Normal components given by surfnorm function do not seem to be the correct ones

Hi,

I am trying to get the normal components of a surface using the surfnorm function ([U, V, W] = surfnorm(X,Y,Z)) but when I try to plot them using quiver3 (quiver3(X,Y,Z,U,V,W)), I realize that they are not the correct values.

I compare my plot to the one using surfnorm directly to plot (surfnorm(X,Y,Z)).

The full and simple code to show the issue is this:

[X,Y] = meshgrid(-2:0.2:2,-1:0.2:1);

Z = X.* exp(-X.^2 - Y.^2);

figure;

[U,V,W] = surfnorm(X,Y,Z);

surf(X,Y,Z);

hold on

quiver3(X,Y,Z,U,V,W);

figure;

surfnorm(X,Y,Z);

If I zoom in to the surface I can see that in the first plot, the vector coming from the surface are not actually normal to it. The second looks correct, but from this I cannot extract the normal components.

Thanks.

[David Goodmanson]

Hi Jose,

try 'axis equal' after the plot commands and see what that looks like.

[Jose Enrique Almanza]

I added 'axis equal' to both figrues. Now the first one seems to be right.

Is there any reason why they don't look the same? The code is the same, just adding 'axis equal' after plotting.

[David Goodmanson]

Hi Jose,

it's easier to see in 2d because when you rotate a 3d object, things can get foreshortened, which is the basic reason that angles appear to change. In 2d, when you stretch one of the axes in a plot, angles look different although the data is the same.

figure(1)
quiver([0 0],[0 0],[1 1],[-1 1],0)
axis equal        % angles look like they should
grid minor

figure(2)
% same data
quiver([0 0],[0 0],[1 1],[-1 1],0)
xlim([-.75 1.75])
ylim([-2 2]) 
grid minor

[Jose Enrique Almanza]

Hi David,

Thanks for your response and for clarifying.