JPS61233796A - Voice accumulator - 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 [Field of Industrial Application] The present invention relates to an audio storage device using a disk memory. In particular, audio data related to messages such as operational guidance, such as informing the user how to use the audio storage device, etc., is stored in a disk memory, and the recorded content is edited and played back, which outputs audio using the so-called recording editing method. This invention relates to a type of audio storage device.

[Prior Art] In conventional audio storage devices, in order to reproduce guidance sentences according to the content of the service, in order to reproduce guidance sentences according to all contents, it is necessary to assume all cases. A voice response device was needed and equipped to prepare guidance sentences, which are created voice information, select the necessary guidance sentences, and output and respond with voice (Reference 1: ``Speech storage system configuration method''). "Study" Technical Research Report of Institute of Electronics and Communication Engineers, pp. 5E82-56.58, Published in 1933.
(September 17, 2017).

In addition, among the guidance sentences stored in disk memory,
Some devices stored only fixed phrases, and variable words such as time and number 1 were stored in memory for voice information (semiconductor memory), and responses were made by recording and editing them.
Reference 2: [Configuration of voice storage device for telephone network] Institute of Electronics and Communication Engineers, Technical Research Report, page 5E84-70.53, published October 19, 1980).

[Problems to be solved by the invention] In the former case (Reference document 1), it is necessary to have a voice response device including a speech synthesis device and its control device, and in the case of the latter (Reference document 2), Since all the variable words are stored in a memory for voice information, a voice memory (semiconductor memory) is required, resulting in a problem that the device becomes complicated and expensive.

[Means for Solving the Problems] In order to solve these problems, the present invention stores necessary guidance information (elements constituting a guidance sentence) in a disk memory and reads it out through a buffer 1 memory. It is something.

[Function] By directly editing and reading out the necessary guidance information from multiple pieces of guidance information stored in the disk memory onto the buffer memory, it is possible to edit all the necessary guidance sentences, and the hardware By preventing an increase in the amount of data and complicating the control, it is possible to eliminate the need for a voice response device that is expensive and requires complicated control, and a large-capacity blind spot memory including a semiconductor memory.

[Example] An example of the present invention will be described with reference to FIGS. 1 and 2.

1 is an audio storage device, and audio interface 2 (28-2
n~2p), direct memory access control circuit 3. Disk memory control circuit 4. Disk memory5. Control memory 6, control circuit 7. It includes a common bus 8. Reference numeral 10 denotes an exchange, which is connected to the voice storage device 1 through lines 9 (9a to 9n to 9p), and is also connected to a telephone set 12 through a line 11. Here, telephone 12. Switchboard 10. or audio interface 2 (2a~2n~2
p) includes a D/A and A/D converter, and all signal exchange within the audio storage device 1 is performed using digital signals.

FIG. 2 shows an example of the circuit configuration of the audio interface 2n, which includes an audio interface control circuit 21n1, a buffer memory 22An and a buffer memory 22Bn, and buffer changeover switches 24n and 25n, which are logical switches that operate in conjunction with the line interface 23n1. There is. Buffer memory 22 for audio interface 2n
The reason why the buffer memory 22Bn and the buffer memory 22Bn have two sides is because of the difference in transfer speed between the audio information and the disk memory 5, and the time it takes for the access head included in the disk memory 5 to move to the address of the record desired to be accessed. This is to perform a buff 1 operation to absorb
Two buffer memories 22An and 22Bn are used alternately on the line interface 23n side and the common bus 8 side. Since these two buffer memories 22An and 22Bn are logical, one buffer memory may be divided and used.

These buffer changeover switches 24n and 25n are controlled by an audio interface control circuit 21n, and the buffer 1 changeover switch 24n is on the upper side (buffer memory 22A).
When the buffer selector switch 25n is on the lower side (buffer memory 22Bn side), the buffer selector switch 25n is on the lower side (buffer memory 22Bn side), and when the buffer selector switch 24n is on the lower side (buffer memory 22Bn side), the buffer selector switch 25n is on the upper side (buffer memory 22Bn side). memory 22An side).

