JP3091574B2 - Data conversion circuit and data conversion method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital data conversion circuit and a conversion method for obtaining output data having an exponential function with respect to input data.

[0002]

2. Description of the Related Art In recent years, a technique of converting an analog signal into a digital signal and performing signal processing by a DSP (digital signal processor) has been actively performed. For example, when a DSP is used in an audio device, processing of a graphic equalizer, reverberation, and the like can be easily performed digitally without deterioration in sound quality.

[0003] When processing an audio signal, a process of converting an input signal into an output signal having an exponential function relationship is often performed. In order to perform such a conversion, the exponential function must be calculated based on an approximate expression such as Taylor approximation or Chebyshev approximation. However, this calculation requires a large number of program steps and the processing time is long,
The effect on other programs was significant.

Therefore, conventionally, a method of referring to a ROM table without performing the above-described approximate calculation has been adopted.
As shown in FIG. 4, the possible range of the input data is finely and evenly divided, and the minimum value (or the average value) of the exponential function curve for each divided input is stored in the memory using the input data as an address. In this method, the conversion data is read from the ROM table using the input data to be converted as an address. However, according to this method, the conversion can be easily performed. However, as shown in FIG. 4, since the input data within one divided area are all converted into the same data, the conversion value of the true exponential function is obtained. Is large, especially when the input data is large. For this purpose, the number of divisions of the input data must be further reduced.
There is a dilemma that the utilization efficiency of OM deteriorates.

Therefore, as a further improved method, FIG.
Method was conceived. In this method, the range of input data can be divided into N, and points of the exponential function curve corresponding to the division points are connected by straight lines and approximated. That is, the straight lines in each of the divided ranges are represented as Y _i = a _i X + b _i (i = 1 · 2... N), and the coefficients a _i and b _i of the N straight lines are stored in the ROM. . Then, in the conversion reads the coefficients a _i and the coefficient b _i corresponding to the input data from the ROM, Y = a _i
A method of obtaining the conversion calculates and outputs X + b _i. According to this method, the error between the exponential function curve and the approximate straight line is reduced even if the number of divisions of the range that the input data can take is reduced, and the number of coefficients stored in the ROM is also reduced.

[0006]

According to the data conversion method shown in FIG. 5, the number of coefficients to be stored in the ROM table is certainly smaller than in the prior art. When the number of divisions is 32, the number of coefficients is 64, and the number is still large.

[0007]

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and divides a range which can be taken by input data into N ranges, and a coefficient corresponding to each of the divided ranges. c _j (j = 1 · 2... N) is further set, each divided range is divided into M, and an exponential function curve corresponding to the divided range is represented by a straight line Y = c _j ( a _i X + b _i ) (i
= 1,2... M), and the coefficients c _j , a _i ,
And a memory for storing the coefficient b _i , and the coefficient c _j , coefficient a _i , and coefficient b _i according to input data from the memory.
Comprising an address counter for reading out the coefficient c _j read from the memory, the coefficient a _i, and, a calculation means for calculating a _{Y = c j (a i X} + b i) on the basis of the input data and the coefficient b _i A data conversion circuit is provided.

Further, using the data conversion circuit, the uppermost p bits of the input data are set in the address counter to read out the coefficient _cj, and the uppermost p bits of the input data are read out.
After setting the next r bits from the bits in the address counter and reading the coefficients a _i and b _i , the data of the remaining bits of the input data and the read coefficients c _j , coefficients a _i , and, by a factor b _i Y = c _j (a
_i X + b _i ) is provided by the calculating means.

[0009]

According to the above means, the range of the input data is set to N
A coefficient c _j is set in each of the divided ranges, and
The divided range is further divided into M, an approximate straight line is provided for each of the M, and the equation of the straight line is represented by Y = c _j (a _i X
+ With b _i), since the coefficients a _i and the coefficient b _i is common to each of the N range, the number of coefficients to be stored in ROM becomes a N + 2M number less than the conventional example of FIG. 5 can do.

