next up previous contents
Next: 16.1.4 Rendering Data Values Up: 16.1 Scalar Field Visualization Previous: 16.1.2 Representing Data Values   Contents

16.1.3 Interpolating Data Values

Be careful about interpolating between sampled scalar values. Most interpolation techniques assume a linear relationship between sampled data values, which might not be the case. A linearly interpolated set of colors applied to non-linear data can create misleading and incorrect images. There are two basic approaches to avoiding this problem. First, don't interpolate between data values at all. Only display the data values at the sample points. This results in no interpolation problems, but possibly makes the data harder to interpret. The other possibility is to ensure adequate sampling. If the data is known to be linear between sample points, within an acceptable error, then interpolation can be safely done. What ``acceptable error'' means depends on the application.