JPS60117292A - Voice response system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a voice response system that employs a recording/editing method.

[Technical field of invention]

In voice response systems that use the conventional recording and editing method, the smallest units of voice data called phrases, such as words, numbers (values), dates, hours, or short sentences, are converted into AD converters (hereinafter referred to as voice data). A common configuration is to store and maintain phrases called phrases in an external storage device.In addition, the number of phrases to be included varies depending on the main purpose of the voice response system. Become.

In a voice response system, a unique code is assigned to each of the phrases that the device has, and all phrases are transferred from an external storage device to a high-speed storage device such as bulk memory (hereinafter referred to as bulk memory) at startup of the device. Load the phrase audio data. At this time, the address and length (side information) in the bulk memory of each phrase that has been quartered is held in the main memory of the voice response system as an address and length conversion table (hereinafter referred to as a conversion table). Information in the conversion table can be referenced by the code assigned to each phrase.

[Problems with background technology]

The central processing unit of the voice response system combines some of the phrases that the system has to form a single sentence (broadcast sentence), and outputs this broadcast sentence as audio from a loudspeaker, etc. be. Next, the central processing unit performs a process of arranging the phrase-specific codes according to the combination of the phrases, and sequentially refers to the resulting arrangement code with a conversion table provided on the main storage device. 5. Using the addresses and lengths obtained in this way, predetermined audio data on the bulk memory is read out by the central processing unit. , after being subjected to DA conversion, the phrase combination is output to the outside as audio.

Therefore, in conventional voice response systems, the arrangement of codes corresponding to the combination of phrases constituting one broadcast sentence is
Each time, the central processing unit of the system must perform a process of converting the audio data into an address and length on the bulk memory using software. For this reason,
This has the disadvantage that the load on the central processing unit is large, the audio output processing is slow, and the output audio is delayed, making it difficult to listen to.

[Purpose of the invention]

SUMMARY OF THE INVENTION In view of the above drawbacks, an object of the present invention is to provide a voice response system capable of outputting natural and easy-to-hear voices.

[Summary of the invention]

In the present invention, the same audio data loaded from the external storage device to the high-speed storage device is stored in the same location until the device is stopped. Focusing on the fact that the information in the length conversion table does not change, in the voice output mode, the voice data divided into phrases stored in the external storage device via the central processing unit is transferred to the high-speed storage in the voice response device. In the voice response system that outputs voice by transmitting the voice data to the high-speed storage device, the voice response device includes a control device that includes a memory that stores an address and a length conversion table in which the address and length of the voice data on the high-speed storage device are written. The voice output processing speed can be increased by causing the control device to perform a process of converting the code array data corresponding to the combination of phrases created by the central processing unit into the address and length of the voice data on the high-speed storage device. This aims to achieve the above objectives by increasing the

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the voice response system of the present invention. A central processing unit (CPU) 1 includes an external storage device (for example, a magnetic disk, etc.) 2 for storing audio data.
and voice response device 3 are connected. This voice response device 3 has an address and length conversion table 310, and a control device (AMC) that controls the entire voice response device.
31, a high-speed storage device (for example, bulk memory, etc.) 32 in which audio data is stored during normal operation;
A plurality of audio output channels (AR
C) A plurality of D/A converters (composed of a digital/analog converter 34) that converts the digitized audio data sent from the 33 and the audio output channel 33 into analog audio data and outputs it to the outside as voice. Note that the control device 31 controls the writing and reading of audio data to and from the high-speed storage device 32 and the output of audio data to the audio output channel 33.

Next, the operation of this embodiment will be explained. First, the voice loading operation at startup will be explained using FIG. 2. (1) The central processing unit 1 is the control device 3 of the voice response device 3.
Set 1 to write tree mode.

(2) Next, the central processing unit 1 retrieves the audio data phrase by phrase from the external storage device 2, and records the code, address, and length information data as shown at 21 in FIG. 2 and the data shown at 22 in FIG. The audio data of the phrase itself is sent to the control device 31.

(3) When the control device 31 receives the code address and length information 21 from the L1 central processing device 1, it sequentially stores the code, address, and length in the address and length conversion table 310 provided in the control device 31. Write data such as.

At the same time, in order to write the audio data 22 into the high-speed storage device 32, the control device 31 passes the address and length data to the high-speed storage device 32, and also sends the audio data 22 sent from the central processing device 1 to the high-speed storage device 32. is sent to the high speed storage device 32.

