1
$\begingroup$

per this post, this formula

$Z = 1 - \sqrt{X^{2} + Y^{2}}$

generates a cone where the point is at (0, 0, 1) and it spreads out below that. It meets the x-y plane at the unit circle

enter image description here

I am trying to reproduce with Python

ax = plt.figure().gca(projection='3d')
xx, yy = np.meshgrid(np.arange(-1,1.1,.1), np.arange(-1,1.1,.1))
zz = 1 - np.sqrt(xx**2, yy**2)
ax.plot_surface(xx, yy, zz, alpha=.5)

and get this "roof"

enter image description here

each of xx, yy, zz is a 21 by 21 matrix, even if I increase them to 210 by 210, nothing changes, what am I missing?

| improve this question | |
$\endgroup$
  • $\begingroup$ Looks like your mesh is undersampled, in that it only has 6 vertices so its an approximate cone.. Can you increase the resolution some how ? $\endgroup$ – PaulHK Oct 15 '19 at 2:01
  • $\begingroup$ @PaulHK each of xx, yy, zz is a 21 by 21 matrix, even if I increase them to 210 by 210, nothing changes $\endgroup$ – JJJohn Oct 15 '19 at 2:13
  • $\begingroup$ What is np.sqrt(xx**2, yy**2)? I would expect something like np.sqrt(xx**2 + yy**2). Otherwise you're getting back an array of square roots, right? $\endgroup$ – user1118321 Oct 15 '19 at 2:21
  • $\begingroup$ @user1118321 Thanks a lot! Please move your comments to answer, I'll accept it. $\endgroup$ – JJJohn Oct 15 '19 at 2:26
2
$\begingroup$

You need to change the call to sqrt from:

np.sqrt(xx**2, yy**2)

to:

np.sqrt(xx**2 + yy**2)

Otherwise you're passing an array of values to the function and will be returned an array of values for the answer.

See here for details:

>>> np.sqrt([1,4,9])

array([ 1., 2., 3.])

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.