Although numerous methods of calculating digital Fourier spectra are known today, they are all variations of the Danielson-Lanczos (symmetry of cosine functions) or of the Cooley-Tukey algorithm (successive decomposition of one-dimensional data strings into two-dimensional arrays). Hardware implementation of an FFT algorithm can result in higher spectrum generation speed, higher real-time bandwidth of data analysis, faster refresh rate of control functions in servo-applications, or in instruments of limited flexibility geared for special applications. The cost and performance characteristics of any FFT hardware depend on how the major tasks are implemented: complex multiplication, phasor generation, data and phasor indexing, and memory management. For completeness, the less glamorous, but equally important, parts of signal analysis systems should also be considered: overall system control, data acquisition, user interface, retrieval of results, etc. Only in the context o{ the complete system can the FFT hardware be fully evaluated.