KR0152012B1

KR0152012B1 - Method and apparatus of adaptive interpolation for digital signal according to signal

Info

Publication number: KR0152012B1
Application number: KR1019920005946A
Authority: KR
Inventors: 정제창
Original assignee: 김광호; 삼성전자주식회사
Priority date: 1992-04-10
Filing date: 1992-04-10
Publication date: 1998-10-15
Also published as: KR930022883A

Abstract

본 발명은, 취사과정에 의해 다운샘플링되어 있는 디지탈시퀀스신호를 보간하여 원신호를 재현하는 디지탈신호의 보간방법 및 그 장치에 관한 것이다. 본 보간방법 및 장치에서는, 1차 라그랜지보간과 3차 라그랜지보간을 결합계수를 사용하여 결합하고, 그 결합계수를 신호특성에 따라 가변시킨다. 이에 의해 단순한 과정을 통해 원신호를 적응적으로 충실히 재현할 수 있다.The present invention relates to a digital signal interpolation method and apparatus for reproducing an original signal by interpolating a down-sampled digital sequence signal by a cooking process. In the present interpolation method and apparatus, the first and second lagrange interpolation are combined using a coupling coefficient, and the coupling coefficient is varied according to signal characteristics. As a result, the original signal can be adaptively and faithfully reproduced through a simple process.

Description

Adaptive Interpolation Method and Interpolation Device of Digital Signal According to Signal Characteristics

제1도는 종래의 취사(Decimation)/보간(Interpolation)에 의한 디지탈신호처리 시스템의 블록도.1 is a block diagram of a conventional digital signal processing system by Decimation / Interpolation.

제2도는 제1도의 로우패스필터의 특성도로서, (a)는 주파수 영역전달함수 및 (b)는 시간영역임펄스응답.2 is a characteristic diagram of the low pass filter of FIG. 1, (a) is a frequency domain transfer function, and (b) is a time domain impulse response.

제3도는 1차 라그랜지보간(a) 및 3차 라그랜지보간(b)의 설명도.3 is an explanatory diagram of a first lagrange interpolation (a) and a third lagrange interpolation (b).

제4도는 본 발명에 따른 보간장치의 일실시예의 개략적 블록도.4 is a schematic block diagram of an embodiment of an interpolation apparatus according to the present invention.

제5도는 결합계수결정부를 포함하는 취사기(Decimator)의 블록도.5 is a block diagram of a cooker including a coupling coefficient determination unit.

* 도면의 주요부분에 대한 부호의 설명* Explanation of symbols for main parts of the drawings

10, 70 : 취사기 11, 22, 71 : 로우패스필터10, 70: Cooker 11, 22, 71: Low Pass Filter

12 : 다운샘플링부 20 : 보간기12: downsampling unit 20: interpolator

21 : 엎샘플링부 30 : 지연부21: sampling unit 30: delay unit

31, 32, 33 : 지연기 40 : 1차 라그랜지보간부31, 32, 33: delay 40: first lagrange interpolator

41, 42, 51, 52, 53, 54 : 증폭기41, 42, 51, 52, 53, 54: amplifier

43, 55, 56, 57 : 가산기 50 : 3차 라그랜지보간부43, 55, 56, 57: Adder 50: 3rd Lagrange Interpolator

61, 62 : 승산기 63 : 가산기61, 62: multiplier 63: adder

64 : 멀티플렉서 75 : 결합계수결정부64: multiplexer 75: coupling coefficient determination unit

본 발명은, 취사(取捨, Decimation)과정에 의해 다운샘플링되어 있는 디지탈시퀀스신호를 보간(Interpolation)하여 원신호를 재현하는 디지탈신호의 보간방법 및 그 장치에 관한 것이다.BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital signal interpolation method and apparatus for interpolating down-sampled digital sequence signals by a deciding process to reproduce original signals.

