JPH0368400B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a voice storage method in an exchange that provides voice mail services, voice output services, and the like.

Traditionally, switchboards have provided instantaneous connections between arbitrary telephone sets, providing direct communication between callers and called parties. However, in real situations, the person you want to talk to may be absent, or may be talking on another telephone, or the other party's telephone may be busy, so you may not be able to complete the task in one go.
This problem is particularly important in business communications, where the probability of successfully communicating with the desired party is 30.
There are also research reports that show that it is about %. Additionally, many calls are not complex or urgent enough to require the caller to come up to the telephone and speak directly.
Therefore, a voice storage device is installed in the exchange, and when the other party is absent or busy, or the message does not need to be spoken directly, the message is stored as voice, and the exchange sends the message back to the caller at the desired time. It is expected that a service will be provided that will deliver the message to the other party on behalf of the other party, or have the other party come and listen to the message at a time convenient for the other party. This is called a voice mail service, and is considered to be one of the important service functions that future exchanges should have.

Conventionally, tone signals such as a dial tone and a busy tone have been used to inform the caller of the progress status of a call from an exchange.
However, as the functions and services of switching equipment become more sophisticated in the future, the types of information that should be or would be useful to be communicated from the switching equipment to callers will rapidly increase, and it is thought that it will be difficult to identify them using only tone signals. becomes extremely difficult. Therefore, the switching equipment must have a function to output such information in the form of voice. In order to realize the voice output function, it would be most ideal to use a voice synthesizer, but currently the sound quality is still not sufficient. Therefore, the most practical method is to store the audio information to be output in advance in the exchange and read it repeatedly as needed.

As mentioned above, whether it is a voice mail service or a voice output function, it is required to have a voice storage device in the exchange, but in that case, the biggest problem is the storage method of the storage device and It is capacity. As a storage method, recording on a tape recorder using an analog method may be considered, but this is not suitable for repeated random reading and writing. Therefore, it is advantageous to digitize audio and treat it in the same way that data and files are processed in a computer. Considering that digital exchanges will become mainstream in the future, it is even more advantageous to handle digital voice. However, in that case, storage capacity is a problem. According to the world standard method for digitizing audio signals (PCM method), audio signals have an amount of information of 64 kbits per second, that is, 8 kbytes. Therefore, it takes 480k bytes, or about 0.5M, to store one minute of voice messages.
Byte storage capacity is required. Considering that an extremely large number of messages are accumulated in the entire exchange, measures must be taken to somehow reduce this accumulation capacity.

One of the countermeasures is to change the digitization method from the PCM method to a modulation method that requires less information, such as the adaptive delta modulation (ADM) method.
This is a method of performing so-called band compression, converting to an adaptive differential PCM (ADPCM) method, etc., and then storing the data. Even now, it is 1/2 the speed of PCM method, 32kb/s
I've been able to get pretty good sound quality with the
It is expected that 16kb/s will also be available.

Another measure to reduce the storage capacity is to remove silent periods between human sounds and store them. It is thought that a considerable amount of storage capacity can be saved by eliminating silent sections that exist between words or sentences, and silent sections that exist when the speaker is thinking.

However, when silent sections are removed and stored in this way, a method for reproducing the silent sections is required when reproducing them. Naturally, the lengths of silent sections and sound sections are not constant, so data such as the position and length of silent sections, the boundary points of adjacent sound sections on the storage device, etc., must be stored and adequately managed. There must be.

FIG. 1 shows a conventional example of a sound storage method in which silent sections are removed, stored, and reproduced as described above. (Bell System Technical Journal,
October 1980 p.1383〜1395“An Experimental
In the speech storage system shown in FIG. 1, the digitized speech code applied to input terminal 10 has a frequency equal to the sampling clock of the speech code applied to input terminal 11. The clock signal is used as a DMA (Direct Memory Access) clock, and the buffer memory 13 or 14 provided on the main memory of the computer 12 receives the DMA.
mode and are written sequentially. The buffer memory 13 or 14 to be written is switched by a switch 16 based on a command from the control section 15 of the computer 12. However, switch 16 is a logical switch. Assuming that the switch 16 is now in the state shown in FIG. 1 and the voice code has been written to the buffer memory 13 to the end,
An interrupt is generated to the control unit 15 to notify it. In response to this interruption, the control section 15 turns the switch 16 to the buffer memory 14 side to continue writing the voice code. Meanwhile, during this time, the control unit 15 performs various calculations based on the audio codes written in the buffer memory 13 and determines whether these audio code sequences correspond to a sound section or a silent section. Determine. As a result, if it is determined that it is a sound section, the control section 15 transfers the audio code written in the buffer memory 13 to the magnetic disk 17. At that time, the switch 18 is set to the buffer memory 13.
It's falling to the side. Switch 18 is also a logical switch and operates in conjunction with switch 16 to select different buffer memories.

