JPS648935B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a digital √AGC method, and more specifically, the present invention relates to a digital √AGC method, and more specifically, a characteristic in which the amount of attenuation distortion substantially monotonically increases or decreases with respect to frequency (a √ characteristic that is well known as exhibiting this characteristic) Assuming that, below,
The present invention relates to a method for automatically compensating for distortion according to the amount of attenuation distortion on a transmission line having a characteristic (referred to as √ characteristic).

Due to recent developments in semiconductor technology, signal processing in the audio band is rapidly moving toward digital processing. Particularly in the field of communications, where the transfer characteristics between transmitter and receiver are a fundamental issue, digital filters and FFTs (fast Fourier transformers) using digital signal processing LSIs that perform real-time processing are used. Data modems are one of the devices that have been efficiently implemented into LSI by applying these digital processing technologies. This modem (especially 4800 ~
The 9600bps high-speed modem uses a Nyquist filter to shape the transmitted baseband digital signal, then shifts the frequency to the voice band (0.3KHz to 3.4KHz) and transmits it. At this time, various distortions occur depending on the line. Distortion on the line is what is called a line characteristic, and there are those that give attenuation and group delay within the band, and those that have an attenuation that depends on the square root of the frequency (asymmetric attenuation), such as unloaded cables. distortion). The former can be baseband equalized using an automatic equalizer after demodulation, but the latter is difficult to equalize when the distortion is large due to its asymmetry.
For this reason, a correction filter is inserted in the passband.

Conventionally, correction filters with fixed characteristics called fixed equalizers have been used to compensate for asymmetric attenuation distortion in voice band data transmission, so the number of correction filters can be increased or decreased depending on the transmission frequency. I needed to. Therefore, there is a problem that a complicated process of calculating the number of correction filters according to the transmission frequency is required, and the price increases as the number of correction filters increases.

In view of the problems in the prior art described above, an object of the present invention is to provide a correction filter for compensating for asymmetric attenuation distortion in voice band data transmission, based on the concept of changing the characteristics of the correction filter according to the line characteristics. The purpose is to reduce the cost of the correction filter by requiring only one.

In order to achieve the above object, the present invention provides a method for correcting attenuation distortion using a correction filter on a transmission path having a characteristic that the amount of attenuation distortion increases or decreases substantially monotonically with respect to frequency.
A parameter representing the amount of attenuation distortion in the transmission path is extracted, and according to the parameter, a coefficient that changes the characteristics of the correction filter is set based on a predetermined relational expression. Decide on a value that approaches -1 as
A digital √AGC method is provided which is characterized in that the characteristics of the correction filter are changed according to this coefficient.

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 is a schematic block circuit diagram of a correction filter applied to the present invention, and shows a first-order recursive filter. In FIG. 1, voice band data sent from a transmitting section (not shown) of the other party's modem is sent to a multiplier 1-1 as data x via a line 1.
The signal is inputted to the correction filter 2 via. data x
is generally subjected to various distortions by the line 1, so after being corrected by the correction filter 2,
The data y is delivered from the correction filter 2 to the modem receiving section (not shown). In the correction filter 2, the data passing through the adder 3 is inputted to the multiplier 5 through the feedback element 4, and in the multiplier 5,
According to the present invention, a correction value for distortion received on the line 1 is obtained by a correction coefficient b that changes depending on the transmission frequency,
By inputting this to the adder 3, distortion-corrected data y is obtained as the output of the adder 3. On the other hand, the correction coefficient 1-b is given to the multiplier 1-1,
Thereby, the DC gain of the correction filter 2 is kept constant. In the present invention, by changing the above-mentioned correction coefficient b according to the transmission frequency, distortion correction suitable for the transmission frequency is automatically performed. The transmission loss |y/x| is expressed as |y/x|=1/(Hb ² −2bcos2πk/N) ^1/2 . Here, N is the number of samples of the digital signal, and k is the wave number of the digital signal.

FIG. 2 is a graph showing the relationship between digital frequency and transmission loss when correction coefficient b is used as a parameter. As shown in FIG. 2, the larger the absolute value of the correction coefficient b, the smaller the transmission loss at frequency π, and the larger the transmission loss at frequencies 0 and 2π.

FIG. 3 is a graph showing the frequency characteristics when the characteristics of the correction filter 2 are superimposed on the received signal x of the sample value series. In FIG. 3, the audio band A is subjected to √ attenuation distortion near the frequency π, but this distortion is compensated by the characteristic B of the correction filter.

