This document summarizes a lab experiment on performing fast Fourier transforms (FFTs). The lab examines using FFTs to analyze digital signals and view their spectral contents. It discusses how FFTs provide an efficient algorithm for calculating the discrete Fourier transform, requiring only Nlog2(N) computations compared to N2 for the discrete Fourier transform. The document outlines how to properly perform FFTs on sampled signals using a digital signal processor, including setting the sampling frequency, block size, and applying a multiplication factor and windowing to obtain accurate spectral results. Finally, it provides tasks for generating an embedded sine wave and building a virtual instrument to acquire an audio signal, view its spectrum using FFTs, and output the signal.
This document summarizes a lab experiment on performing fast Fourier transforms (FFTs). The lab examines using FFTs to analyze digital signals and view their spectral contents. It discusses how FFTs provide an efficient algorithm for calculating the discrete Fourier transform, requiring only Nlog2(N) computations compared to N2 for the discrete Fourier transform. The document outlines how to properly perform FFTs on sampled signals using a digital signal processor, including setting the sampling frequency, block size, and applying a multiplication factor and windowing to obtain accurate spectral results. Finally, it provides tasks for generating an embedded sine wave and building a virtual instrument to acquire an audio signal, view its spectrum using FFTs, and output the signal.
This document summarizes a lab experiment on performing fast Fourier transforms (FFTs). The lab examines using FFTs to analyze digital signals and view their spectral contents. It discusses how FFTs provide an efficient algorithm for calculating the discrete Fourier transform, requiring only Nlog2(N) computations compared to N2 for the discrete Fourier transform. The document outlines how to properly perform FFTs on sampled signals using a digital signal processor, including setting the sampling frequency, block size, and applying a multiplication factor and windowing to obtain accurate spectral results. Finally, it provides tasks for generating an embedded sine wave and building a virtual instrument to acquire an audio signal, view its spectrum using FFTs, and output the signal.
This document summarizes a lab experiment on performing fast Fourier transforms (FFTs). The lab examines using FFTs to analyze digital signals and view their spectral contents. It discusses how FFTs provide an efficient algorithm for calculating the discrete Fourier transform, requiring only Nlog2(N) computations compared to N2 for the discrete Fourier transform. The document outlines how to properly perform FFTs on sampled signals using a digital signal processor, including setting the sampling frequency, block size, and applying a multiplication factor and windowing to obtain accurate spectral results. Finally, it provides tasks for generating an embedded sine wave and building a virtual instrument to acquire an audio signal, view its spectrum using FFTs, and output the signal.
Oct 2012 – Feb 2013 CECOS College of Engineering and IT dawar@cecos.edu.pk Fourier transform Fourier transform is used to view the spectral composition of analog signals Discrete Fourier transform (DFT) is used to view the spectral contents of digital signals
N2 complex operations are needed to calculate the DFT
DFT can be efficiently calculated using fast Fourier transform (FFT)
CECOS College of Engineering and IT dawar@cecos.edu.pk
Fast Fourier Transform FFT make use of some clever algorithms to do the same thing as DFT does, but in much less time FFT requires Nlog2 (N) computations In calculating FFT two parameters play very important role 1. The sampling frequency (fs) 2. The block size or the number of points (N)
CECOS College of Engineering and IT dawar@cecos.edu.pk
Fast Fourier Transform For block size of N, N time domain samples are used to calculate N frequency domain samples The 0th sample will represent 0 Hz and (N-1)th sample will represent fs Hz The multiplication factor for the FFT should be fs/(N-1) For sampling frequency of fs Hz the highest frequency contained in the signal can be fs/2 Hz This means that the useful information lye between 0 and (N/2)-1 samples
CECOS College of Engineering and IT dawar@cecos.edu.pk
Fast Fourier Transform
CECOS College of Engineering and IT dawar@cecos.edu.pk
Fast Fourier Transform
CECOS College of Engineering and IT dawar@cecos.edu.pk
Fast Fourier Transform
CECOS College of Engineering and IT dawar@cecos.edu.pk
Fast Fourier Transform
CECOS College of Engineering and IT dawar@cecos.edu.pk
FFT using DSP
Embedded sine waveform
CECOS College of Engineering and IT dawar@cecos.edu.pk
FFT using DSP
CECOS College of Engineering and IT dawar@cecos.edu.pk
FFT using DSP The graph showing the spectrum has no idea about the x-axis
CECOS College of Engineering and IT dawar@cecos.edu.pk
FFT using DSP Alter the graph properties using the multiplication factor
1000/127
CECOS College of Engineering and IT dawar@cecos.edu.pk
FFT using DSP Now the correct spectrum can be seen
CECOS College of Engineering and IT dawar@cecos.edu.pk
FFT using DSP You may limit the frequency axis to the significant portion of spectrum
CECOS College of Engineering and IT dawar@cecos.edu.pk
FFT using DSP The useful portion of the spectrum
CECOS College of Engineering and IT dawar@cecos.edu.pk
FFT using DSP Windowing reduces the spectral leakage in calculating FFT using DSPs
No window used Hanning window used
Spectral Leakage
CECOS College of Engineering and IT dawar@cecos.edu.pk
Task 1. Generate a 50Hz Embedded sine wave, sampled at 8000Hz and plot its 128, 256, 512 and 1024 points FFT 2. Design a VI that acquires an audio signal from the analog input of the kit, shows its spectrum and routes the audio to the analog output of the kit
CECOS College of Engineering and IT dawar@cecos.edu.pk