Further, the upper p bits of the input data are represented by a coefficient c.
_{j can be used as} a read address, and r bits following the upper p bits can be used as read addresses of the coefficients a _i and b _i , and the read coefficients c _j , a _i , and b
when calculating the provided credentials _{Y = c j (a i X} + b i) to _i, operation easy since the data excluding the upper p bits and r bits subsequent input data to the value of X can be used become.

[0011]

FIG. 1 is a block diagram showing an embodiment of the present invention. 1 is a ROM for fixedly storing the coefficient c _j , coefficient a _i , and coefficient b _i , 2 is an address counter for specifying the address of the ROM 1, 3 is a 16-bit register holding input data to be converted, 4 is a switching circuit for setting the address counter 2 switches the upper p bits and r bits subsequent thereto in the register 3, the coefficient c _j read from the ROM1 is 5, the coefficient a _i, and the coefficients b _i and the register 3
Y = c _j (a _i X + b) based on the input data held in
_This is an arithmetic circuit for performing the calculation of _i ).

The coefficients c _j , a _i ,
And the coefficient b _i is determined by an exponential function. Therefore, the coefficient c _j, coefficients a _i, and, for the coefficients b _i will be described with reference to FIG. In FIG. 2, the horizontal axis is input data, the vertical axis is output data after conversion, and the exponential function is Y = 2 ^X
/ 2. Here, N is set to 2 and M is set to 2 for easy understanding of the coefficient c _j , coefficient a _i , and coefficient b _i . That is, the range “8” that can be taken by the input data on the horizontal axis is 2
Evenly divide the coefficients in the range from “0” to “4” to c ₁ ,
The coefficient in the range from “4” to “8” is c ₂ . Furthermore, the range from "0" to "4" and the range from "4" to "8" are each bisected. Then, the exponential function curves in the equally divided ranges are approximated by straight lines Y ₁ , Y ₂ , Y ₃ , and Y ₄ .
When this straight line is expressed by an equation, the Y-axis is input “0”,
"2", "4", is expressed on the assumption that the _{"6", Y 1 = X / 2 +} 1 Y 2 = X +2 Y 3 = 2X + 4 = 4 (X / 2 + 1) Y 4 = 4X + 8 = 4 (X + 2), which is in the form of Y = c _j (a _i X + b _i ).

Therefore, in the case of FIG.₁Is 1, coefficient
c_TwoIs 4, and the coefficient a₁Is 1/2, coefficient a_TwoIs
1, coefficient b₁Is 1, coefficient b_TwoIs 2 and the number of coefficients is the sum
There are six. Figure 2 shows the case of approximation with four straight lines
However, when approximating with more than four straight lines,
Therefore, the number of coefficients changes. For example, 32 straight lines
When N is 2, as shown in FIG.
Coefficient c_jIs two, coefficient a_iAnd coefficient b_iAre 16 each
Thus, a total of 34 coefficients are obtained. The number of coefficients is minimal
Is the case where N is 8, in which case the coefficient c_jIs 8
Pieces, coefficient a_iAnd coefficient b _iBecomes 4 each, a total of 16
become. Therefore, the approximation accuracy of the exponential function is large when the number of straight lines is large.
As the approximation accuracy increases, the N
The number of coefficients depends on the choice
Needs to be set so that is minimized.

The ROM 1 shown in the embodiment of FIG. 1 stores 16 coefficients when the curve is approximated by 32 straight lines and N is 8 as described above. For example, from address “0” to address “7”, coefficients c ₁ , c ₂ ,.
.. C ₈ are stored, and coefficients a ₁ , b ₁ , a ₂ , b ₂ ,... A ₄ , b ₄ are sequentially stored at addresses “8” to “15”.

The conversion operation in the conversion circuit of FIG.
Will be described. First, the input data to be converted (16 bits
Is held in the register 3. And switch
The upper 3 bits of this input data (p
= 3) is the lower bit B of the address counter 2₀, B₁, B
_TwoAnd the remaining bits are set to “0”.
You. As a result, the address is changed from the address “0” of the ROM 1 to the address.
Of the input data is accessed.
Coefficient c according to 3 bits _jIs read out and sent to the arithmetic circuit 5.
Is output. That is, the upper 3 bits of the input data are 8 minutes
Since it shows which range of the horizontal axis divided
is there.