FIG. 4 is a block diagram showing one embodiment of a circuit for realizing the digital √AGC method according to the present invention. In FIG. 4, received data x is input to a correction filter 2 and a Δ detection circuit 41
is also input, and the √attenuation distortion amount Δ is detected in the Δ detection circuit 41. Detected √attenuation distortion amount Δ
The correction coefficient circuit 42 calculates the correction coefficient b based on
is determined, and this correction coefficient is given to the correction filter 2. As a result, output data y is obtained in which the output of the correction filter 2 is compensated for the √ attenuation distortion.

There are the following methods for extracting the √attenuation distortion amount Δ in the Δ detection circuit 41.

The difference between the signal components at both ends of the signal band is √
Let it be the amount of attenuation distortion. This method works if the signal is √
This is based on the fact that if the signal is subjected to attenuation distortion, the slopes of attenuation at both ends of the audio band A shown in FIG. 3, for example, should be different.

The difference between the original component of the signal when there is no distortion in the transmission path and the signal component at the edge of the band on the high band side is √
Let f be the amount of attenuation distortion.

Let the difference between the original component of the signal and the component of the received signal be the amount of attenuation distortion.

In the above and above, the signal component is basically electric power, but it may be anything that can represent the amount of distortion.

The correction coefficient b corresponding to the amount of attenuation distortion Δ is obtained as follows. That is, basically √
If the f-attenuation distortion amount Δ is zero, the correction coefficient b should be zero, and the larger the √ attenuation distortion amount Δ, the closer the correction coefficient b is to −1. √Attenuation distortion amount Δ
Since the correction coefficient b and the correction coefficient b can be determined in advance for various √ characteristics, they can also be determined by looking up a table in a read-only memory (ROM) or the like. However, if the capacity of the ROM is not sufficient, the correction coefficient b can also be obtained by directly calculating using some predetermined relational expression instead of using the ROM. Since this processing can be performed digitally, it can be treated as similar to the original functional processing. The above relational expression can generally be approximated by a higher-order expression of Δ as follows.

b _o = − _M 〓 ⁱ⁼⁰ α _i Δ ⁱ , 0≦|b _o |<1 As is clear from the above explanation, the present invention automatically corrects asymmetric attenuation distortion according to the frequency using a single correction filter. This is especially useful for voice band data transmission. Furthermore, if accurate suppression of the variation in the gain of each of the correction filters is required, it is also possible to absorb it in the flat AGC 43 provided afterwards.

[Brief explanation of the drawing]

Figure 1 is a schematic block circuit diagram of the correction filter applied to the present invention, Figure 2 is a graph showing the relationship between digital frequency and transmission loss when correction coefficient b is used as a parameter, and Figure 3 is a correction to the received signal. A graph showing the frequency characteristics when the filter characteristics are superimposed, and FIG. 4 shows the digital √
FIG. 2 is a block diagram showing one embodiment of a circuit for implementing the AGC method. 1: Line, 2: Correction filter, 3: Adder,
4: Feedback element, 5: Multiplier, A, A': Audio band,
B: Characteristics of correction filter, 41: √ Attenuation distortion amount detection circuit, 42: Correction complement circuit, b: Correction coefficient,
x: received signal, y: output signal.

Claims (1)

[Claims] 1. In a system in which the attenuation distortion is corrected by a correction filter on a transmission path having a characteristic that the amount of attenuation distortion substantially monotonically increases or decreases with respect to frequency, it represents the amount of attenuation distortion in the transmission path. A parameter is extracted, and a coefficient for changing the characteristics of the correction filter is determined according to the parameter to a value that is zero when the parameter is zero and approaches -1 as the parameter becomes larger. A digital √AGC method characterized by changing the characteristics of the correction filter according to the coefficient. 2. The digital √AGC method according to claim 1, wherein the parameter representing the amount of attenuation distortion is a difference between a signal component at a high end of a signal band and a signal component at a low end of a signal band. 3. The digital √AGC method according to claim 1, wherein the parameter representing the amount of attenuation distortion is a difference between a high band component of a signal band and a component originally included in the signal. 4. The digital √AGC method according to claim 1, wherein the parameter representing the amount of attenuation distortion is a difference between a component of a received signal and a component originally included in the signal. 5. The digital √AGC system according to claim 1, wherein the correction filter is a high-pass digital filter for √ compensation.