음성 영상 등의 디지탈신호처리기법중의 하나로서 시퀀스신호의 취사/보간방법이 널리 이용되고 있다. 이 방법에서는, 취사과정에서 신호의 샘프링율을 줄여 데이터량을 감소시켜 전송하는 한편, 보간과정에서 신호의 샘플링율을 늘려 샘플들사이의 값들을 보간해줌으로써 원신호를 충실히 재현한다. 이러한 취사/보간은, 음성이나 영상신호 등의 많은 디지탈신호처리에 있어서 데이터량을 증감시키기 위해서나, 서로 다른 클럭으로 동작하는 멀티레이트(Multirate)시스템에서 상호 동기를 일치시키기 위해서도 이용되고 있다.As one of digital signal processing techniques such as an audio image, a sequence signal cooking / interpolation method is widely used. In this method, the sampling rate of the signal is reduced to reduce the amount of data during the cooking process, while the sampling rate of the signal is increased during interpolation to faithfully reproduce the original signal by interpolating values between samples. Such cooking / interpolation is also used to increase or decrease the amount of data in many digital signal processing such as audio and video signals or to synchronize mutual synchronization in a multirate system operating with different clocks.

종래의 취사/보간에 의한 디지탈신호처리시스템을 나타낸 제1도를 참조하면, 취사기(10)에서의 다운샘플링과정(12)은 예를 들어 M:1의 경우 M개의 입력샘플이 들어 올 때마다 하나의 샘플을 취하며, 반대로 보간기(20)에서의 1:M 업샘플링(21)에서는 각 입력샘플들사이에 (M-1)개의 0을 등간격으로 넣어 하나의 입력이 들어 올 때 마다 M개의 출력이 나오도록 한다. 이때 취사기(10)에서는 다운샘플링되는 고주파성분이 저주파성분처럼 재생되는 소위 얼라이어징(aliasing)을 방지하기 위하여 입력신호중의 고주파성분을 미리 차단하는 로우패스필터(11)를 사용하며, 보간기(20)에서도 저주파성분이 고주파와 저주파의 합성신호처럼 보이게 되는 소위 이미징(imaging) 현상을 방지하기 위해 로우패스필터(22)를 사용하여 고주파성분을 차단한다.Referring to FIG. 1, which shows a conventional signal processing / interpolation digital signal processing system, the downsampling process 12 in the cooker 10 is performed when M input samples are input in the case of M: 1. On the contrary, in the 1: M upsampling 21 in the interpolator 20, (M-1) zeros are equally spaced between each input sample when one input comes in. Get M outputs every time. In this case, the cooker 10 uses a low pass filter 11 which blocks the high frequency components in the input signal in advance in order to prevent so-called aliasing in which down-sampled high frequency components are reproduced like low frequency components. Also at (20), the low frequency component is cut off by using the low pass filter 22 to prevent a so-called imaging phenomenon in which the low frequency component looks like a composite signal of high frequency and low frequency.

이러한 취사기(10)와 보간기(20)에 사용되는 로우패스필터(11, 22)는 모두 제1도(a)와 같은 주파수영역전달함수를 가지며, 이에 해당하는 시간영역임펄스응답은 제2도(b)와 같은 싱크(sinc) 함수의 모양을 갖는다. 흔히 이 싱크함수에 윈도우(window)를 씌워 유한임펄스응답(FIR)필터를 구현한다. 그런데 특히 보간기에 있어서는, 제2도(b)와 같은 싱크보간기에 윈도우를 씌운다 하더라도 대개의 경우 필터의 탭수가 길어져 보간과정에서의 복잡도가 커지게 된다는 문제점이 있다.The low pass filters 11 and 22 used in the cooker 10 and the interpolator 20 both have a frequency domain transfer function as shown in FIG. 1A, and the corresponding time domain impulse response is shown in FIG. It has the shape of a sink function as shown in FIG. Often, this sink function is windowed to implement a finite impulse response (FIR) filter. However, especially in the interpolator, even if the window is covered by the sink interpolator as shown in FIG. 2 (b), the number of taps of the filter is generally long, which increases the complexity in the interpolation process.

