http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
개별검색 DB통합검색이 안되는 DB는 DB아이콘을 클릭하여 이용하실 수 있습니다.
통계정보 및 조사
예술 / 패션
<해외전자자료 이용권한 안내>
- 이용 대상 : RISS의 모든 해외전자자료는 교수, 강사, 대학(원)생, 연구원, 대학직원에 한하여(로그인 필수) 이용 가능
- 구독대학 소속 이용자: RISS 해외전자자료 통합검색 및 등록된 대학IP 대역 내에서 24시간 무료 이용
- 미구독대학 소속 이용자: RISS 해외전자자료 통합검색을 통한 오후 4시~익일 오전 9시 무료 이용
※ 단, EBSCO ASC/BSC(오후 5시~익일 오전 9시 무료 이용)
A method on the hardware implementation of the Finite Impulse Response (FIR) digital filter using Digital Quarter Square Multiphcation (DQSM) algorithm is proposed. ROM requirement can be reduced by using the double precision algorithm and the absolute value circuit, and also execution speed can be improved by reducing logic level steps of absolute value circuit. The frequency response of FIR digital filter implemented by the above method is quite agreeable with the frequency response simulated by Remez exchange algorithm. If coefficients should be stored in the RAM instead of the ROM, this method is especially useful in the case of FIR digital filter with time varying coefficients.
This paper describes an analysis method for the pitch periods of Korean /consonant+vowel/type pronounced by a male speaker. After sampling the voiced speech at a sampling rate of 10㎑, we stored these data in the memory of the microcomputer through the 8 bit A/D converter. An algorithm was used for center-infinite peak clipping and simplified autocorrelation analysis to detect the pitch period. As a result of experiment, it shows that the pitch period of Korean speech is changed variously depending on the kinds of consonant and vowel.
This paper proposes an approach to the approximate realization of two-dimensional FIR digital filters by using the decomposition of their coefficient matrices. This approach is based on the design method of multistage separable planar filters, and reduces with very small error the number of required computations(number of multiplications/additions), register coefficients and memories applying the intimate relationship between the eigen values of every stage and ranks of the matrices.
A method on the implementation of FIR(Finite Impulse Response) digital filter using stored square ROM multiplier is proposed. This paper describes that ROM requirement can be reduced by using the absolute value circuit and split address technique, and also execution speed can be improved by using the pipelining technique for stored square ROM multiplier. The frequency response of FIR digital filter simulated and implemented by the above method is quite agreeable with theoretical results computed by Remez exchange algorithm. All results stated assume the use of fixed point parallel arithmetic.
In implementing a digital filter with fixed-point arithemetic, additions may cause overflows to occur. In order to avoid overflows, the input signal may be attenuated suitably so that no overflows occur at any point in the filter. But this results in a low output signal-to-noise ratio. Lp Norm is the theoretical basis for scaling the signal level at different node in a digital filter to avoid overflows. The purpose of this paper is to describe a systematic procedure for scaling wave digital filter to avoid overflow problems while maintaining a signal-to-noise ratio as large as possible. The scaled WDF have an improved signal-to-noise ratio over unscaled WDF without overflow.