High Frequency Compensation for Compressed Digital Audio using Sampled-Data Control

Abstract

Compressed digital audio technologies such as MD and MP3, cut down high frequency component data which have less influence on hearing. However, with an increase of compression rate, the increase in sound unclearness is observed. Therefore, a technology to realize high-quality sound is required to respond to the increasing demands for high-quality sound reproduction. In order to meet this challenge, a method is considered, in which the removed high frequency components are reproduced. Focusing attention on the sampled-data control theory to establish an optimum signal processing system using the original sound (CD sound) frequency characteristics, we have applied this theory to high frequency compensation of compressed digital audio signals in the form of “interpolating data points between discrete data to improve signal quality.” As a result, high frequency components of compressed digital audio signals are compensated so that they are close to the frequency spectrum of CDs, which enables sound quality improvement with a remarkably small amount of calculation load. This technology has been implemented on semiconductors which have been commercialized in LSIs for MD players.