Real time implementation of audio spectrogram on field programmable gate array (fpga)

An audio spectrogram is a visual representation of sound. Spectrogram in general is defined as the time varying spectral representation which shows the variation of spectral density of a signal with respect to time. Spectrograms (also known as voicegrams, sonograms or spectral waterfalls) are typically used to identify phonetic sounds, to analyze the cries of animals and also in other fields like speech processing, sonar, seismology, etc. Spectrograms can be created using two methods, using a series of band pass filters to form the approximated filter bank and short-time Fourier transform (STFT) calculated from the time signal. This thesis concentrates on using the short-time Fourier transform (STFT) to obtain the audio spectrogram. The objective of this research is the real time implementation of real time spectrogram of an audio signal on a video monitor using Xilinx Virtex-5 ML506 Evaluation Board. The Xilinx ML506 Virtex-5 Evaluation Board has powerful audio and video capabilities, which are utilized in this research. The input audio signal is processed in the FPGA to calculate STFT of the signal. Once the STFT is calculated, it must be converted into a form suitable for displaying by the video monitor. The video signal sent to the monitor shows the real time audio spectrogram. This research has several applications in scientific and commercial devices. The design is written is Verilog & VHDL, simulated using Xilinx ISE Web pack and programmed on the Xilinx Virtex-5 ML506 Evaluation Board.