应用声学
應用聲學
응용성학
APPLIED ACOUSTICS
2001年
3期
31-33,10
,共4页
罗万伯%张露%李志蜀%罗霄峰%罗霄岚%Duanpei Wu
囉萬伯%張露%李誌蜀%囉霄峰%囉霄嵐%Duanpei Wu
라만백%장로%리지촉%라소봉%라소람%Duanpei Wu
语音识别%语音信号%非线性时间规正%实时实现
語音識彆%語音信號%非線性時間規正%實時實現
어음식별%어음신호%비선성시간규정%실시실현
本文提出了语音信号的一种时域-频域-能量表示,并给出了算法,可用于孤立词语音识别.这种时域-频域-能量表示有两个特点:基于短时能量梯度的非线性时间规正,可保留语音信号频域的过渡特性,丢掉其稳态特性;计算量小,适于实时应用.
本文提齣瞭語音信號的一種時域-頻域-能量錶示,併給齣瞭算法,可用于孤立詞語音識彆.這種時域-頻域-能量錶示有兩箇特點:基于短時能量梯度的非線性時間規正,可保留語音信號頻域的過渡特性,丟掉其穩態特性;計算量小,適于實時應用.
본문제출료어음신호적일충시역-빈역-능량표시,병급출료산법,가용우고립사어음식별.저충시역-빈역-능량표시유량개특점:기우단시능량제도적비선성시간규정,가보류어음신호빈역적과도특성,주도기은태특성;계산량소,괄우실시응용.
A time-frequency-energy representation of speech signal and its algorithm are introduced to isolated-word speech recognition. It can be characterized by two aspects: (1) non-linear time normalization, based on the gradients of short-time energy in a specific number of frequency bands, retains the transient portions and ignores the steady-state portions of speech signal in frequency domain. (2) real-time implementation due to low computational load.