The Power Spectrum image is compute directly from the results of passing the surface data through a Fast Fourier Transform (FFT) algorithm.  The FFT algorithm outputs real and imaginary components which can be used to obtain amplitude and phase information of the surface.  The power spectrum image is computed as the from the real and imaginary components of the FFT output according to the equation:  Sqr((Real*Real) + (Imag*Imag)).  The low frequency components of the surface are displayed in the center of this image.  The power spectrum image displays repeated patterns as narrow peaks, the co-ordinates of which describe their periodicity and direction.  Please note in the images below how the measurement of a sinusoidal surface is indicated in the power spectrum image by very narrow spikes in the image.

 

There are two options for displaying the power spectrum: Linear and Logarithmic.  Selecting the logarithmic option is often convenient because it tends to enhance the weaker components of the surface making them visible.