JPS62245434A - Waveform generating device for electronic musical instrument - Google Patents

Abstract

PURPOSE:To simplify a circuit constitution, to reduce the capacity of a memory, and to apply an instrument for a delay device in which no waveform is stored, by providing a decimal value generating means which generates decimal values (d) and (1-d) in logarithmic values, an inverse logarithmic transform means, and an adding means. CONSTITUTION:In a waveform 2 desired to reproduce,a waveform generating means which generates the logarithmic values logA, and logB of adjacent amplitude values and A and B, and a decimal value generating means which generates the logarithmic values log(1-d) of the decimal values (d) and (1-d), are prepared to calculate and generate the amplitude D of the time (d) [time elapsed by (d) from (a)] between times (a) and (b), assuming the fundamental amplitude values of the times (a) and (b) as A and B. And the adding means which performs the addition of logA+log(1-d), and the addition of logB+log(d), is provided. The inverse logarithmic transform means which obtains A(1-d) and Bd by making those added values into an inverse logarithm, is provided, and the adding means which adds those A(1-d) and Bd, is provided. By performing an interpolation arithmetic operation by those means, the D can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a device that performs calculations based on data and generates a waveform signal, and particularly relates to a device that uses linear interpolation for calculations.

[Prior art]

Conventionally, the above-mentioned waveform generator includes, for example, the Japanese Patent Publication No. 58-5
There is one disclosed in Japanese Patent No. 1307.

In this invention, means for generating data consisting of an integer part and a decimal part and whose values change sequentially, a first memory storing basic amplitude values obtained by sampling a waveform, and a first memory for each basic amplitude. and a second memory that stores the difference between the values. Then, in FIG. 2, the basic amplitude value A consisting of the first memory is read out based on the integer part of the data, and at the same time, the difference between the basic amplitude value A and the next basic amplitude value B is calculated based on the integer part of the data. (B-A), calculate A+(B-A)Xd based on the decimal part d of the data, these basic amplitude values A, and the difference (B-A), and calculate according to the value of the decimal part of the data. , the amplitude values between the basic amplitude values A and B are interpolated.

[Problem that the invention seeks to solve]

However, in the above device, the first memory stores the basic amplitude values themselves, and the second memory stores the difference between the basic amplitude values. Therefore, since it is necessary to store one difference for one basic amplitude value, it is necessary to perform multiplication. Generally, a multiplier circuit has a more complicated structure than an adder circuit, and therefore, the structure of this device using the multiplier circuit itself has also become complicated. In addition, the above device interpolates the amplitude value between the basic amplitude values A and B according to the value of the decimal part of the data, so it inputs audio, temporarily stores it in memory, and changes the pitch. It could not be applied to delay devices that produce output.

[Means for solving problems]

In order to solve the above-mentioned problems, this invention
A configuration as described later is adopted. Before explaining this configuration, the present invention will first be explained in detail with reference to FIG.

In the same figure, 2 is the waveform to be reproduced, A and B
Let us consider the case where the basic amplitude values at times a and b are calculated and generated by the amplitude value at time d (a time d has elapsed from a) between 8% and 5. D is obtained by linear interpolation. If b-a is selected to be 1, D=A+(B-A)d, and if this is transformed, D=A+Bd-Ad=A(1-d)+Bd. Here, when A(1-d) is logarithmized, log
A(1-d)=log A-1-log(1-d
), similarly, when Bd is logarithmized, log Bd = log B + log d, and multiplication is converted to addition.

So, 10gA+l! og(1-d) and log B
+ log d is performed, these calculated values are anti-logarithmized to obtain A(1-d) and Bd, and when these are added, D is obtained.

The above is the principle of this invention, and based on this, this invention is configured as follows. First, adjacent amplitude values A
, a waveform generating means for generating logarithmic values log A and log B of H, and decimal values d and logarithmic value logd of (1-d).
, 10g(1-d). And logA+I! Addition means is provided for adding og(1-d) and adding log B + log d. These added values are anti-logarithmized and A(
An anti-logarithm conversion means for obtaining A(1-d) and Bd is provided, and an addition means for adding these A(1-d) and Bd is provided. These means perform interpolation calculations and obtain D.

Note that as (1-d), an approximate value obtained by inverting each bit of d is also used. This is because the linear interpolation method that forms the basis of this invention is a rough interpolation method as shown in Figure 2, so even if the approximate value of (1-d) is used, a large error will not occur, and This is because the circuit configuration can be simplified by using .

〔effect〕

As is clear from the above description, the present invention does not use any multiplication, but only addition, so the circuit configuration can be simplified. Furthermore, when a memory is used as the waveform generation means, it is only necessary to store in this memory the logarithmized value of the basic amplitude value, and in addition to the basic amplitude value as in the conventional system described above, the basic amplitude value Since there is no need to store the difference between them, the memory capacity can be reduced, and it can also be applied to delay devices that do not store waveforms.

〔Example〕

FIG. 1 shows a first embodiment in which the present invention is applied to a musical tone generator for an electronic keyboard instrument. In the figure, 4 is an address generator that generates an address consisting of an integer part 4a and a decimal part 4b. The value of this address changes to the beginning every time the clock signal M goes to H level. The integer part 4a is an adder 6
is supplied to.