For example, during general playback, when the line interface 23n is reading out the audio information from the buffer memory 22Bn via the buffer changeover switch 24n, the buffer memory 22An buffer? Changeover switch 25
When the contents of the disk memory 5 are read out via the buffer memory 22Bn and the entire buffer memory 22Bn is reproduced, the audio interface control circuit 21n switches the buffer changeover switch 24n.

25n is switched to the opposite side to perform a buffer operation.

Furthermore, each time the buffer changeover switches 24n and 25n are switched, the audio interface control circuit 2In transfers the buffer memory 22An to the control circuit 7.

and requests an update of 22Bn. Note that the control circuit 7 can also serve as the audio interface control circuit 21n. In general, the audio interface control circuit (2
1a to 21n to 21p) can also be integrated into one audio interface control circuit.

For example, when outputting guidance from the voice storage device 1 to the telephone @12, the control circuit 7 controls the disk memory control circuit 4 and the direct memory access control circuit 3, and controls the direct memory access from the disk memory 5. The audio information for guidance is read out to the buffer memory 22An or 22Bn under the control of the circuit 3, and the audio information is buffered at the audio interface 2n. It is transmitted to telephone 12 via line 11. Here, the disk memory control circuit 4 moves the access head included in the disk memory, reads out the target information as serial data from the area where it is stored, writes serial data to the area, and writes the desired information to the area. If the bus 8 is a parallel bus, it also performs serial-to-parallel conversion.

Figure 3 shows the guidance information q1 to qll.
The correspondence relationship between the guidance control block array 40 on the control memory 6 and the guidance areas 31-1 to 31-m on the disk 30 included in the disk memory 5 when reproducing a guidance sentence consisting of Show one row. This example shows that the guidance g1 is stored for the length indicated by the data length d1 from the address on the disk 30 indicated by the disk address block A1. When reproducing a guidance sentence, depending on the editing content, guidance controller 1 to roll block row 4 consisting of necessary guidance q1 to qI, l and an end mark are required.
0 is configured according to instructions from the control circuit 7, and the audio information on the disk 30 is configured according to instructions from the audio interface control circuit 2.
Each update request from 1n is sequentially transferred from the disk memory 5 to the buffer memory 22An or 22Bn.

Here, the data length d1 to dIIl is the buffer memory 22
The size is not necessarily an integral multiple of the size of An or 22Bn, and may be smaller than the size of the buffer memory. Therefore, the buffer memories 22Aa to 22Ap or 22Ba to 22A from the audio interface control circuits 21a to 21p
In order to update a maximum of P buffer memories 22A or 22B from a to p for each 22 Bp update request, it may be necessary to access the disk memory 5 multiple times.

When accessing the disk memory 5, the 5CAN method (Reference 3: "Performance Analysis of Operating Systems" Information Processing Society of Japan, 49 Page, published August 30, 1982), the buffer time Tbuf is the time required for the line interface 23 to read or write one buffer memory 22A or 22B once. , the number of simultaneous accesses M (Tfi, j includes buffer memory 2)
The maximum number of times to read from or write to the disk memory 5 until the buffer size of 2A or 22B is filled. ), the following formula holds.

T −2T M≦ buf sk (1>T
acc "set" Wait However, T: Time for the line interface 23 to read or write the contents of 22A or buf or 23B once Tsk "In the disk 30, the outermost circumference of the information areas stored on the same disk The time required for the access head to move between the area and the innermost circumferential area T, ...; the time required to read or write information from one guidance area 31 of the disk memory 5 (in the example of equation (1), (Variation due to amount is ignored and the averaged value is used.) T ; Time required for stabilization after the access head moves - °Walt after the access head moves and stabilizes
' Rotational waiting time until the beginning of the target guidance area 31 on the disk 30 reaches the access head.

Assuming that the number of lines that simultaneously access the voice storage device 1 is P, it is not enough to play only one guidance q, and by editing two or more guidance Qs, the disk memory 5 can be accessed to record and edit guidance sentences. The maximum number of edits MAXED is MAXED=MP.

