesian coordinate system by just doing a simple rotation before exporting, so it really doesn’t matter to me.

( oddly enough though i think that in OpenGL the default is Z for depth )


(theeth) #10

Or better, coordinates transformations, for example:

Theta = atan2(Y, X)
R = sqrt(X ** 2 + Y ** 2)
Z = Z

that would defined cylindrical coordinates from the cartesian coordinates. That way, you could define any type of coordinates you want, as long as you can transfer from cartesian.

Martin