This adder 6 is also supplied with the clock signal /, and when the clock signal y is at the H level, the adder 6 supplies the integer part 4a as it is to the logarithmic waveform memory 8, but when the clock signal y is at the L level. At this time, the value obtained by adding 1 to the integer part 4a is supplied to the logarithmic waveform memory 8. Logarithmic waveform memory 8
For example, logarithmic amplitude values obtained by logarithmizing the basic amplitude value of the waveform 2 shown in FIG. 2 are stored in order at each address corresponding to each integer part 4a sequentially generated by the address generator 4a. Therefore, when the clock signal y is at the H level, a logarithmic amplitude value, for example, log A, is read from the address, for example, a, indicated by the integer part 4a at that time, and when the clock signal y is at the L level, the integer part at that time is read. The value obtained by adding 1 to the value of logk (this is the same as the value when reading logk) indicates the address b to logk.
g Read B.

The fractional part 4b is supplied to an inverter 10. Inverter 10
When the clock signal y is also supplied with the clock signal y, when the clock signal y is at the H level, the decimal part 4b is inverted and supplied to the logarithmic conversion ROM 12, and when it is at the L level, the decimal part 4b is directly sent to the logarithmic conversion ROM 12. supply

When the clock signal / is at H level, what is originally required is the number (1-d
). Since ld is the two's complement of d, 1 must be added after d is inverted. However, as already explained in the section on means to solve the problem,
Since the linear interpolation method itself is a rough interpolation method, there is no significant error even if 1 is not added. Therefore, in order to simplify the circuit, logarithmic transformation R is performed by simply inverting each bit of d.
It is supplied to OM12.

The logarithmic conversion ROM 12 converts d supplied from the inverter 10 and the inverted value d into logarithmic values logd and log.
generate g d.

log Allog B from the logarithmic waveform memory 8 and log E and A'ogd from the logarithmic conversion ROM 12 are supplied to an adder 14. However, when the clock signal G is at H level, log A and log d are supplied to the adder 14, and when the clock signal M is at L level, the adder 14 is supplied with log A and log d.
l for 4! og B and logd are supplied. The adder 14 is also supplied with a value log M obtained by logarithmizing a modulation value such as an envelope added to the waveform 2 shown in FIG. 2, for example, from the amplitude modulation calculation section 16. Therefore, adder 14
When the clock signal is at H level, lOgA and lo
Add g d and log M, and when the clock signal y is at L level, add log B, log d, and log M, and obtain the summed value log A + log d +
log M.

Log B + log d + log M are sequentially supplied to the inverse logarithm conversion ROM 18.

The inverse logarithm conversion ROM 18 is supplied with l.

A+l! og'a+A'ogM, 10gB+I! og
Perform antilogarithmic transformation of d+logM, A-d-M,
B-d·M and supplies it to the addition register 20. As a result, the value of the addition register 20 becomes A-d-M+B-d
・With M, the value is the amplitude value plus the envelope.

As is clear from the above description, in this example, 10
g A + log d ten log M and log B
+ log d + division.

Note that the amplitude modulation unit 16 writes the logarithmized envelope of the envelope in memory, reads out the logarithmized envelope corresponding to the sum of the integer part 4a and decimal part 4b from the address generator 4, and adds the logarithmized envelope. In some cases, the logarithmization envelope is calculated by a computer and then fed to the adder 14.

In the above explanation, it is assumed that the basic amplitude value is already stored in the memory 8 as a logarithmized value.
It is also possible to interpolate a value obtained by inputting the converted value into a logarithmic conversion memory as a value obtained by logarithmizing the basic amplitude value.

FIG. 3 shows a second embodiment. This embodiment differs from the first embodiment in that the inverter 10 is removed and instead the two logarithmic transforms R
The difference is that OMs 12a and 12b are used. When the clock signal y is at H level, the logarithmic conversion ROM 12a converts the log based on the value of the decimal part 4b at that time, for example, d.
(1-d), adder I J lr & word eggplant ten and word eggplant six word sumtater one logarithm conversion ROM 12
b is configured to generate log d based on the value of the decimal part 4b at that time, for example, d, when the clock signal y is at L level, and supply it to the adder 14. As a result, while the first embodiment uses log d, the second embodiment uses log (1-a), improving the accuracy of interpolation.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a first embodiment of a waveform generator according to the invention, FIG. 2 is a diagram showing the principle of the invention, and FIG. 3 is a block diagram of a second embodiment. 4.6.8... Waveform generation means, 4.10, 12...
Decimal value generation means, 14... Addition means, 18... Anti-logarithm conversion means, 20... Addition means.

Claims (3)

[Claims] (1) Waveform generating means that generates adjacent amplitude values A and B as logarithmic values, decimal value generating means that generates decimal values d and (1-d) as logarithmic values, and addition logA+log(1-d)
and logB+logd, antilogarithmic conversion means for antilogarithmically converting an added value, and addition means for adding antilogarithmic values, and generating a waveform by performing an interpolation operation. (2) As the value (1-d), an approximate value obtained by inverting each bit of the value d using an inverting circuit is used. . (3) The waveform generator for an electronic musical instrument according to claim 1, wherein the logarithmic value is obtained by reading out a logarithmic conversion memory.