JPS60245315A - D/a converting circuit - Google Patents

Abstract

PURPOSE:To minimize a switching noise, and to improve an S/N ratio by providing a pair of constant-current sources set to a constant-current of weight of 3<n>, and executing a D/A conversion by its combination. CONSTITUTION:A pair of constant-current sources I, (I'), 3I, (3I'),...3<n>I (3<n>I') set to weight of 3<n> are connected to a pair of output terminals (out) (out') through changeover switches S0, (S0'), S1, (S1')... Sn (Sn'). In case of a D/A conversion, for instance, a ternary notation is adopted so that 1 of a decimal number, 2, 3, 4, 5,...9 become I, I+I', 3I, 3+I, 3I+I+I'...6I+3I, and a binary input digital signal is converted to a ternary digital signal and supplied to a switching means. In this way, by converting a code constituted of a ROM, a noise generated when switching the switch can be set to the least significant bit (LSB), and the S/N ratio cam be improved.

Description

[Detailed Description of the Invention] [Technical Field] The present invention relates to a D/A conversion circuit, for example,
The present invention relates to techniques that are effective when used to form audio signals.

[Background technology]

Conventionally, as one of the D/A conversion methods, a weighted constant current source method is known, for example, as described in "R Linear IC Practical Circuit Manual" written by Yojibe Yokoi, published by Radio Engineering Co., Ltd., page 322.

This weighted constant current source method is, for example, as shown in FIG.
...2" I is connected to the output terminal out (out) by manipulating the switch means SO to Sn, and the analog output signal is formed by passing it through a load resistor (not shown). It is known that the following problems occur with the current source method: As shown in the input/output characteristic diagram in Figure 2, when generating an AC signal such as an audio signal, This generates relatively large switching noise near the midpoint voltage of 0.

The reason for this is that the most significant bit (MSB) set to the current value of the maximum constant current source 2" I is switched with the midpoint voltage O as the boundary. The current value of 2" I will have a large error current value according to the above weighting with respect to the variation of the element, so
Noise (N) will be generated during this switching. In the audio signal (AC signal) as described above, the signal level (S) is small near the midpoint voltage, so the signal-to-noise ratio (S/N) becomes extremely poor. As a result, for example, if a weighted constant current source type D/A conversion circuit as described above is used in a speech synthesizer, a large amount of noise will be generated near the area where the auditory sense is the highest, which is a serious problem. It becomes problematic.

[Purpose of the invention]

An object of the present invention is to provide a D/A conversion circuit that improves the S/N ratio for AC signals.

The above and other objects and novel features of this invention include:
It will become clear from the description of this specification and the accompanying drawings.

[Summary of the invention]

A brief overview of typical inventions disclosed in this application is as follows.

That is, D/A conversion is performed using a pair of constant current sources having a weight of 3n.

[Embodiment 1] FIG. 3 shows a circuit diagram of an embodiment of the present invention. Although not particularly limited, each circuit in the figure is formed on a single semiconductor substrate such as single crystal silicon by a known semiconductor integrated circuit manufacturing technique.

In this example, a pair of constant current sources I N' ), 31 (31' ) ...3'' whose weight is set to 3η
I (3"I') is used. These constant current sources are not particularly limited, but the changeover switch 5
o(Soo), Sl(31°)...Sn(
Sn') is connected to a pair of output terminals out. Note that the output terminal out (π) is provided with a load resistance, although not shown.

The constant current source and the changeover switch are configured by MOSFETs (insulated gate field effect transistors), although they are not particularly limited. That is, although not shown, by setting the sizes of MOSFETs that commonly receive a constant reference voltage according to the weights, a constant current of each weight is formed. Further, as the changeover switch, a multiplexer constituted by a transmission gate MOSFET is used. The D/A conversion operation of this embodiment is, for example, decimal 1 is ■, 2 is I+F, 3 is 3I, 4 is 3
I+I, 5 is 31-11+1', 6 is 31+31.
7 is 31+31' +1.8 is 31+31' +I+I
", 9 can be formed in the same way as 61 + 31 and below. In order to adopt such a ternary system, a binary digital signal is converted into a ternary digital signal as described above, and the switch means converts the binary digital signal into a ternary digital signal. Such code conversion is, but is not particularly limited to, code conversion configured by ROM (read only memory) (not shown).

As shown in FIG. 4, the input/output characteristics of this embodiment are such that the most significant bit (MSB) is switched at the trisection point of the input signal Din.

That is, in the left region of the three equally divided points indicated by dotted lines in the same figure, the maximum current 3ηI, 3″ is applied to the output terminal out.
The switch means of ``■'' is in the OFF state, in the center region the switch means of 3Tl■ is in the ON state and the switch means of 3"I" is in the OFF state, and in the right region 3"I, 3"
The switch means of I' are both turned on. As a result, the noise generated by the largest error current is generated at the three equal division points of the full scale of the output signal.

