KR840002113B1 - Sound synthetic system - Google Patents

Abstract

A sound synthesizer for a paging system consists of a keyboard(8) for input, a CPU, an automatic reset circuit(10), a memory board(30) for storing the sound data, a memory decoder that interfaces between the CPU and memory board, a sound reproducing circuit(50) to convert sound data into a sound signal according to the keyboard input, an output control circuit(60), and a speaker(70).

Description

Call and notification voice synthesizer

1 is a schematic diagram showing an overall calling and notification voice synthesizer according to the present invention.

2 is a detailed view of the automatic reset circuit shown in FIG.

3 is a detailed circuit diagram of the audio reproducing section and the output control circuit shown in FIG.

4A is an output waveform of the audio reproducing unit shown in FIG.

4 (b) is an output waveform diagram of the output control circuit.

5 is a diagram representing a speech data collection algorithm used in the present invention.

* Explanation of symbols for main parts of the drawings

10: automatic reset circuit 20: central processing unit

30: memory board 50: voice playback unit

60: output control circuit 70: speaker

80: key input device

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a speech synthesizer for calling and notification, and more particularly, to a speech synthesizer for synthesizing and outputting speech from speech data stored in a memory board classified by a computer system.

Background Art [0002] Conventionally, in order to call and notify the public, a person directly uses a microphone or a voice, or otherwise, a simple voice is recorded on a recording tape and played back. However, this call and notification method has a problem that the procedure is very cumbersome and passive. Therefore, there has been a desire to simplify calling and notification procedures and to allow them to be done automatically. Such a desire can be met by providing a speech synthesis apparatus that synthesizes, reproduces, and makes calls and notifications by a computer method. Accordingly, automatic service is required at a service window where a customer should be called, for example, a bank or a public office. Can call the customer, and can provide the farewell guidance automatically instead of the crew in the elevator, and install the place where the voice signal in other daily life should occur, to reduce the effort, avoid congestion and make a significant contribution to maintaining order. can do.

It is therefore an object of the present invention to provide a digital synthesizer for active calling and announcements.

Automatic reset circuit for automatically resetting the entire apparatus for realizing the above object, a central processing unit for controlling the entire apparatus, a voice data memory board, a memory decoder for performing an interface function between the central processing unit and the memory board A voice reproducing section for converting voice data stored in the memory board into a voice signal according to a key input, and a voice output control circuit for amplifying and filtering the output of the voice reproducing section to an appropriate level, and the like. A voice synthesizer for call and notification is configured to decipher keystrokes at the central processing unit according to simple keystrokes and, based on the digital voice data embedded in the memory board, bring the appropriate voice data blocks in the voice datarom to the central processing unit. Changed to appropriate voice data considering the state before and after After compatibility, it is sent to D / A converter. That is, when the "ENTER" key is pressed after the key input of "1" and "2", the voice data of "10" is read out, converted and outputted accordingly, and the voice data of "Yi" is read out and subjected to the above process. The desired sound is output through the voice output control circuit. Now, the present invention will be described in detail with reference to the accompanying drawings.

Fig. 1 is a general view showing the calling and notification voice synthesizer according to the present invention as a whole. As shown in FIG. 1, the speech synthesizer of the present invention is an automatic reset circuit 10 which resets the entire system to an initializing state, a central processing unit 20 that performs arithmetic operations on the entire system, and a speech synthesizer. A memory board 30 including ROMs 31a and 31b incorporating control programs and voice data classified and stored by a computer, and a RAM 32 for storing arbitrary data by a user; 20) and a memory decoder circuit for performing a demultiplex function between the memory board 30 and a voice coupled to the central processing unit 20 to convert digital voice information embedded in the memory board 30 into an analog signal. A reproducing section 50, an output control circuit 60 for amplifying and filtering the analog signal output from the voice reproducing section 50 into an appropriate signal, and the audio signal output from the output adjusting circuit 60 It is composed of the speaker 70, and a key input boards 80, which is the area of the interface with the user for radiating to the outside.

Thus, when the user switches on to operate the call and notification voice synthesizer of the present invention, the entire system is reset to the initial state by the automatic reset circuit 10 (to be described in detail with reference to FIG. 2). When the necessary key input is performed by the key input board 80, the digital voice data stored in the memory board 30 by the operation of the central processing unit 20 is again passed through the central processing unit 20. 50) is converted into an analog signal, and then amplified and filtered through the output control circuit 60 to reproduce the voice information through the speaker.