보다 단순화된 보간기로서는, 보간하고자 하는 점주위의 복수입력신호들을 연결하는 1차 혹은 3차 등의 다항식을 구하여 이 다항식으로부터 보간값을 구하는 라그랜지(Lagrange)보간기가 공지되어 있다.As a simpler interpolator, a Lagrange interpolator which obtains an interpolation value from a polynomial such as a first order or a third order connecting multiple input signals around a point to be interpolated is known.

제3도(a) 및 (b)는 각각 2:1 업샘플링의 1 경우 1차 및 3차 라그랜지보간의 예를 도시한 것으로서, 편의상 홀수번째의 샘플들이 주어진 입력값이고 짝수번째 샘플들을 보간한다고 가정한다. 제3도(a)에서의 1차(Linear)라그랜지보간은 이웃하는 두 홀수번째의 샘플을 연결하는 1차 곡선(ax + b)을 구하여 해당보간위치에서의 보간값을 구한다. 즉 1차 라그랜지보간에서는 인접한 두 샘플의 평균치를 취하는 것이다. 결국 1차 보간값 fL(0)는, fL(0)=1/2[f(-1)+f(1)]로 표현되게 된다. 여기서 f(-1), f(1)은 주어진 입력값이다.3 (a) and 3 (b) show examples of 1st and 3rd order lagrangian interpolation in the case of 2: 1 upsampling, respectively, and for convenience, odd-numbered samples are given input values and even-numbered samples are interpolated. Assume that The linear Lagrangian interpolation in FIG. 3 (a) obtains an interpolation value at the corresponding interpolation position by obtaining a first-order curve (ax + b) connecting two neighboring odd-numbered samples. In other words, in the first lagrange interpolation, the average of two adjacent samples is taken. As a result, the primary interpolation value fL (0) is represented by fL (0) = 1/2 [f (-1) + f (1)]. Where f (-1) and f (1) are given input values.

한편, 3차(Cubic)라그랜지보간에서는 이웃하는 네 개의 홀수번째 샘플들을 연결하는 3차곡선( ax³+ bx²+ cx + d)를 구하여 보간하고자 하는 위치에서의 값을 취하게 된다. 즉 3차 보간값 fc(0)는, fc(0)=1/16[-f(3)+9f(-1)+9f(1)-f(3)]과 같이 얻어지게 된다.In the cubic Lagrangian interpolation, a cubic curve (ax ³ + bx ² + cx + d) connecting four neighboring odd-numbered samples is obtained to take a value at a position to be interpolated. That is, the third interpolation value fc (0) is obtained as fc (0) = 1/16 (-f (3) + 9f (-1) + 9f (1) -f (3)).

그런데, 전술한 1차 라그랜지보간은 매우 간단하다는 장점이 있지만, 보간되어지는 곡선이 매끄럽지 못하고, 3차 라그랜지보간은 매끄러운 곡선으로 보간되지만 그 과정이 복잡하고, 또 계단파의 경우에는 오버슈터(overshoot)와 언더슈트(undershoot)가 발생하게 된다는 문제점이 있다.By the way, although the above-described primary Lagrange interpolation has an advantage of being very simple, the interpolated curve is not smooth, and the third-order Lagrange interpolation is interpolated with a smooth curve, but the process is complicated, and in the case of a step wave, the overshooter (overshoot) and undershoot (undershoot) is a problem that occurs.

본 발명은, 전술과 같은 종래 보간기의 문제점을 해결하기 위해 안출된 것으로서, 1차 라그랜지보간과 3차 라그랜지보간을 적절히 결합하고, 그 결합계수를 신호의 특성에 따라 변화시킴으로써 원신호를 충실히 재현할 수 있는 적응적 보간방법 및 이 방법의 실시를 위한 장치를 제공하는 것을 목적으로 한다.The present invention has been made to solve the problems of the conventional interpolator as described above, by properly combining the first and third lagrange interpolation, and by changing the coupling coefficient according to the characteristics of the signal to produce the original signal An object of the present invention is to provide an adaptive interpolation method that can be faithfully reproduced and an apparatus for implementing the method.

