JPS63299516A - Digital/analog converter - Google Patents

Abstract

PURPOSE:To accurately convert the processing result into an analog form in case the voice data is expressed in the floating points having no cancelling by using a 1st D/A converter which converts an exponent part of the floating point value into an analog form and a 2nd D/A converter which applies the D/A conversion to a mantissa part of the floating point value with use of the output of the 1st D/A converter as the reference value. CONSTITUTION:A 1st D/A converter 12 converts an exponent part 11b of the floating point value 11 into an analog signal I0. While a 2nd D/A converter 13 uses the signal I0 outputted from the converter 12 as the reference value to apply the D/A conversion to a mantissa part 11a of the value 11 to convert the value 11 into the analog value. Thus it is possible to convert accurately the digital numerical value expressed in the floating points into the analog value and also to express the digital sound signals in the floating points. Then the signal processing is possible with no characteristic deterioration.

Description

DETAILED DESCRIPTION OF THE INVENTION <Field of Industrial Application> The present invention relates to a digital-to-analog converter, and more particularly, to a digital-to-analog converter for converting floating point values into analog.

<Prior Art> There is an audio signal processing device that includes a digital signal processor that performs predetermined digital processing and outputs digital audio data each time it is input. According to such a digital signal processor, multiple sets of coefficient data are set in advance according to various characteristics such as volume characteristics, equalizing characteristics, bass/treble characteristics, loudness characteristics, etc., and the operation members ( For example, by operating a knob, coefficient data corresponding to the position of the knob is selected, and digital processing is performed using the coefficient data to change the characteristics.

<Problems to be Solved by the Invention> By the way, digital audio data is expressed in fixed-point numbers for convenience of DA conversion. However, when digital processing is performed using fixed-point numbers, precision loss occurs, leading to deterioration of characteristics.

From the above, it is an object of the present invention to provide a digital-to-analog converter that can convert the processing result with high accuracy when audio data is expressed in floating point numbers without loss of digits.

Figure 1 is a schematic explanatory diagram of the present invention, in which 11 is a digital number expressed in floating point, lla is the mantissa part, 11b is the exponent part, and 12 is the first DA. The converter 13 is a second DA converter.

Function> The first DA converter 12 converts the exponent part 1 of the floating point value 11 into
1b is an analog signal ■. and the second DA converter 1
3 is an analog signal output from the first DA converter! .

The mantissa part 11a of the floating point value is set to D as a reference value.
A conversion is performed to convert the floating point value 11 to an analog value.

<Example> Fig. 1 is a schematic explanatory diagram of the present invention, in which 11 is a digital number (n.2') expressed in floating point, 11a is a mantissa part (n), llb is an exponent part (m), 12 is the index n1eD
The first DA converter 13 performs A conversion, and the second DA converter 13 performs DA conversion of the mantissa m.

FIG. 2 is a circuit diagram when the second DA converter 13 is configured as an integral type. In the figure, 21 is a clock signal P whose mantissa n is set by the stomach enable signal LD and which has a constant speed. 22 is a flip-flop that is set by the load enable signal LD and reset when the counter content becomes zero, 23 is an integrating circuit, and reference current source 23a1:+ :z capacitor b, 7:z p23c,
.

- Switch 23d, which is instantaneously closed upon generation of the toe enable signal LD; switch 23e, which is closed when the flip-flop is set, and opens when it is reset;
is provided.

4- When the toe enable signal LD is generated, the mantissa n is set in the counter 21, and the switch 23d is momentarily closed to discharge the charge in the capacitor 23b remaining as the previous value, and then the flip-flop 22 is set. do. The switch 23 is set by setting the flip-flop 22.
e is closed, the capacitor 23b of the integrating circuit 23 is charged by the reference current I0 from the current source 23b, and in parallel with this, the counter 21 outputs a lock pulse P. Count down the contents each time it occurs. Then, when the contents of the counter 21 become zero, the counter 21 flips 70 knobs 2.
2, thereby opening the switch 23e and ending the charging of the capacitor 23b. At this time, if the voltage across the capacitor 23b is E and the charging time is T, then E=T - T. Since the charging time T is proportional to the mantissa n, the voltage E at both ends becomes E=k·■·n(1), and an analog signal proportional to the mantissa n is output.

By the way, the first DA converter 12 shown in FIG.
It corresponds to the reference current source 23a (see FIG. 2) of the A converter 13, and when the index m is input, it outputs a current I proportional to 2''.In other words, the second DA converter 1
3 is the current I shown by the following formula % formula % (2). It is designed to output .

Therefore, by substituting equation (2) into equation (1), E=-n
・2” (31 (however, ・k2)
Therefore, an analog output E proportional to the floating point value n·2'' is obtained from the second DA converter 13.

<Effects of the Invention> According to the present invention, it is possible to accurately convert a digital value expressed in floating point to an analog value, and therefore, a digital audio signal can be expressed in floating point, and signal processing without characteristic deterioration can be performed. Ta.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic explanatory diagram of the present invention, and FIG. 2 is a circuit diagram of a second DA converter in FIG. 1. 11... Digital numerical value, 11a... Mantissa part, llb... Exponent part, 12... First DA converter, 13... Second DA converter patent applicant Alvine Co., Ltd. agent Patent attorney Chiki Saito Digital input (n)

Claims (1)

[Claims] In a digital-to-analog converter that converts a numerical value expressed in floating point to analog, a first converter that converts an exponent part of the floating point numerical value to analog;
A digital-to-analog converter comprising: a DA converter; and a second DA converter that converts the mantissa part of the floating point value into a DA converter using the output from the first DA converter as a reference value. .