Crest factor minimization in the frequency domain
M. Solomou, C. Evans, D. Rees
Abstract
M. Solomou, C. Evans, D. Rees
Abstract
This paper examines the minimization of the signal crest factor using a frequency-domain method. This method is based on the observation that low crest factor signals possess properties that reduce the effect of system nonlinearities. The main idea of this method is that instead of setting the signal crest factor as the cost function to be minimized, multisine signals are designed with their harmonic phases selected to minimize the distortion introduced by a cubic nonlinearity on the measurement of a frequency response function. The resulting signals are compared with other types of low crest factor signals. It is found that the method does not produce the lowest possible crest factors but that the signals do minimize the nonlinear distortion introduced by a cubic nonlinearity, while possessing a low crest factor. The signals are thus termed minimum distortion multisines.
OpenAlex reports 16 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This paper examines the minimization of the signal crest factor using a frequency-domain method. This method is based on the observation that low crest factor signals possess properties that reduce the effect of system nonlinearities. The main idea of this method is that instead of setting the signal crest factor as the cost function to be minimized, multisine signals are designed with their harmonic phases selected to minimize the distortion introduced by a cubic nonlinearity on the measurement of a frequency response function. The resulting signals are compared with other types of low crest factor signals. It is found that the method does not produce the lowest possible crest factors but that the signals do minimize the nonlinear distortion introduced by a cubic nonlinearity, while possessing a low crest factor. The signals are thus termed minimum distortion multisines.
Key concepts: Crest factor, Crest, Distortion (music), Total harmonic distortion, Nonlinear distortion, Frequency domain, Nonlinear system, Minification