상기 목적은, 본 발명에 따라, 취사과정에 의해 다운샘플링되어 있는 디지탈시퀀스신호를 보간하여 원신호를 재현하는 디지탈신호의 보간방법에 있어서, 다운샘플링되어 입력된 디지탈시퀀스신호의 1차 라그랜지보간단계, 상기 입력디지탈시퀀스신호의 3차 라그랜지보간단계, 상기 1차 라그랜지보간 및 3차 라그랜지보간의 상호 결합계수를 결정하는 단계, 상기 결합계수 및 상기 결합계수의 1에 대한 여수(餘數)를 상기 1차 및 3차 라그랜지보간단계로부터의 해당시퀀스신호와 각각 승산하는 단계, 상기 각 승산단계로부터의 시퀀스신호를 상호 가산하는 단계, 및 상기 가산된 신호와 소정시간 지연된 입력디지탈시퀀스신호를 소정의 클럭에 응하여 교호적으로 출력하는 내삽단계를 포함하는 것을 특징으로 하는 신호특성에 따른 디지탈신호의 적응적 보간방법에 의해 해결된다.The above object is, according to the present invention, in the interpolation method of a digital signal which reproduces an original signal by interpolating a down-sampled digital sequence signal by a cooking process, the first-order lagrang interpolation of the down-sampled input digital sequence signal Step 3, determining the third-order Lagrangian interpolation of the input digital sequence signal, determining the mutual coupling coefficient between the first-order Lagrangian interpolation and the third-order Lagrangian interpolation, ) Is multiplied by the corresponding sequence signals from the first and third lagrange interpolation steps, the sequence signals from the respective multiplication steps are mutually added, and the input digital sequence signal delayed by the added signal by a predetermined time. And adaptively outputting the digital signal according to the signal characteristic, characterized in that it comprises an interpolation step of alternately outputting the signal in response to a predetermined clock. It is solved by interpolation method.

최적의 결합계수는, 바람직하게 취사기에서 다운샘플링전 상태의 시퀀스신호와 상기 보간완료된 출력신호간의 배교에 의해 결정되어 상기 각 승산기로 전송되며, 이때 상기 보간완료된 출력신호는 별도의 보간과정에서 얻어지는 것이 더욱 바람직하다. 그리고 가장 근사한 결합계수를 구하기 위해 평균자승오차법이나 평균절대오차법이 사용될 수 있다.The optimum coupling coefficient is preferably determined by cross-linking between the sequence signal in the pre-sampling state and the interpolated output signal in the cooker and transmitted to each multiplier, wherein the interpolated output signal is obtained in a separate interpolation process. More preferred. And the mean square error method or the mean absolute error method can be used to find the closest coupling coefficient.

그리고 보간과정으로 전송되는 결합계수는 시퀀스신호를 일정한 블록단위로 나누어 매 블록마다 한 번씩 내보내거나, 혹은 결합계수가 크게 변동될 경우에만 내보낼 수도 있다.In addition, the coupling coefficient transmitted through interpolation may be divided into a predetermined block unit and exported once every block, or only when the coupling coefficient is greatly changed.

