JPS6220559B2 - - Google Patents

Description

[Detailed description of the invention] TECHNICAL FIELD The present invention relates to a voice response device.

In voice response devices that use conventional digital recording methods, in order to simplify the audio editing method, audio is divided into certain basic unit lengths, and audio that is longer than this basic unit length is divided into integral multiples of this basic unit length. The length of the audio is stored in the audio storage device.
In such a case, for example, if the basic unit length is 0.5 seconds, the audio that is actually 0.6 seconds long will be stored in the storage device as twice the basic unit length, that is, 1 second of audio, so it will be equivalent to 0.4 seconds. Digital storage elements will be wasted. The amount of digital storage elements that are ineffectively used is theoretically equivalent to the number of spoken words x 0.5 x 1/2 seconds, and the well-known pulse code modulation (PCM) method is used to record the audio. Assuming that the number of spoken words is 100, the amount of storage elements equivalent to 25 seconds, or 1.6×10 ⁶ bits, will be wasted.

Especially in a voice response device that uses an expensive memory such as an IC memory as a voice storage element, this waste cannot be ignored in terms of cost. This waste will be reduced as the basic unit length of the audio is made shorter, but on the other hand, the editing control of the audio becomes more complicated, and the amount of hardware for the editing control device that specifies the read address of the audio storage device increases. This results in an increase in the load on the processing device that instructs the editing control device about the audio to be sent. For example, the basic unit length is
In the case of 125 microseconds (μs), the processing device is
Since the editing control device must be instructed on the audio to be transmitted every 125 μs, the processing load increases rapidly. To prevent this, if, for example, 0.5 seconds worth of audio is to be commanded at once, the editing control device would need hardware to store as many as 0.5 seconds/125 μs = 4000 audio addresses.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a voice response device that eliminates the above-mentioned drawbacks, improves the efficiency of use of a voice storage device, and allows simple voice editing control.

The present invention adds a specific code to the end of each voice code to indicate the end of the voice, stores it in a voice storage device as voice of any length not limited by the basic unit length, and reads out these voices. When outputting,
The termination code added to the end of each voice is detected to identify the end of that voice, and from then on, the code generator is installed separately from the voice storage device to generate a By outputting a code corresponding to silence, the waste of the audio storage device is saved, and the editing control of the audio is performed for each basic unit length, thereby simplifying the editing control.

FIG. 1 shows an embodiment of the voice response device of the present invention, and FIG. 2 shows an example of the voice code arrangement of the voice storage device of FIG.

In FIG. 1, the audio code is stored in the audio storage device 2, and the first storage location address where the audio code to be output from the processing device to the audio editing control device 1 is stored and the line number information to be sent are Given. The digital voice code read from the voice storage device 2 is given to the digital-to-analog converter 6 via the switch 5 and converted into analog (PAM).
converted to code. Then, under the control of the audio editing control device 1, the signal is distributed through the distributor 7 to the demodulators 8-1 to 8-n provided corresponding to the output lines, and is demodulated into audio and output. The silence code generator 3, end code detector 4, and detailed operations will be described later.

FIG. 2 shows a partial example of the audio code array in the audio storage device 2 described above. Audio is recorded in digital codes using the well-known pulse code modulation (PCM) method, and one second of audio is 8000 bytes, or
It is stored in the audio storage device 2 as a 64000-bit digital code. A semiconductor integrated circuit memory (IC memory) is used as the storage element of the audio storage device 2, and 8-bit audio codes are simultaneously read out in one read operation. Thus, in the example shown in FIG. 2, voices "A" and "B" each having a length of substantially 0.4 seconds are each stored as 3200-byte voice codes, with a specific bit pattern at the end of each voice code. A termination code is added. The audio codes "A" and "B" are assigned a read address for each byte corresponding to 125 μs of audio, and the audio code "A" is the audio code in the order of addresses a ₁ , a ₂ , ... a ₃₂₀₀ . “B” is address b ₁ , b ₂ , …b ₃₂₀₀
are read in this order.

Next, an outline of the operation will be explained when the audio unit length for audio editing control is set to 0.5 seconds.

First, in order to output continuous, uninterrupted audio on n lines simultaneously, one byte of audio code is read out from the audio storage device during 125/n microseconds.
The premise is to convert it into a PAM code and transfer it to the demodulator.

In FIG. 1, the processing device instructs the audio editing control device 1 about the start address of the audio to be sent to each line for n lines every 0.5 seconds. The audio editing control device 1 temporarily holds the start address of each line instructed by the processing device, and then updates it at another timing.
Every 125/n microseconds, the addresses to be sent to each line in the order of 1, 2, . . . n lines are read and sent to the voice storage device 2. Now, the voice “A” shown in Figure 2
shall be transmitted in the first 0.5 seconds, and audio "B" shall be transmitted in the next 0.5 seconds. In this case, the audio editing control device 2 sends the address _a1 to the audio storage device 2 at the timing assigned to the line 1 during the first 125 microseconds of the first 0.5 seconds. This results in
The first byte _a1 of the audio code "A" is read out, transferred to the demodulator 8-1 via the switch 5, digital-to-analog converter 6, and distributor 7, and output as 125 microsecond audio. . When the next cycle of 125 microseconds arrives, the audio editing control device 1
The voice address is automatically incremented by one step from _a1 to _a2 , and this is sent to the voice storage device 2 in the same manner as described above, and the second byte _a2 of the voice code "A" is read and output. Thereafter, in the same manner, the audio code "A" up to the 3200th byte a ₃₂₀₀ is sequentially read out and output. When this is detected by the end code detector 4 stored next to a ₃₂₀₀ , the switch 5 is switched to the silence code generator 3 side, and the silence code generator 3 is switched on for the next 0.1 seconds.
A specific code corresponding to silence is transmitted. 0.5
When the transmission of the silence code is completed after a second has elapsed, the switch 5 switches to the audio storage device 2 side again, and the processing device transfers the start address b ₁ of the audio “B” to the audio editing control device 1. Thereafter, the audio code "B" is read out and output in the same manner. When the end code added to the end of the voice code "B" is detected, a silent code is sent out for 0.1 seconds in the same way.

As explained above, according to the present invention, it is possible to save the memory element for the silent code shown by the broken line in FIG. This makes it possible to reduce the load on the processing device, simplify voice editing control, and provide an inexpensive voice response device.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing one embodiment of the present invention;
FIG. 2 is a diagram showing an example of the audio code arrangement in the audio storage device of FIG. 1. DESCRIPTION OF SYMBOLS 1... Audio editing control device, 2... Audio storage device, 3... Silence code generator, 4... End code detector, 5... Switching device, 6... Digital-to-analog converter, 7... Distribution 8-1 to 8-n... demodulator.

Claims (1)

[Claims] 1. In a voice response device that stores voice codes in multiple memory locations and edits and outputs them for each certain basic unit length, each voice code is serialized with a termination code at the end. a storage means for storing in a memory location, a means for detecting an ending code attached to the voice code, a means for generating a code corresponding to silence, and a storage means for storing the ending code at the beginning of the storage means until the ending code is detected by the detecting means. After the end signal is detected by the detection means, the code corresponding to silence is read out from the code generation means for the remaining time of the constant basic unit length. A voice response device characterized by comprising: means for taking out and outputting.