Search | Korea Science

Min, So-Yeon;Bae, Myung-Jin
- Proceedings of the KAIS Fall Conference
- /
- 2006.05a
- /
- pp.364-367
- /
- 2006
음성신호를 관찰하였을 때 성문특성으로 인해서 고주파 쪽 특성이 약화되는 경향이 있다. 약화된 고주파 특성을 보상하기 위하여 프리 엠퍼시스 필터를 통해 보상하고 있다. 프리 엠퍼시스 필터를 간단한 수식으로 표현하면 y(n)=s(n)-As(n-1)와 같이 차분 방정식으로 나타낼 수 있다. 여기서 A값은 보통 0.9에서 1사이의 값을 사용한다. 본 논문에서는 QMF 필터를 이용하여 입력신호를 고주파와 저주파의 2개의 대역으로 분할하여 각 밴드에 프리 엠퍼시스 필터를 적용하여 약화되어진 특성을 정확히 보상하는 방법을 제안한다.
PDF

민소연;강은영;배명진
- Proceedings of the IEEK Conference
- /
- 2000.11d
- /
- pp.203-206
- /
- 2000
Since the amplitude of voiced fall off at about -20dB/decade, dynamic range is often compressed prior to spectral analysis so that details at weak, high frequencies may be visible. Preemphasizing the speech, either by differentiating the analog speech s$\sub$a(t) prior to A/D conversion or by differencing the discrete-time s(n)=s$\sub$a(nT), compensating for falloff at high frequencies. The most common form of preemphasis is y(n)=s(n)-As(n-1), where A typically lies between 0.9 and 1.0 and reflects the degree of pre-emphasis. In This paper, we proposed that A is adjusted at each time by measuring the slope of envelope in frequency domain.
PDF

Lim, Ji-Sun
- Proceedings of the KAIS Fall Conference
- /
- 2010.05a
- /
- pp.253-256
- /
- 2010
Since the amplitude of voiced fall off at about -20dB/decade, dynamic range is often compressed prior to spectral analysis so that details at weak, high frequencies may be visible. Preemphasizing the speech, either by differentiating the analog speech $s_a$(t) prior to A/D conversion or by differencing the discrete-time s(n)=$s_a$(nT), compensating for falloff at high frequencies. The most common form of preemphasis is y(n)=s(n)-As(n-1), where A typically lies between 0.9 and 1.0 and reflects the degree of pre-emphasis. In this paper, we proposed that A is adjusted at each time by measuring the slope of envelope in frequency domain.
PDF

Kim, Dong-Jun;Park, Sang-Hui
- The Journal of the Acoustical Society of Korea
- /
- v.16 no.3
- /
- pp.18-25
- /
- 1997
This study describes a modified preemphasis formula, what we call energy-dependent preemphasis(EDP). This uses the normalized short-term energy of speech signal, with the assumption that the source characteristics of the glottal pulses and the radiation characteristics of the lips are approximately proportional to the energy of speech signal. Using this method, speech analyses, such as AR spectrum estimation, formant detection, are performed for nonstationary starting parts of 5 Korean single vowels. The results are compared with the conventional two preemphasis methods. We found that the proposed preemphasis gave enhanced spectral shapes and more accurate formant frequencies and avoided overlapping phenomenon of adjacent two formants.
PDF