한편, 전술한 본 발명의 보간방법을 실시에 적합한 보간장치는, 다운샘플링되어 입력된 디지탈시퀀스신호를 단계적으로 지연시키는 복수의 지연기, 상기 단계적으로 지연된 신호를 1차 라그랜지보간하는 제1보간기, 상기 단계적으로 지연된 신호를 3차 라그랜지 보간하는 제2보간기, 상기 1차 라그랜지보간 및 3차 라그랜지보간의 상호 결합계수를 결정하는 결합계수결정부, 상기 결합계수 및 상기 결합계수의 1에 대한 여수를 상기 제1 및 제2 보간기로부터의 해당 시퀀스신호와 각각 승산하는 제1 및 제2승산기, 상기 각 승산기로부터의 시퀀스신호를 상호 가산하는 가산기, 및 상기 가산된 신호와 소정시간 지연된 입력디지탈시퀀스신호를 소정의 클럭에 응하여 교호적으로 출력하는 멀티플렉서를 포함한다.On the other hand, an interpolation apparatus suitable for implementing the above-described interpolation method of the present invention includes a plurality of delayers for stepwise delaying downsampled and input digital sequence signals, and first interpolation for first-order lagrange interpolation of the stepwise delayed signals. And a second interpolator for interpolating the stepped delay signal to a third lagrange interpolation, a coupling coefficient determining unit for determining a mutual coupling coefficient between the first lagrange interpolation and the third lagrange interpolation, the coupling coefficient and the coupling coefficient. First and second multipliers for multiplying the free number for 1 by the corresponding sequence signals from the first and second interpolators, an adder for mutually adding sequence signals from each of the multipliers, and the added signal and the predetermined signal. And a multiplexer for alternately outputting a time delayed input digital sequence signal in response to a predetermined clock.

이하에서 일실시예에 관한 첨부도면을 참조하여 본 발명을 상세히 설명한다.Hereinafter, with reference to the accompanying drawings, an embodiment of the present invention will be described in detail.

제4도는 설명 및 이해를 용이하게 하기 위해 2:1 업샘플링으로 보간을 수행하는 본 발명에 따라 보간장치를 블록도로서 도시한 것이다. 본 도면에서 볼 수 있는 바와 같이, 본 보간장치는, 대체적으로 다수의 순차적 지연소자(31, 32, 33)를 구비한 지연부(30)와, 지연신호를 각각 1차 및 3차 라그랜지보간을 수행하는 제1 및 제2보간기(40, 50), 1차 및 3차 라그랜지보간부(40, 50)의 출력 시퀸스신호를 결합계수(a)로 결합하는 제1 및 제2승산기(61, 62)와 가산기(63) 및 가산기(63)에서 가산된 신호와 소정시간 지연된 입력 디지탈시퀸스신호를 교호적으로 출력하는 멀티플렉서(64)로 이루어진다.4 is a block diagram of an interpolator in accordance with the present invention for performing interpolation with 2: 1 upsampling to facilitate explanation and understanding. As can be seen from the figure, the interpolator includes a delay unit 30 having a plurality of sequential delay elements 31, 32, and 33, and first and third lagrange interpolation of delay signals, respectively. First and second multipliers for combining the output sequence signals of the first and second interpolators 40 and 50 and the first and third lagrange interpolators 40 and 50 into a coupling coefficient (a). 61 and 62, an adder 63, and a multiplexer 64 alternately outputting the signal added by the adder 63 and the input digital sequence signal delayed by a predetermined time.

제1보간기(40)는 지연부(30)의 제1 및 제2지연소자(31, 32)에서 순차지연된 입력시퀸스신호를 각각 0.5배 증폭기(41, 42)에서 감폭하여 가산기(43)에서 가산한 다음 출력한다. 즉1차 라그랜지보간에 의해 보간될 위치의 전후 샘플을 평균한 보간값을 제공한다.The first interpolator 40 attenuates the input sequence signals sequentially delayed by the first and second delay elements 31 and 32 of the delay unit 30 by 0.5 times the amplifiers 41 and 42, respectively, and adds them to the adder 43. Add and print. That is, an interpolation value obtained by averaging the before and after samples of the positions to be interpolated by the first lagrange interpolation is provided.

