surf plot behaves strange!

Hi,

i generate a surface with surf. The problem is, that the coloring does not match the actual Data.

The Data is shown by the Z-Value in the surface (as usual). I uploaded the created Image. There i pointed out what i mean exactly:

http://i.imgur.com/XRgBl.png

How can the same color be attached to such different z-values?Is this a bug? Has anybody an idea how i can fix this?

Thanks for the attention

Answer 1

[Kevin Holst]

Surf attaches a color to a face based on one point on that face, and each face will be created by 4 points, so that color may not make sense for 3 of those points. This is especially true when there is a lot of variation point to point.

http://www.mathworks.com/help/techdoc/ref/surf.html

What I typically do is set shading to interp either within the call to surf itself or after the call.

[Miro]

You say that one color possibly makes no sense for the points, but how can they then be displayed correctly when i do not use the meshgrid (top picture in the link?

Includeing shading interp makes it even worse, cause it takes the colors, which already seem to be interpolated (pic 2 and 3) as basis.

[Kevin Holst]

Your top picture shows surf(data) which plots each data sample versus its row/column numbers. So your data matrix must be something close to size of 50x560. The reason that when you use meshgrid, you're getting a different looking plot is because it appears that there aren't 560 distinct values in the y-direction. It appears that some are repeated due to the resolution of your y-variable.

One color makes sense (programatically) for those two points you show because of how the surf function attaches a color to a surface face by default. It requires 4 points to make a face, and it chooses 1 of those 4 points to base the color of the face on.

Answer 2

[Walter Roberson]

The colors of a surf() plot or a pcolor() plot are not directly derived from the Z value of individual points. Instead they are interpolated from nearby points. See http://www.mathworks.com/help/techdoc/ref/surf.html under "Algorithms"

[Kevin Holst]

What you described is what happens when using shading interp, which is not the default. The default is faceted in which "each mesh line segment and face has a constant color determined by the color value at the endpoint of the segment or the corner of the face that has the smallest index or indices.'

http://www.mathworks.com/help/techdoc/ref/shading.html

[Miro]

exactly. And this works fine if i plot the data without a specified meshgrid. But when i use the dimensioned grid, the problem occurs. The y-Axis is a time vector, and the time is not continuous, could that possibly be the reason?