Here, if recording editing is performed exceeding the maximum number of edits MAXED, 22A or 22B of the buffer memory is
Since the update of the audio will not be completed within the buffer time 'buf and will not be updated, during playback, the same audio that was previously read from the buffer memory 22A or 22B will not be updated within the buffer time Tb.
It will be output at intervals of uf. In the above description, for example, when an update request to the buffer memory 22An or 22Bn is received from the audio interface control circuit 21n and requires multiple accesses to the disk memory 5, the first access is not counted as the number of edits.

Therefore, in the present invention, when it becomes necessary to edit more than the maximum number of edits MAXED, the previous buffer 1 memory 22
In order to prevent the same sound that was played back from A or 22B from being played back again, the silence or white noise information stored in the control memory 6 is transferred to the buffer memory 22A or 22B, and the sound is reproduced as sound. Control is performed from the control circuit 7 to maintain continuity and prevent deterioration in the quality of the guidance sentences.

Next, flowcharts are shown in FIGS. 4A to 4F, and diagrams for explaining the relationship between the flowchart and the buffer 1 size are shown in FIGS. 4G and 4H.
Explain the operation.

The first program, the 1-1st program, the second program, and the third program are running in parallel processing, and the second
After a reset start, the program (STEPs 71 to 77) monitors the presence or absence of a transfer request from the disk memory 5 to the buffer 1 memory 22A or 22B issued by the first program. The editing counter included in the control memory 6 is reset (STE
P81). The first program performs guidance recording and editing processing. Note that the parallel processing and the monitor that starts the program according to the "buf cycle" are well-known techniques, so the explanation will be omitted.

When outputting a guidance sentence consisting of guidance g1 to g (Fig. 3), as described above, the guidance control block (GCB) corresponding to the line n to be output is
It is abbreviated as ), set the address on the disk 30 of each guidance (disk address block Ai) and each size (data length d,) in the guidance control block string (abbreviated as GCBSn) 40.
After performing initialization (STEP 52, 54), check for audio information data (B) that has not been transferred to the buffer memory 22An or 22Bn among the guidance audio information specified by the GCB being processed (STEP 55). ), the amount of which is the amount of information that has not been updated after the buffer memory update request from the audio interface control circuit 21n (TRCNT
2rl), the buffer memory 22An
Or the untransferred voice information data (B) to 22Bn
request the second program (STEP 71 to 77) to transfer the data to the buffer memory 22An or 22Bn (STEP
EP57).

Here, the operations of 5TEP55, 57 and 64 are explained in the 4th G.
This will be explained again using figures.

4G and 4H, the buffer size of the buffer 22A or 22B is shown, and the audio information of guidance gi""i+1, which is a component of the guidance sentence, is stored in this buffer 122A or 22B.

Guidance con 1 showing the editing content in Figure 4G
~ Guidance pointer (GP
Data length d of the voice information of the guidance Q instructed by n)
, the remaining untransferred audio information data (B) after subtracting the data transferred to the buffer memory 22A or 22B (TRCNT1 n) is calculated (STEP 55), and the amount of unupdated information (TRCNT1 n) in the buffer memory 22A or 22B is calculated.
2n) is processed with 5TEP56), the case shown in Fig. 4G (a) is processed with 5TEP57 or less (Fig. 4B), and the case shown in Fig. 4G (b) is processed with 5TEP64 or less (4th
Figure C). - In the case shown in FIG. 4G, it is necessary to access the disk memory 5 multiple times in order to update the buffer memory 22A or 22B.

Since this operation is performed, the number of edits is counted for all lines in FIG. 4G, and (STEP 60 (4th
(Figure B)), if this value exceeds the maximum number of edits MAXED, the silent or white noise information in the control memory 6 is transferred to the buffer memory 22A or 22B, and the amount of unupdated information (Figure 4G (a) ) of TRCNT2nv inus B, transfer only TRCNT2n of Fig. 40 (・b) (S
TEP63). Since this transfer does not involve accessing the disk memory 5, it can be performed at high speed, and since it is possible to access and multiplex the disk memory 5 of other lines, messages on other lines can be played back and recorded without any problems. becomes possible. The audio information corresponding to the guidance gi+1 to g is updated and reproduced as shown in FIG. 4H. Roughly speaking, for the guidance sentences (q1 to gIll), silence or white noise is simply inserted between each guidance (Q, Q2...qIIl), and the guidance that makes up the guidance sentences spans two buffer memory sides. Even if there is a problem, there is no break in the guidance and it does not interfere with the communication of the content (Figure 4H).

