Peak harmonic distortion due to quantization
Year: 1971
Type: dissertation
Abstract: In this thesis, the problem of quantization noise 1s presented, and recent efforts 1n this area are reviewed. With the motivation for further Investigation Into the problem explained, the purpose of this thesis 1s stated to be the determination of peak harmonic distortion due to quantization for predominantly single frequency Inputs. Two cases were examined with pure sinusoid and sine wave plus Gaussian bandlimited white noise. The method used was to simulate the quantization process on the computer, and to use a Fast Fourier Transform algorithm to analyze the spectra of the quantized signals. For pure sinusoidal Inputs, the location of the peak harmonic distortion 1n the quantization noise spectrum was found to be very sensitive to the degree of loading of the quantizer. However the magnitude of the peak distortion when plotted as a function of the number of bits totally used by the Input was fitted very well with a straight line of slope -6dB/b1t. Moreover the largest component 1n the quantization noise spectrum was observed to be about 4dB above the average noise spectral density across the entire frequency band of observation. The addition of noise to the sine wave was anticipated to have a smoothing effect on the quantization noise spectrum. This phenomenon was observed for a specific set of Input noise samples but the results are not conclusive, becuase after further investigation of the noise generation mechanism, the statistical properties of the synthesized noise signal were found to be unsuitable for analysis of power spectra. However since the pure sinusoidal inputs represent the worst case condition for harmonic distortion due to quantization, the measured peaks will provide the upper bounds necessary for specifications in engineering system designs.
Author: On-Ching Yue
Institution: Rochester Institute of Technology
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W593044123