Next, the switching circuit 4 selects two bits (r = 2) of the fourth and fifth bits from the upper bits of the input data held in the register 3, and sets the bit B _{1 of the} address counter 2 to while set B _2, bit B
₃ is set to "1", the remaining bits including bits B ₀ is set to "0". As a result, any one of “8”, “10”, “12”, and “14” is accessed as the address of the ROM 1, so that the coefficient a _i is read from the ROM 1 and output to the arithmetic circuit 5. Then, by incrementing the address counter 2, since the address is advanced one, accessing the ROM1 at this address, the coefficient b _i is read and output to the arithmetic circuit 2.

The arithmetic circuit 5 uses the coefficient c _j , coefficient a _i , coefficient b _i read from the ROM 1 and the lower 11 bits of the input data held in the register 3 to obtain Y = c _j (a _i
X + b _i ) is calculated. Usually, since a DSP or the like has a multiplier and an ALU, the calculation of Y = c _j (a _i X + b _i ) can be easily performed. The calculation result is output as conversion data, stored in a RAM or the like (not shown), and used for subsequent signal processing.

[0018] Incidentally, in this embodiment, first reads the coefficients c _j, then I read the coefficient a _i and the coefficient b _i, initially reads the coefficient a _i and the coefficient b _i, then the coefficient c
_j may be read.

[0019]

As described above, according to the present invention, the number of coefficients stored in the ROM is significantly reduced as compared with the prior art.
For example, in the method shown in FIG. 5, when an exponential function curve is approximated by 32 straight lines, 64 coefficients must be stored, but according to the present invention, 16 coefficients are stored. Just do it. Therefore, there is an advantage that the use efficiency of the ROM is further improved.

The upper p bits of the input data are addressed.
Coefficient c_jIs read, and r following the upper p bits is read.
Coefficient a by bit_iAnd coefficient b_iIs read
Thus, access to the ROM is simplified. Further, Y = c_j
(A_iX + b_i)), The upper p of the input data
Bit and data excluding r bits following this and coefficient c
_j, Coefficient a_iAnd the coefficient b_iHas the advantage of simplifying the operation of
is there.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a diagram illustrating a function of output data with respect to input data.

FIG. 3 is a diagram showing the number of coefficients.

FIG. 4 is a diagram showing a conventional approximation method.

FIG. 5 illustrates an improved approximation method.

[Explanation of symbols]

 1 ROM 2 address counter 3 register 4 switching circuit 5 arithmetic circuit

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H03M 7/50 G06F 7/556

Claims (2)

(57) [Claims] In a method of converting linear input data into output data of an exponential function, a range that the input data can take is divided into N ranges, and a coefficient c _j (j is _assigned to each of the divided ranges. .., N), and further divides each of the divided ranges into M, and plots an exponential function curve corresponding to the divided ranges as a straight line Y = c _j (a _i X + b _i) ) (I
= 1,2... M), and the coefficients c _j , a _i ,
And a memory for storing the coefficient b _i , and the coefficient c _j , coefficient a _i , and coefficient b _i according to input data from the memory.
Comprising an address counter for reading out the coefficient c _j read from the memory, the coefficient a _i, and, a calculation means for calculating a _{Y = c j (a i X} + b i) on the basis of the input data and the coefficient b _i Data conversion circuit. 2. The coefficient c _j is read by setting upper p bits of the input data in the address counter,
Wherein from the upper p bits of the input data by setting the following r bits to the address counter coefficients a _i and the coefficient b _i
And Y = c _j (a _i X + b _i ) is calculated by the calculation means using the remaining bits of the input data and the read coefficients c _j , a _i , and b _i . 2. A data conversion method using a data conversion circuit according to claim 1, wherein: