发明名称 音声信号処理装置
摘要 PROBLEM TO BE SOLVED: To improve an audio sound quality when a narrow band audio sound signal such as a mobile phone is subjected to band spreading.SOLUTION: When vocal track transfer characteristics extracted by performing a linear estimation analysis for an input narrow band audio signal is subjected to wideband spreading, an interpolation filter 32 designed by using a sample value Hcontrol which takes frequency characteristics of an original voice into a consideration is utilized. With a filter coefficient for voiced sound and a filter coefficient for unvoiced sound prepared as the filter coefficient of the interpolation filter 32, a coefficient selection part 7 switches a coefficient based on a result discriminated by a voiced sound/unvoiced sound discrimination part 4 which discriminates whether an input audio sound is the voiced sound or the unvoiced sound. The accuracy of the discrimination is improved and the quality of the wide band audio voice obtained by the band expansion is also improved by adding not only the periodicity of the original audio voice but also the discrimination of the whole inclination of the frequency spectrum for the discrimination between the voiced audio sound and the unvoiced audio sound.
申请公布号 JP2015206958(A) 申请公布日期 2015.11.19
申请号 JP20140088752 申请日期 2014.04.23
申请人 发明人
分类号 G10L21/0388;G10L25/93 主分类号 G10L21/0388
代理机构 代理人
主权项
地址