제2보간기(40)는 보간될 위치 전후의 네 개 샘플을 지연전 상태에서 및 제1내지 제3지연소자(31 내지 33)에서 지연시켜 각각 -1/16, 9/16, 9/16, -1/16배 증폭기(51 내지 54)에서 감폭한후, 각 가산기(55 내지 57)에서 가산하여 3차 라그랜지보간값을 출력한다.The second interpolator 40 delays four samples before and after the position to be interpolated in the pre-delay state and in the first to third delay elements 31 to 33, respectively, -1/16, 9/16, and 9/16. , -1/16 times amplifiers amplify (51 to 54), and then add in each adder (55 to 57) to output the third-order Lagrangian interpolation value.

제1 및 제2보간기(40, 50)에서 각각 1차 및 3차 라그랜지보간된 신호fL(n)은, 후술한 결합계수결정부에서 전송되어온 결합계수(a)에 의해 결합된다. 0~1사이의 값을 가지는 결합계수(a)가 1차 라그랜지보간신호 fL(n)에 곱하여지고, 결합계수(a)의 1에 대한 여수(餘數) 즉 (1-a)는 3차 라그랜지보간신호 fc(n)에 곱하여져 가산기(63)에서 가산되어 결합보간신호 F(n)로서 출력된다.The first and third lagrange interpolated signals fL (n) of the first and second interpolators 40 and 50, respectively, are combined by the coupling coefficient a transmitted from the coupling coefficient determination unit described later. The coupling coefficient (a) having a value between 0 and 1 is multiplied by the first-order Lagrange interpolation signal fL (n), and the free power of 1 of the coupling coefficient (a), that is, (1-a) is the third order. The Lagrange interpolation signal fc (n) is multiplied and added by the adder 63 and output as the combined interpolation signal F (n).

이때, 1차 라그랜지보간신호 {fL (n)와 3차 라그랜지보간신호 [fc(n)를 먼저 감산하고 감산된 신호[fl(n)-fc(n)]를 상기 결합계수와 승산한 다음, 승산된 신호 (a×[fL(n)-fc(n)])와 상기 3차 라그랜지보간신호 fc(n)를 가산하여도 동일한 결과를 얻을 수 있다.In this case, the first Lagrange interpolation signal {fL (n) and the third Lagrange interpolation signal [fc (n) are first subtracted and the subtracted signal [fl (n) -fc (n)] is multiplied by the coupling coefficient. Next, the same result can be obtained by adding the multiplied signal a × [fL (n) -fc (n)] and the third order lagrangian interpolation signal fc (n).

멀티플랙서(64)는 결합보간신호F(n)외에 제 2지연소자(32)에서 지연된 입력시퀸스신호를 입력받아 이들을 교대로 스위칭하여 출력한다. 즉 2:1의 업샘플링이므로 제1 혹은 제2보간기(40, 50)에서의 클럭의 2배에 해당하는 클럭에 응하여 스위칭함으로써 보간될 신호를 입력시퀸스신호에 내삽함으로써, 보간완료된 시퀸스신호 y(n)를 출력하게 된다.The multiplexer 64 receives an input sequence signal delayed by the second delay element 32 in addition to the combined interpolation signal F (n) and alternately switches them and outputs them. That is, since it is 2: 1 upsampling, the interpolated sequence signal y is interpolated by interpolating the signal to be interpolated to the input sequence signal by switching in response to a clock corresponding to twice the clock of the first or second interpolators 40 and 50. will print (n).

제 5도는 1차 라그랜지보간신호 {fL(n)와 3차 라그랜지보간신호fc(n)을 결합하기 위한 결합계수(a)를 결정하는 결합계수결 정부(75)를 포함한 취사기(70)를 나타낸 것이다. 본 도면에서 볼 수 있는 바와 같이, 취사될 입력신호 {∝(n)}는 로우패스필터(71)를 거쳐 2:1 다운샘플링(72)한 후 취사된 신호 f(n)로 전송된다.5 is a cooker 70 including a coupling coefficient determining unit 75 for determining a coupling coefficient (a) for combining the primary lagrange interpolation signal {fL (n) and the tertiary lagrange interpolation signal fc (n). ). As can be seen in the figure, the input signal {∝ (n)} to be cooked is transmitted to the picked signal f (n) after 2: 1 downsampling 72 via the low pass filter 71.

