지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학위논문
- 저자정보
- 지도교수
- 이인성
- 발행연도
- 2016
- 저작권
- 충북대학교 논문은 저작권에 의해 보호받습니다.
이용수2
이 논문의 연구 히스토리 (2)
- 이 논문의 후속연구가 궁금하신가요?
- 연관 학술논문 또는 학술발표를 통해 보다 발전된 연구결과를 확인하실 수 있습니다.
- 이 논문의 연구 히스토리 확인하기
초록· 키워드
오류제보하기
This paper proposes a noise suppression method using Signal-to Noise Ratio estimation in the two microphone array environment of close spacing.
The conventional method uses a noise suppression method using a gain function obtained through the Signal-to Noise Ratio estimation based on coherence function from full band. Input signals received from two microphone and time-domain signal converts frequency-domain signal and calculate the coherence function between the signals. Estimating a Signal-to-Noise Ratio is estimated using the real and imaginary parts of the coherence function. The gain function is obtained by applying the Wiener filter to estimates and noise suppressed output signal obtain through synthetic and restore process of gain function and input signal.
However, this method cause performance decreased by the noise damage that affects all the feature vector component.
So, We propose a noise suppression method. First, input signals obtained from two microphones separate the voice sound and the unvoice sound by using VAD. Window function apply according to the respective characteristics. Time-domain signals convert frequency-domain signal using Fast Fourier Transform. allocates a frequency-domain signal into N constant multi frequency band each frequency band gets a gain function through SNR estimation based on coherence function.
Performance evaluation of the proposed method is shown by comparison with PESQ(Perceptual Evaluation of Speech Quality) value which is an objective quality evaluation method provided by the ITU-T.
The conventional method uses a noise suppression method using a gain function obtained through the Signal-to Noise Ratio estimation based on coherence function from full band. Input signals received from two microphone and time-domain signal converts frequency-domain signal and calculate the coherence function between the signals. Estimating a Signal-to-Noise Ratio is estimated using the real and imaginary parts of the coherence function. The gain function is obtained by applying the Wiener filter to estimates and noise suppressed output signal obtain through synthetic and restore process of gain function and input signal.
However, this method cause performance decreased by the noise damage that affects all the feature vector component.
So, We propose a noise suppression method. First, input signals obtained from two microphones separate the voice sound and the unvoice sound by using VAD. Window function apply according to the respective characteristics. Time-domain signals convert frequency-domain signal using Fast Fourier Transform. allocates a frequency-domain signal into N constant multi frequency band each frequency band gets a gain function through SNR estimation based on coherence function.
Performance evaluation of the proposed method is shown by comparison with PESQ(Perceptual Evaluation of Speech Quality) value which is an objective quality evaluation method provided by the ITU-T.
목차
I. 서 론 1II. 음성품질 향상 기법 32.1 통계 모델을 기반으로 한 음성향상 기법 32.2 인간의 청각 특성을 고려한 음성향상 기법 7III. Coherence function 기반의 전 밴드 잡음억제 시스템 102.1 Coherence function 정의 102.2 신호 대 잡음비 추정 12Ⅳ. Coherence function 기반의 다중 주파수 밴드 잡음억제 시스템 173.1 유/무성음에 따른 window함수 적용 183.2 다중 주파수 밴드 분할 213.3 다중 주파수 밴드 신호 대 잡음비 추정 22V. 성능평가 265.1 실험 환경 265.2 PESQ를 이용한 실험 결과 275.3 스펙트로그램을 이용한 실험 결과 34Ⅵ. 결론 45참고문헌 47감사의 글 51
최근 본 자료
댓글(0)
0