FIG. 2 is a circuit diagram showing the automatic reset circuit 10 shown in FIG. 1 in detail. Referring to FIG. 2, the automatic reset circuit 10 is connected to the reset terminal of the CPU 20. FIG. To make the whole system in an initial state, specifically, the configuration, that is, the output Q terminal of the monostable multivibrator 18 is connected to the reset terminal of the central processing unit 20, and the A input of the multivibrator 18. The terminal is connected to ground, and the B input terminal thereof is via the Schmitt-trigger inverter 17 to emitter ground transistor T1 and diodes 14, 11, resistors 12, 15, and capacitor 13, The trigger pulse supply circuit (16) was interconnected. A resistor 19a and a capacitor 19b for determining the time constant of the pulse period generated at the Q terminal of the monostable multivibrator 18 are respectively coupled between the Rxy terminal and the Cx terminal of the monostable multivibrator 18. As a result, the period Tw of the reset pulse introduced into the central processing unit 20 is determined as Tw = 0.7 × (resistor 19a) × (capacitor 19b).

The memory decoder circuit shown in FIG. 1 is composed of four 16 demultiplexers and other logic gates coupled to four address terminals of the central processing unit 20 to make four inputs 16 outputs. Depending on the memory capacity of the ROM, 16 inputs of 4K bytes are selected by input of 4 address lines and output of 16 lines, which takes an AND logic on the memory request signal of the central processing unit and the memory read and read signals. The memory is hard-accessed.

3 is a detailed view of the voice reproducing unit 50 and the output control circuit 60 shown in FIG. 1, and the voice reproducing unit 50 receives the memory board 30 introduced through the central processing unit 20. As shown in FIG. A digital to analog converter (D / A) is combined with an output port (D _{0 to} D ₇ ) of the central processing unit 20 to convert the digital voice data of the digital signal into an analog signal. The output of is connected to the output control circuit (60). The output control circuit 60 outputs the operational amplifier 61 and the output of the amplifier 61 to amplify the analog signal output from the voice reproducing unit 50 to a size sufficient to drive the speaker 70 and to filter it with an appropriate signal. The resistor 62 connected between the (-) input terminal and the filter circuits of the resistors 63, 64, 66 and the capacitors 65, 67 coupled to the output of the amplifier 61 Another amplifier 68 is connected through.

The voice reproducing unit 50 described above is driven by an 8-bit D / A converter and has a resolution of 1/256. The input of the audio reproducing section 50 is input through the 8-bit line of the output ports b _{0 to} b ₇ of the central processing unit 20, D ₀ = 2 °, D ₁ = 2 ¹ , D ₂ = 2 ^It is a binary value represented by ² D ₃ = 2 ³ , D ₄ = 2 ⁴ , D ₅ = 2 ⁵ , D ₆ = 2 ⁶ , and D ₇ = 2 ⁷ . Since the voice reproducing section 50 is a precise current control circuit, it is influenced by the load of the output shaft. In order to prevent the influence of the load on the output shaft, the voice reproducing unit performs amplification of 1: 5 by the operational amplifier to reproduce the voice. Part 50 had full B-class characteristics. At this time, the output of the operational amplifier is represented by a graph as shown in FIG. 4 (a), which is composed of resistors 63, 66, and capacitor 65 so that distortion of voice data does not occur when the waveform is reproduced by a speaker. The low pass filter induces the complete audio signal output as shown in FIG. 4 (b).

Referring now to FIG. 5, a method for storing voice data in the call and announcement speech synthesizer of the present invention will be described.

변조)한다. 이때 변위량을 몇 비트로 표현하느냐에 따라 줄어드는 정보량이 결정되며, 단순한 변조에 의한 데이터의 수집은 입력 신호의 진폭이 큰 경우나 경사가 큰 경우에는 표현할 수 있는 변위량의 제약에 의해 원래의 신호와는 다른 더 작은 값만의 데이터를 갖게 된다. 이렇게 얻어진 데이터로는 양질의 음질을 바랄 수 없으므로 원음에 가까운 신호를 재생하기 위해서는 실제 변위량이 표현 가능한 변위량보다 큰 경우에는 변위량 표현 비트를 확장하고 전후에 이를 표시하는 코드를 삽입하므로써 변조오차를 없앤다.In the present invention, the voice data is a kind of digital recording method, but an original method is used to reduce the amount of memo notes occupied at this time. Most digital recordings use delta modulation or PCM modulation, or the former has a disadvantage in the amount of information in the sound quality. In the present invention, the combination of the above two uses the HDPCM method, the HDPCM method is designed for the storage of analog voice data, the audio signal is first collected by using the A / D converter at a predetermined time interval (PCM) and then the amount of information To calculate displacement by computer

Modulation). In this case, the amount of information to be reduced is determined by how many bits the displacement is expressed, and the data collection by simple modulation is different from the original signal due to the limitation of the displacement that can be expressed when the amplitude of the input signal is large or the slope is large. You will have only small data. Since the data obtained in this way cannot be expected to be of good quality, in order to reproduce a signal close to the original sound, the modulation error is eliminated by extending the displacement expression bit and inserting a code indicating it before and after the actual displacement is larger than the displacement.

In this case, the modulated speech data is subjected to sufficient reproduction experiment by the demodulator, and is corrected to high quality sound, and is tabled in syllable units and recorded in the ROM memory by a computer system.

