Thisstudyproposesapost-processortoimprovetheharmonicstructureofavowelinanenhancedspeech,enablingthespeechqualitytobeimproved.Initially,aspeechenhancementalgorithmisemployedtoreducethebackgroundnoiseforanoisyspeech.Hencetheenhancedspeechispost-processedbyahybrid-medianfiltertoreducethemusicaleffectofresidualnoise.Sincetheharmonicspectraareimpactedbybackgroundnoiseandaspeechenhancementprocess,thequalityofavowelisdeteriorated.Aharmonicregeneratedmethodisdevelopedtoimprovethequalityofpost-processedspeech.Experimentalresultsshowthattheproposedmethodcanimprovethequalityofpost-processedspeechbyadequatelyregeneratingharmonicspectra.