이때 로우패스필터링된 신호 yo(n)는 지연기(73)를 거쳐 결합계수결정부(75)로 제공되고, 동시에 다운샘플링된 신호 f(n)는 전술한 제4도의 보간장치와 동일한 보간기(74)를 거쳐 보간완료된 신호y(n)로서 결합계수결정부(75)에 입력된다.At this time, the low pass filtered signal yo (n) is provided to the coupling coefficient determiner 75 through the delay unit 73, and at the same time, the downsampled signal f (n) is the same as the interpolator of FIG. The interpolation completion signal y (n) is input to the coupling coefficient determination unit 75 via 74.

결합계수결정부(75)는 로우패스필터링된 원신호 yo(n)의 지연신호와 보간된 신호 y(n)를 상호 비교하여 최적의 결합계수(a)를 결정한다. 이때 보간기(74)에는 0~1의 결합계수(a)값이 적절한 구간 예를 들어 0.1단위로 입력되며, 결합계수결정부(75)는 평균자승오차법이나 평균절대오차법 등의 방법에 의해 보간기(74)의 출력과 지연된 원신호 yo(n)를 비교하여 최소의 오차로 되는 결합계수(a)값을 선택하여 출력한다.The coupling coefficient determination unit 75 compares the delayed signal of the low-pass filtered original signal yo (n) with the interpolated signal y (n) to determine an optimal coupling coefficient a. In this case, the coupling coefficient (a) of 0 to 1 is input to the interpolator 74 in an appropriate section, for example, 0.1 unit, and the coupling coefficient determining unit 75 is applied to a method such as the mean square error method or the average absolute error method. As a result, the output of the interpolator 74 is compared with the delayed original signal yo (n) to select and output a coupling coefficient (a) value that is the minimum error.

결합계수결정부(75)에서는 입력시퀸스신호를 등간격으로 구획하여 일정주기로 전송하거나, 결합계수(a)의 값에 큰 변동이 있는 경우에만 전송하도록 할 수도 있다.The coupling coefficient determination unit 75 may divide the input sequence signal at equal intervals and transmit the signal at a predetermined period or only when there is a large variation in the value of the coupling coefficient a.

이상의 도시 및 설명한 실시예에서는 2:1 업샘플링의 경우를 예로서 설명하였지만 임의의 M:1 업샘플링에 대해서도 본 발명이 적용될 수 있음은 물론이다. 또한 전술 실시예에서는 1차원 데이터를 예로 들어 설명하였으나 2차원 이상의 데이터에 대하여도 마찬가지로 적용된다.Although the illustrated and described embodiments have described the case of 2: 1 upsampling as an example, the present invention can be applied to any M: 1 upsampling. In addition, in the above embodiment, the one-dimensional data has been described as an example, but the same applies to the two-dimensional or more data.

이상 설명한 바와 같이, 본 발명에 따른 보간방법 및 보간장치는, 1차 라그랜지보간과 3차 라그랜지보간을 결합계수를 사용하여 결합하고, 그 결합계수를 신호특성에 따라 가변시킴으로써 단순한 과정을 통해 원신호를 적응적으로 충실히 재현할 수 있다는 우수한 효과를 제공한다.As described above, the interpolation method and the interpolation apparatus according to the present invention combine the first and third lagrange interpolation using the coupling coefficient, and change the coupling coefficient according to the signal characteristics. It provides an excellent effect that the original signal can be adaptively and faithfully reproduced.

Claims

A digital signal interpolation method of reproducing an original signal by interpolating a downsampled digital sequence signal by a cooking process, comprising: a first-order Lagrangian interpolation step of the downsampled input digital sequence signal, and three of the input digital sequence signal Interpolating signal combining the first and third lagrange interpolated signals, and outputting the first and third lagrange interpolated signals by the coupling coefficients; And an interpolation step of alternately outputting the combined interpolated signal and the input digital sequence signal delayed by a predetermined time in response to a predetermined clock.