Referring to FIG. 5, a diagram illustrating a pitch data collecting algorithm of the present invention is described as follows according to each step of a flowchart.

al: In this step, the 8-bit data obtained by the A / D converter is compared with the previous data to obtain a comparison value.

a2: Check whether this comparison value (data change value when T ° and T ⁺ of data) is 4 bits or more.

a3: When the change value is 4 bits or less, it is preserved as a difference between the previous data (T °) and the current data (T ⁺ ).

a4: The current data (at T ⁺ ) is now considered as the previous data (T °).

a5: Asks if the data obtained by all samplings have been processed and finishes the operation.

b1: Since the difference in data values is 4 bits or more, first insert an excess code (indicating that the data after this code is a complete data value, not a difference in value between each sampling).

b2: Preserves the 8-bit complete data value rather than the 4-bit comparison value data.

b3: The current data is set as the previous data, and the next data is regarded as the current data and prepared for the next step.

b4: Check whether the difference between the current data and the previous data exceeds 4 bits.

b5: If 4 bits are exceeded, go to step b2 to continue processing voice data.

b6: The difference is 4 bits or less, and an excess release code (a subsequent data means that the difference in values between samplings is 4 bits or less) is inserted, and the processing continues to step a3.

Next, referring to FIGS. 1, 2, 3, and 4, the overall operation of the system will be described. When the user switches on to use the call synthesizer of the present invention, the B + voltage is applied to the entire system. Is applied, and the capacitors 13 and 19b of the automatic reset circuit 10 of FIG. 2 are charged. While charging, transistor T1 is turned off so that the B ⁺ voltage is applied at low level to the input of monocular mulvibrator 18 via resistor 15, inverter 17, but the charge voltage of capacitor 13 is diode When the potential barrier of (14) is over 0.7 V, which is sufficient for the device, the transistor T1 is in a conductive state, and the B ⁺ voltage is applied to the B input of the monostable multivibrator 18 through the inverter 17 at a high level. At this time, the Q output of the monostable multivibrator 18 drops to a low level and the duration of the lowered level is maintained for a time determined by the resistor 19a and the capacitor 19b. The reset pulse formed in this manner is introduced into the reset terminal of the central processing unit 20 to drive the central processing plant 20.

As described above, when the central processing unit 20 is driven, the control signal of the central processing unit 20 is operated according to the embedded program to select the memory chip, and the central processing unit 20 includes the address bus and the data bus. And communicate with memory board through buffer. At this time, the voice data corresponding to the user's key input 80 is called from the memory board according to the command of the central processing unit 20, the output port (D ₀ ~ D ₇ ) through the central processing unit 20 appears, The voice data is then applied to the inputs D _{0 to} D ₇ of the voice reproducing section 50, which is a D / A converter (see FIG. 3).

In FIG. 3, the voice data in the digital form introduced into the voice reproducing unit 50 is converted into an analog signal through the digital / analog converter. In other words, when the voice data to be reproduced enters the CPU 20, the CPU 20 selects a corresponding voice data block in consideration of the type of sound before and after. The data taken out from the selected data block is continuously added and subtracted (demodulated) while taking into account the situation of modulation, and converted to the original signal (digital), and then output as a stair wave through the D / A converter. The analog signal thus changed is amplified to a certain level through the operational amplifier 61A (see FIG. 4A), and then output as a complete analog signal from which the stepped components are removed through the lowpass filter 63A (4b). See also) to ring the speaker 70.

As described above, the voice synthesizer for call and announcement of the present invention converts the digital voice data embedded in the memory board into an analog signal by using a key input, and amplifies and filters it to a certain level. As a result, the inconvenience of having to call and notify people by hand training is eliminated.

Claims (3)

An automatic reset circuit 10 for resetting the system to an initial state and making it operable; a central processing unit 20 coupled to the automatic reset circuit via a reset terminal and performing arithmetic operations of the system; And digital voice data coupled to the memory board 30 and the central processing unit 20, which are composed of ROMs 31a and 31b having a voice synthesizer control program and voice data and a random data memory RAM 32. A voice reproducing unit 50 for converting the signal into an analog signal, an output control circuit 60 for amplifying and filtering the output of the voice reproducing unit 50 into a predetermined zone, and a speaker for reproducing the output of the output control circuit into voice ( 70) and a voice synthesizer for call and notification, characterized in that it is composed of a key input board (80) coupled to the central processing unit (20) to perform key input. A monostable multivibrator (18), an inverter (17) connected to the B input terminal of the multivibrator, transistors T1, diodes (11), (14), and a resistor (12) connected to the input of the inverter. Trigger circuit of capacitors 13 and 16, resistor 19a and capacitance 19b connected between Rxy and Cx and B ⁺ terminals of monostable multivibrator 18 A voice synthesizer for calling and notification that uses a settling unit (10) as a penalty. The operational amplifier (61A) is composed of a resistor (62) and an amplifier (61), and the resistors (63), (64), (66), the capacitors (65), (67), and the amplifier (68). And a low pass filter (63A) to configure the output control circuit (60).

Images