(5) The high-speed memory unit 32 stores the audio data sent from the control device 31 at the position specified by the address and length data sent in the same way.

When this storage is completed, the control device 31 issues a completion notification to the central processing unit 1. By repeating the operations (2) to (5) above for all the phrases that make up the broadcast text, the trap setting in the control device 31 at startup, the truncated address, the setting of the length conversion table 310, and the high-speed memory Audio data can be loaded into the device 32.

Next, the sound output operation during normal operation will be explained using FIG.

(Li) The central processing unit 1 is the control device 31 of the voice response device 3.
Set to read mode.

(2) The central processing unit 1 performs a process of combining phrases for audio output, creates code array data indicated by reference numeral 23 in FIG. 3 corresponding to these seven races, and passes this to the control unit 31.

(3) The control device 31 sequentially extracts codes from the beginning of the received code array data 23, and searches the address and length conversion table 310 provided in the control device 31 based on these codes to find the same code.

(4) When the control device 31 finds the same code, it reads out the address and length information integrated with this code and sends it to the high speed storage device 32.

(5) The high-speed storage device 32 reads audio data of the length specified by the length from the position specified by the address, and sends it to the control device 31 (6) The control device 31 is a high-speed storage device The audio data sent from 32 is transferred to the audio output channel 33. The audio data transferred to the audio output channel 33 is sent through this channel to the D/A converter 34, where the analog audio data is converted and the audio output for one phrase is performed. When the voice output for l phrases is completed, the voice output channel 33 outputs a voice output completion notification to the control device 31.

(7) The control device 31 repeats the operations (3) to (6) above until the data of the code array data 23 sent from the central processing unit 1 is completed (determination is made using an end code, etc., if necessary).

As a result of such an operation, a broadcast sentence (one sentence) is outputted to the outside as a voice, and the control device 31 notifies the central processing unit 1 of the completion of voice output. The audio output of the code array data constituting one broadcast sentence passed to the control device 31 of No. 3 is completed.

According to this embodiment, an address and length conversion table is provided in the control device 31 constituting the voice response device 3, and the code array data 23 is stored in the control device 31 in the high-speed storage device 3.
Since the process of converting to the address and length of the audio data above 2 is performed, the CPU
All you have to do is create the code sequence data 23 corresponding to the broadcast text using the software that operates it.
It is possible to significantly reduce the software processing load on the CPU and improve the audio output processing speed.

For this reason, the outputted voice does not become elongated or awkward, making the voice more natural and easier to listen to.Also, since the voice response device 3 outputs one broadcast sentence, the voice The connection between phrases at the time of output can be improved more than before.In other words, there is no break between phrases.Furthermore, the voice response device of this embodiment can output voice using a normal input/output device. Therefore, it can be implemented immediately without any other configuration changes.

〔Effect of the invention〕

According to the voice response system as described above, the address/length conversion table that stores the address and length on the high-speed storage device of voice data loaded from the external storage device to the high-speed storage device is converted into a voice response system. The fir!' is provided in the control device of the device, and performs operations such as determining the address and length on the high-speed storage device corresponding to the code sequence (combination of phrases) constituting one broadcast sentence from the address/length conversion table. #By allowing the device to do this, it has the effect of outputting natural and easy-to-listen audio.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an embodiment of the voice response system of the present invention, Fig. 2 is a chart showing the voice data loading operation at startup in the system of this embodiment, and Fig. 3 is a block diagram showing an embodiment of the voice response system of the present invention. FIG. 3 is a chart diagram showing the audio output operation during normal operation of the system. 1...Central processing unit, 2...External storage device, 3.-
・Voice response device, 31...Control device, 32...High speed storage device Patent attorney Noriyuki Chika (and 1 other person) Figure 1, Figure 2

Claims (1)

[Claims] In the audio output mode, the audio data divided into phrases stored in the external storage device via the central processing unit is
In a voice response system that outputs voice by transferring it to a high-speed storage device in a voice response device, a control device comprising a memory that stores an address and a length conversion table in which the address and length of voice data on the high-speed storage device are written. is provided in the voice response device, and the control device performs a process of converting code array data corresponding to a combination of seven races created by the central processing unit into an address and length of voice data on the high-speed storage device. A voice response system characterized by allowing users to perform actions.