2. The adaptation of a digital signal according to a signal characteristic according to claim 1, wherein the coupling coefficient is determined by a comparison between the sequence signal of the down-sampling state and the interpolated output signal in the cooker and transmitted to each of the multipliers. Ever interpolation method.

The method of claim 2, wherein the interpolated output signal is obtained in a separate interpolation process.

3. The method of claim 2, wherein the coupling coefficient is determined by a mean square error method.

The method of claim 2, wherein the coupling coefficient is determined by an average absolute error method.

The adaptive interpolation method according to any one of claims 1 to 5, wherein the coupling coefficient is transmitted in a cycle of a predetermined sequence block unit.

6. The method according to any one of claims 1 to 5, wherein said coupling coefficient is transmitted only when the change is large.

2. The method of claim 1, wherein in the interpolation signal combining step, each of the multiplied sequence signals is multiplied by multiplying the coupling coefficient and a free number of one of the coupling coefficients with the first and third order lagrange interpolated signals, respectively. Adaptive interpolation method of a digital signal according to the signal characteristics, characterized in that.

The method of claim 1, wherein in the interpolation signal combining step, one signal of the first and third lagrange interpolated signals is subtracted by another signal to be multiplied by the coupling coefficient, and the multiplied sequence signal and the other signal are mutually reduced. Adaptive interpolation method of digital signal according to signal characteristics, characterized in that the addition.

A digital signal interpolator for reproducing an original signal by interpolating a digital sequence signal downsampled and transmitted by a cooker, the apparatus comprising: a plurality of delayers for delaying the downsampled input digital sequence signal step by step; A coupling coefficient for determining a mutual coupling coefficient between the first interpolator for performing the first lagrange interpolation, the second interpolator for the third delayed lagrange interpolation, and the first and second lagrange interpolation. Decision unit; An interpolation signal combiner for combining the first and third lagrange interpolated signals with the coupling coefficients and outputting the combined signal, and a multiplexer alternately outputting the added signal and an input digital sequence signal delayed by a predetermined time in response to a predetermined clock. Adaptive interpolation device according to the signal characteristics, characterized in that it comprises a.

11. The adaptive interpolation of digital signals according to the signal characteristic of claim 10, wherein the coupling coefficient determiner determines and transmits the coupling coefficient by comparing a sequence signal in a state before downsampling with the output signal which has been interpolated. Device.

12. The apparatus of claim 11, wherein the interpolated output signal is input to the coupling coefficient determiner from a separate interpolation process.

12. The apparatus of claim 11, wherein the coupling coefficient determination unit determines the coupling coefficient by an average square error method.

12. The apparatus of claim 11, wherein the coupling coefficient determination unit determines the coupling coefficient by an average absolute error method.

15. The adaptive interpolation apparatus according to any one of claims 10 to 14, wherein the coupling coefficient determination unit transmits the coupling coefficient at a period of a predetermined sequence block unit.

15. The adaptive interpolation apparatus according to any one of claims 10 to 14, wherein the coupling coefficient determining unit transmits the coupling coefficient only when the change of the coupling coefficient is large.

The multiplier of claim 10, wherein the interpolation signal combiner multiplies each of the multiplied sequence signals by multiplying a coupling coefficient and a free number corresponding to one of the coupling coefficients by corresponding sequence signals from the first and third lagrange interpolation steps, respectively. Adaptive interpolation method of digital signal according to signal characteristics, characterized in that the addition.

The interpolation signal combiner may subtract one of the corresponding sequence signals from the first and third lagrange interpolation steps to another signal to multiply the combination coefficient, and add the multiplied sequence signal and the other signal to each other. Adaptive interpolation method of digital signal according to signal characteristics.