On the other hand, if it is determined that the audio code in the buffer memory 13 corresponds to a silent section, the data is not transferred from the buffer memory to the magnetic disk. Therefore, when the buffer memory 14 is written to the end, writing to the buffer memory 13 starts, and the previous voice code disappears. The operation of the control section 15 with respect to the buffer memory 14 during writing to the buffer memory 13 is exactly the same.

In other words, in the conventional method shown in FIG. 1, the entire speech code written in the buffer memory is treated as one block, and it is determined whether or not it is a sound section in units of blocks, and the block that is determined to be a sound section is (sound block) to a magnetic disk. Also, Batsuhua
The size of the memory is selected to correspond to one sector of the magnetic disk or an integral multiple thereof in relation to the input/output unit of the magnetic disk.

In order to correctly insert silent blocks between sound blocks during playback, the control unit 15 transfers the sound blocks to the magnetic disk 17 and also stores, for example, the serial number from the beginning of the message of that block and the block. Stores addresses (sector numbers, etc.) on the magnetic disk. This information is stored as control data in the main memory or in a portion of the magnetic disk 17 that is different from the voice code.

Now, the stored audio is played back as follows.

That is, the control section 15 reads the number and address of the active block from the control data stored on the main memory or the magnetic disk, takes out the active block from the magnetic disk 17, and inputs the active block into the buffer memory 13 or 14. If the sound block number jumps, it means that there was a silent block between them, so before taking out the sound block from the magnetic disk 17, clear the buffer memory 13 or 14 and set it to the specified value. Insert as many silent blocks as . After that, the sound block is taken out and the same operation continues. Buffer memories 13 and 14 are switch 1
6 and 18, when one is being read out to the input/output terminal 10 in DMA mode, the other is being read out to the control unit 1.
The control in step 5 switches between inputting a sound block from the magnetic disk 17 and clearing it for inputting a silent block, as in the case of audio storage.

As for control data, in addition to the method of using the block number and address for the active block mentioned above, for the active block, the address can be used.
Various methods can be used for silent blocks, such as recording the number of consecutive blocks.

In the conventional audio storage system as described above, the block size is largely determined by hardware factors such as the buffer/memory magnetic disk and the control section. That is, since magnetic disks normally perform input/output in units of sectors, the size of the buffer memory must be the same as the sector or an integral multiple thereof. In addition, in order to secure the total time required for magnetic disk access time, data transfer time, and the time required for the control unit to perform the necessary calculations and processing on the audio code in the buffer memory, input data must be input during that time. There must be buffer memory for the voice codes to be used.

The block size should originally be determined based on the characteristics of the audio signal, and should not be influenced by hardware or the like as described above.
Furthermore, if the silence block compression described above is combined with the band compression described earlier, it is conceivable that the size of blocks corresponding to the same time will decrease as band compression technology advances. Changes in block size require a reconsideration of the entire system, which does not allow rapid adoption of technological advances and is therefore unsuitable.

Note that the block size is
Although it is not impossible to make it independent of the size, it would complicate control and processing and is not practical.

Furthermore, as mentioned earlier, in addition to storing and retrieving the audio codes themselves, the control unit must also store information regarding the sound sections and silent sections, and manage the storage and playback of the audio based on this information. . Performing such detailed operations for each message, as well as managing information related to the message as a whole, such as the sender, destination, and sending time, would make the operation of the control unit significantly more complicated, and the control unit would have to control each individual message. Various restrictions will be added to the functionality.

The present invention eliminates such drawbacks in the prior art, and by adding a circuit between the buffer memory and the input terminal, the block size can be set independent of the buffer memory, magnetic disk, etc. The purpose of this invention is to insert information regarding silence blocks into the audio code, thereby obtaining a method for storing audio that does not require any involvement of the computer control section.

That is, according to the present invention, in an audio storage method that digitizes and stores audio, during storage,
The code sequence of the digitized audio signal is divided into blocks consisting of a fixed number of codes, and it is determined whether each block is a silent block or a sound block, and only the sound block is marked with a flag at the beginning of the block. Information about the time position or the number of silent blocks preceding that block is added and stored, and during playback, a flag is detected from the stored code sequence to find the beginning of the sound block, and the flag is A sound storage method is obtained which is characterized in that a silent block is inserted based on the accompanying time position or information regarding the number of preceding silent blocks, and then the stored sound block is extracted and reproduced.

