1) front-end features
语音前端特征
2) eigen voice
特征语音
1.
Speaker adaptation algorithm based on eigen voice;
基于特征语音的说话人自适应算法研究
3) phonetic features
语音特征
1.
Therefore,learners should master the rule of pronunciation accurately in vocal music learning,combine singing techniques with Italian pronunciation,and better understand the phonetic features of euphonic singing s target,thus enhance their own singing level of vocal music.
因此,在声乐学习中,学习者要准确地掌握发音规则,将意大利语音与歌唱技巧有机结合起来,更好地了解美声歌唱追求的语音特征,从而提高自己的声乐演唱水准。
2.
The geographical distribution of dialectal phonetic features is unbalanced.
方言各语音特征的地理分布并不平衡,根据影响扩散的语言与社会因素,可以推断扩散的方向及其反映的历史层次。
3.
Based upon an analysis of 63 American and British brand names ranked among the world s 100 most valuable brands in 2004,this paper makes a discussion about phonetic features of brands English names with special reference to letter number,syllable number and initial sound.
他们的语音特征为:英语品牌名称多以4~9个字母构成,2~3个音节者居多,且名称起始音多为辅音特别是塞音、擦音和双唇鼻音。
4) phonetic characteristic
语音特征
1.
The present paper makes an analysis of the character, the function,the cause and the phonetic characteristics of light tones.
对轻声的性质、轻声的功能、轻声的原因及轻声的语音特征进行了分析。
5) phonetic feature
语音特征
1.
The essay mainly analyses the phonetic features,constructive forms and grammatical function of the words ending with"zi"in Wu tai Dialect,and make comparisons with mandarin.
文章着重分析了五台方言的“子”尾的语音特征、构成形式和语法功能,并与普通话“子”尾进行了比较。
2.
This paper studies the phonetic features of Yongfeng Town in Shuangfeng county of Hunan province , and have research on several important phonetic features.
五、重要语音特征的专题研究。
6) phonological features
语音特征
1.
The article is actually few that through describing the phonology of a tipical Xiang dialect then to discussing the phonological features of Xiang dialect.
由于种种原因,湘语的语音研究还不够深入,多数论文或著作限于对现实平面语音特征的粗线条勾勒,通过描写一典型湘方言的语音进而探讨湘语语音特征的文章尚属少见。
补充资料:语音识别的特征抽取
语音识别的特征抽取
feature extraction of speech recognition
yuyin shib一e de teZheng ehouqu语音识别的特征抽取(feature extractionofs】珍ech reC雌贝ition)用信号处理技术提取语音信号中可供相互区分的特征参数。特征参数大致分为两类,一类反映语音的频谱包络特性,即声道响应特征,如共振峰参数、线性预测系数、倒谱系数等。另一类反映语音的时域特性,如能量、音调等。特征参数选取对识别性能影响很大,目前用的最多而又有效的参数是倒谱系数。也有用FFq,系数和带通滤波器组的输出作为特征参数。能量是语音信号时域特征,将它用于识别系统,可使性能得到改善。倒谱系数和能量的变化,即它们的动态特性是语音信号的重要特征之一,在识别系统中起重要作用,它和静态特性在很大程度上是无关和互补的,有人建议使用倒谱的回归系数作为动态特性。选用什么特征参数要有利于识别,即在多维特征空间中要易于将不同的识别音区分开。许多研究表明,频域特征参数在频率维作非线性规正,使之符合人耳分析频率的特性,能给识别带来好处。
说明:补充资料仅用于学习参考,请勿用于其它任何用途。
参考词条