Here, the initialization (resetting) of the counter that counts the number of edits included in the control memory 6 is performed in response to an interruption by a clock signal generated by a clock signal generator included in the control circuit 7, with a period of Tbuf interval. A method of generating a signal to
P91 to P94), the audio interface control circuit 2
It is also possible to update the buffer memory 22A or 22B from 1 to 1, by observing the request, ie, the "buf interval."

Furthermore, the movement control of the access head and the scheduling of the movement order are performed in response to a transfer request (STEP55, 5TEP57>) to the buffer memory 22A or 22B.
This is done in the second program (STEP 71-77). Note that numbers, verbs, etc. that are often recorded and edited are placed on the same cylinder on the disk 30 (an area that does not require movement of the access head), and this eliminates the movement of the access head. ,. It is also possible to set 1=O and increase the maximum number of edits.

Further, in the above embodiment, the number of edits is counted for all lines, but it is also possible to allocate the maximum number of edits MAXED to each line and manage the number of edits for each line.

All the control operations in the above embodiments are under program control in the control circuit 7. This program configuration is a software matter that can be configured as appropriate based on the processing flow shown in FIGS. 4A to 4H, and does not limit the technical scope of the present invention.

[Effect of the invention].

As is clear from the above description, according to the present invention, guidance information is stored on a disk memory without the need for a dedicated voice response device or a memory (semiconductor memory) that stores all guidance information. Since it is possible to edit, synthesize, and output the images, it can be realized at low cost.

Furthermore, even if the number of edits exceeds a value determined by the buffer time of the buffer 1 memory and the access speed of the disk memory, it is possible to realize an audio storage device that does not affect other lines that are not playing guidance. Therefore, the effect in use is extremely large.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an embodiment of the present invention, FIG. 2 is an example of the circuit groove configuration of the audio interface 2n, and FIG. 3 is a diagram showing the guidance control block on the control memory 6 and the disk in the disk memory 5. 30, FIGS. 4A to 4F are flowcharts of the control circuit 7 during guidance reproduction, and FIGS. 4G and 4H are diagrams for explaining the flowcharts. 1... Audio storage device, 2... Audio interface,
3... Direct memory access control circuit, 4...
Disk memory control circuit, 5... Disk memory, 6
... Control memory, 7... Control circuit, 8... Common bus, 9... Line, 10... Exchange, 11... Line, 12... Telephone, 21... Voice Interface control circuit, 22A, B...Buffer memory, 23...
・Line interface, 24.25... Buffer changeover switch, 30... Disk.

Claims (2)

[Claims] (1) A disk memory means including a disk memory divided into a plurality of guidance areas, a buffer memory means including first and second buffer memories for reading and writing to and from the disk memory means, and the buffer memory means In the audio storage device, the audio storage device includes a control means for controlling access to write and read data in the disk memory means via the disk memory means, and a control memory means for storing silence information or white noise information. The control means stores audio information for a plurality of guidances for composing a sentence, and the control means stores the audio information necessary from the disk memory means in order to compose the guidance sentence by recording and editing the plurality of guidances. control to read and reproduce audio information for a plurality of guidances onto the first or second buffer memory, and determine the number of guidances for updating the first or second buffer memory; If the value determined by the memory buffer time and the maximum number of lines accessed simultaneously is exceeded, the control memory means outputs an amount of the silent information or white noise corresponding to the guidance exceeding the value. An audio storage device characterized in that information is stored and reproduced on the first or second buffer memory. (2) Among the plurality of guidances, audio information for a plurality of guidances that are frequently edited into one guidance sentence is stored in the same cylinder of the disk memory. audio storage device.