However, near the midpoint of the analog signal (audio signal), there is only a small error current (not shown) generated by switching the noise constant current ■, so the S/N of the small signal is small.
The ratio can be significantly improved. Moreover, since the relatively large switching noise occurs near half the peak value of the audio signal, the switching noise N becomes relatively small for the relatively large audio signal S. In other words, since the noise N is masked by the audio signal S, it is equivalent to the audibly unpleasant noise being erased.

[Embodiment 2] FIG. 5 shows a block diagram of an embodiment of a speech synthesis device to which the present invention is applied. Each circuit block surrounded by a broken line in the figure is formed on one semiconductor substrate such as, but not limited to, single-crystal silicon using known semiconductor integrated circuit manufacturing techniques.

Although the speech synthesis semiconductor integrated circuit device LSI is not particularly limited, it forms synthesized speech using the well-known PARCOR method. This example of a semiconductor integrated circuit bag for speech synthesis! L.S.
I is a voiced sound generator5. A sound source section consisting of an unvoiced sound (noise) generator 4, a K stack 6, a digital filter 7, and a D/
It is composed of an Ai converter 8, a parameter interpolation circuit 3, a parameter decoding circuit 2, a serial/parallel converter 1 constituting an interface, and the like.

The voice data (characteristic parameters) read from the ROM according to the start address of the voice to be generated from an appropriate control device (not shown) is processed as follows. That is, the audio data (feature parameters) is subjected to parallel/serial conversion by the interface l and decoded by the parameter decoding circuit 2. And parameter interpolation circuit 3
, the feature parameter values of all frames and 1 to 5 ms
Interpolation is performed for each feature parameter, and smoothing is performed for discrete changes in each feature parameter makeup.

Then, the pinch information includes the sound source section, the amplitude information and the PARC
The OR coefficient controls the digital filter 7. As a result, the original voice is restored and synthesized, and is uttered via the D/A converter 8 and external speaker SP. As the D/A converter 8 in such a speech synthesis apparatus, the D/A conversion circuit of the embodiment shown in FIG. 3 is used. Although not shown, the output of the D/A converter 8 is provided with a smoothing filter and an output amplifier.

〔effect〕

By providing a pair of constant current sources set to a constant current with a weight of +113" and performing D/A conversion operation in combination, it is possible to convert AC signals such as audio signals in the positive and negative directions around the midpoint voltage. When forming a changing output voltage, the switching noise can be set to the lowest focus (LSB).This allows the switching noise when the signal component is small to be minimized.This reduces the S/N ratio. The effect is that a significant improvement can be achieved.

(2) Since the above (1) reduces distortion due to variations in device characteristics, when built into a semiconductor integrated circuit device, the yield can be improved.

(3) According to (11) above, the D/
When used as a converter circuit, it is possible to obtain a high quality speech synthesis signal by improving the S/N ratio of the output speech signal.

Although the invention made by the present inventor has been specifically explained above based on Examples, it goes without saying that this invention is not limited to the above Examples and can be modified in various ways without departing from the gist thereof. Nor. For example, the circuit elements constituting the A/D conversion circuit may use bipolar transistors. Also, as in the embodiment circuit shown in FIG. 3, a pair of output terminals O can be
When switching to Ut and out, output signals with opposite phases to each other are obtained, so they are outputted through a smoothing filter and an amplifier circuit, respectively, and are used to drive a load.
It may also have a TL circuit configuration. Also, without using a changeover switch, the current value of each current source is 0 or I, O.
or 31. ....

It may be a single-ended circuit configuration that can vary from 0 or 3''I.

[Application field]

INDUSTRIAL APPLICABILITY The present invention can be widely used, for example, in addition to A/A converters used in speech synthesis devices, as well as -1D2/D/A converters.

[Brief explanation of drawings]

Fig. 1 is a circuit diagram showing an example of a conventional D/A conversion circuit, Fig. 2 is its input/output characteristic diagram, Fig. 3 is a circuit diagram showing an embodiment of the present invention, and Fig. 4 is a circuit diagram showing an example of a conventional D/A conversion circuit. FIG. 5 is a block diagram of an embodiment of a speech synthesis device to which the present invention is applied. 1. Interface, 2. Parameter fJ encoding circuit, 3. Parameter interpolation circuit, 4. Shoulder tone generator, 5
・・Voiced sound generator, 6.・KSUFF AC, 7.・Digital filter, 8.・D/Al converter, SP・・Speaker representative Patent attorney Akio Takahashi Figure 3 Figure 4

Claims (1)

[Claims] D characterized in that it includes a pair of constant current sources having a current value with a weight of 1.3", and a switch means for selectively flowing the current of these constant current sources to a common output terminal according to an input digital signal. /A conversion circuit. 2. The above constant current source and switch means are:
2. The D/A conversion circuit according to claim 1, wherein the D/A conversion circuit is formed by three integrated circuits. 3. The D/A conversion circuit according to claim 1, wherein the D/A conversion circuit is built in a semiconductor integrated circuit device for voice synthesis and forms a synthesized voice signal.