资源描述:
《基于基音周期模型的语音识别》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、攀枝花学院本科毕业设计(论文)摘要摘要基音周期是表征语音信号本质特征的参数,属于语音分析的范畴,只有准确分析并且提取出语音信号的特征参数,才能够利用这些参数进行语音识别处理。语音识别率的高低,都依赖于对语音信号分析的准确性和精确性,因此基音周期的研究在语音信号的处理应用中具有十分重要的作用。本论文通过两种算法在Matlab上实现基音周期的检测,分别是短时自相关函数法和短时平均幅度差函数法。通过实验得到的基音周期结果的分析来识别不同的人的语音信号。论文首先介绍了语音基音检测算法在语音识别方面的研究背景极其重要意义。其次对现有的基音检测算法进行了
2、归纳和总结,并详细的介绍本文将用的两种基本基音检测算法的基本原理及实现。最后在Matlab上对语音信号进行基音周期的检测。本设计为语音信号的基音周期检测,采集语音信号,对语音信号进行处理,区分清音浊音,并通过对采样值进行滤波、分帧、求短时自相关函数,得到浊音的基音周期。关键词:预处理,基音检测,自相关函数法,平均幅度差函数法,基音轨迹,语音识别攀枝花学院本科毕业设计(论文)AbstractAbstractpitchischaracterizationparametersofspeechsignalcharacteristics,belongs
3、tothecategoryofspeechanalysis,onlytheaccurateanalysisandextractthecharacteristicparametersofspeechsignalcanusetheseparameterstospeechrecognitionprocessing.voicerecognitionrateofhighandlow,aredependentontheaccuracyoftheanalysisofspeechsignalandaccuracy,sothepitchperiodofthes
4、tudyonspeechsignalprocessingplaysanimportantroleinapplication.inthispaper,throughthetwoalgorithminMatlabpitchperioddetection,respectivelyisshort-timeautocorrelationmethodandshorttimeaveragemagnitudedifferencefunctionmethod.Pitchperiodisobtainedbytheexperimentresultsofspeech
5、signalanalysistoidentifythedifferentpeople.Thecurriculumdesignofpitchdetectionofspeechsignals,speechsignal,thespeechsignalprocessing,distinguishbetweenthevoicedandunvoicedspeechhasdifferentsamplesvaluesinthefiltering,framing,askstheshort-timeautocorrelationfunction,havevoic
6、edsoundpitchperiod.Thethesisfirstintroducestheresearchbackgroundofspeechpitchdetectionalgorithmisveryimportant.Nexttotheexistingalgorithmforpitchdetectionaresummarized,andadetailedintroductiontothisarticlewillusethetwokindsofbasicpitchdetectionalgorithmisthebasicprinciplean
7、drealization.FinallyinMatlabonspeechsignalpitchperioddetection.Keywords:pretreatment,pitchdetection,autocorrelationfunction;theaveragemagnitudedifferencefunctionmethod;pitchcontrail,Speechrecognition攀枝花学院本科毕业设计(论文)目录目录摘要IAbstractII1绪论11.1引言11.2基音周期检测对语音识别的意义21.3基音周期检测现状31.4
8、论文的结构安排42基音检测常用的算法52.1引言52.2常用的基音检测算法及其原理52.2.1自相关函数法52.2.2平均幅度差函数法92.3本章小结113基于Ma