Masters Thesis

Audio/Image processing in frequency domain using 2D FFT

FFT and similar frequency transforms have a lot of different applications in today's advanced technology. FFT is a mathematical method to convert a signal from time domain to frequency domain. Frequency domain transforms like FFT are widely used in image processing and enhancement techniques. Medical devices like MRI and CT scan are using image processing using FFT to process patient's body images. It's also being used in audio and speech processing. The objective of this project is to use FFT to process audio signals and create their Spectrogram in order to differentiate between different music styles and instruments. The work includes processing different animal voices to detect their frequency domain and also find difference between their voices in different situation. Other part is to process different music styles and instruments using spectrogram to see if we could use it to distinguish between different instruments in a play, music styles, or musicians without listening to the music itself. The other area of concentration is using FPGA to implement the spectrogram and adding chipscope IP core to the hardware to be able to test and debug the implemented Spectrogram device and also use the chipscope to show the results on computer screen. Introduction of FPGA was mostly for the purpose of testing and debugging, but now a day because of its fast time to market time and also ease of use, it's being used frequently to design different digital systems in several applications. Hardware modules have been designed using VHDL programming Language and implemented using AtlysTM Spartan-6 Xilinx FPGA Evaluation board.

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.