What are the Basic Differences Between Baseband, Subset, and Zoom FFTs and When Should I Use Each? Primary Software: LabVIEW Toolkits>>Sound and Vibration ToolsetPrimary Software Version: 2.0 Primary Software Fixed Version: N/A Secondary Software: LabVIEW Toolkits>>Spectral Measurements Toolkit
Problem: The LabVIEW Sound and Vibration Toolkit and Spectral Measurements Toolkit offer Baseband, Baseband Subset, and Zoom options for performing Fast Fourier Transform (FFT) calculations. What are the differences between these methods and when is each most appropriate? Solution: The similarities and differences between these three FFT methods are summarized below: Baseband FFT The baseband FFT is the simplest option. Like all FFTs, the baseband FFT requires a time-domain input and returns an output in the frequency domain. The “Number of Lines” defines how many frequency components are present in the spectrum. The baseband FFT will always produce a number of lines equal to n/2, where n is the number of time-domain samples passed to the FFT. The width or “frequency resolution” of each line is equal to (Fs / n), where Fs denotes the sampling frequency and n denotes the number of time-domain points. Alternately, the frequency resolution can be written as (1 / acquisition time). The frequency range of a baseband FFT spectrum is always DC to (Fs / 2). The baseband FFT is the best method when you are interested in a broad range of frequencies. It is appropriate for both on-line and off-line processing. When the frequency range of interest is small when compared to Fs / 2, you can improve performance by going with a subset or zoom FFT. Baseband Subset FFT A second option is the baseband subset operation (hereafter called simply “subset”). The subset FFT returns a narrower frequency range than the baseband FFT. The minimum and maximum range of operation are programmable, and spectral results are only returned within this range. The FFT computation is less intensive than a baseband FFT producing the same frequency resolution. However, this improvement is less than “order n”. Consider a baseband FFT (frequency range DC to Fs/2) and a subset from Fs/8 to Fs/4. The subset covers only ¼ the frequency range of the baseband in this case. It will yield a substantial performance improvement, although the improvement will be less than a factor of 4. It is important to remember that the subset still requires just as much data as the baseband to produce a spectrum with a given frequency resolution. In other words, it yields an improvement in computation time but not acquisition time. The subset FFT is a good choice when you are interested in only a narrow frequency range and would like better performance than the baseband FFT. Like the baseband FFT, the subset operation is appropriate for both on-line and off-line analysis applications. Although generally called an FFT, the subset calculation is actually based on another algorithm known as the Discrete Zak Transform (DZT). Zoom FFT The third option is the zoom FFT. Like the subset FFT, it allows you view a narrower spectrum of interest in order to improve processing time. And like the subset, it provides an improvement in processing time but not in acquisition time. The zoom FFT algorithm requires as much data as the baseband FFT or subset FFT to produce a spectrum with a given frequency resolution. The zoom FFT is most appropriate for applications requiring narrow frequency resolution, on-line analysis, and frequent data updates. Unlike the baseband and subset FFTs, the zoom algorithm maintains an internal state and can be placed in a fast-running loop where the acquisition block size is actually smaller than what is required to produce the desired frequency resolution. When used in such a configuration, the FFT will not yield new data with every iteration. It automatically accumulates new data until enough has been acquired to produce an updated spectrum. The zoom FFT supports a feature known as overlap. The default overlap setting is 0%, where each new spectrum will be computed from completely “fresh” data. An overlap setting of 75% will mean that a given spectrum uses 75% “old” data from the last spectrum, while 25% of the data is new. This setting would generate updates 4 times as often as the 0% option, although each update would have only a marginal change in information content. Overlap allows you to use some processing power to improve the display and “interactivity” of your application when performing long-running acquisitions. Related Links: Attachments:
Report Date: 05/17/2004 Last Updated: 04/09/2006 Document ID: 39GALI2L |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
