| 摘要 |
PROBLEM TO BE SOLVED: To prevent the increase in power consumption that is caused by the increase in the number of computing elements. SOLUTION: The voice feature value obtained in an encoding process at a voice encoding process part 104 is turned into the frequency characteristic data accordant with the frequency distribution of voices via a pitch information processing part 108 and a spectrum information processing part 107. A filter coefficient corresponding to the frequency characteristic data is obtained via a filter coefficient store part 110, and a preprocessing filter part 112 performs its filtering operation based on the filter coefficient. The filtered transmitting signal is used to the detection of voices serving as the elements that decide a noise eliminating operation state. Then the noises of non-voice sections are eliminated at a noise elimination processing part 103 based on the voice detection results, etc. Thus, it's not required to use any computing element to extract the voice spectrum, etc., and accordingly the increase of power consumption is prevented for a telephone set. |