The present invention will be described in detail below with reference to the drawings.
2 and 3 are diagrams showing an embodiment of the present invention, with FIG. 2 showing a portion related to audio storage, and FIG. 3 showing a portion related to audio reproduction.

In FIG. 2, the computer 12, buffer memories 13, 14, magnetic disk 17, etc. are the same as in FIG. 1, and the present invention inserts a circuit 20 between them and the audio code input/output terminal 10. This is what you get. The speech codes applied to terminal 10 are sequentially stored in buffer memory 21.
The size of this buffer memory 21 may be set to a value based on the characteristics of the speech code, and this becomes the unit of a block. However, buffer memory 21
For simplicity, the buffer memories 13 and 14, which are divided into two parts, are shown as one. On the other hand, the voice detection circuit 22 observes the level of the voice code, the number of polarity reversals, etc., determines whether the voice block in the buffer memory 21 is a sound block or a silent block, and detects the end point of the block. output the result. If it is a silent block, the silent block counter 23 is incremented by 1, and the DMA clock control circuit 24 stops sending the DMA clock to the buffer memories 13 and 14 in the computer 12.

Therefore, even if the buffer memory 21 is read, it is not written to the buffer memory in the computer, and no silence blocks are accumulated.

On the other hand, if the voice detection circuit determines that the block is a sound block, based on its output, the DMA clock control circuit 24 determines whether the voice code is 1 or 1 at the beginning of the next block.
Three DMA clocks are driven into the computer 12 during the sampling period. In cooperation with the operation of the DMA clock control circuit 24, the selection circuit 25:
A fixed pattern for the first DMA clock, for example, if the DMA input data is 8 bits, it outputs ////////. This becomes, so to speak, a flag indicating the head of the active block. Then the second
For the DMA clock, the count value of the silent block counter 23 is selected and output. This represents the number of silent blocks that existed between that active block and the previous active block.
After this, the silence block counter 23 is reset. Finally, for the third DMA clock,
The first audio code of the block output from the buffer memory 21 is selected and output. From now on, DMA
The clock control circuit 24 sends out one DMA clock during one sampling period and controls the buffer.
The output of memory 21 is sequentially written to buffer memory 13 or 14 in the computer.

Buffer memories 13 and 14 are input
According to the DMA clock, audio codes, flags, the number of silent blocks, etc. are read without being aware of their contents, and when the memory is full, the stored contents are transferred all at once to the magnetic disk 17 under the control of the control section 15. .

In this way, silent blocks are not stored, and sound blocks are stored on the magnetic disk 17 with a flag and information regarding the silent blocks added to the beginning.

If the same fixed pattern used as a flag exists in the audio code, an error will occur in detecting the beginning of a sound block and extracting information regarding a silent block during audio reproduction. To prevent this, the fixed pattern detection circuit 26 monitors the voice code, and when the same fixed pattern appears, it sends a signal to the DMA clock control circuit.
Two DMA clocks are output during the sampling period, and another fixed pattern is added and output by the selection circuit. If information about silent blocks added after the flag that matches a fixed pattern is prohibited, it is possible to distinguish between the flag and the code that matches the fixed pattern in the audio code using the method described above. It is.

FIG. 4 shows the data structure of one sound block when written to the buffer memories 13, 14 or the magnetic disk 17. However, n is the number of speech codes per block.

The DMA clock control circuit 24 has an input terminal 27
The required DMA clock is created by dividing the clock applied to the DMA clock. Further, the counter 28 counts one block of audio codes and provides the read/write address of the buffer memory 21, as well as the operation timing of each section.

On the other hand, when reproducing audio, a flag detection circuit 30 detects a flag from the code output from the buffer memory 13 or 14 of the computer 12, as shown in FIG. In other words, when the fixed pattern used for the flag is detected, the DMA
One more DMA via clock control circuit
Outputs the clock and retrieves the next code. If it does not match the fixed pattern, the previous fixed pattern is a flag, and this code extracted next is information regarding the number of silent blocks. This is input to the block counter 31, and the selection circuit 32 selects and outputs the noise generation circuit 33 to insert a silent block. Block counter 31 is a subtraction counter, and the above state continues until this counter reaches 0. During this time DMA
No clock is output from the clock control circuit 24, and no code is output from the buffer memories 13 and 14. The noise generating circuit 33 is a circuit that generates a low level noise signal corresponding to background noise in order to eliminate the unnaturalness of silent blocks. Further, the gate 34 is provided to prevent information regarding the fixed pattern and the silent block from being output to the terminal 10 when it appears.

When the block counter 31 becomes 0, it starts again.
The DMA clock control circuit 24 outputs a clock, and the sound block is taken out from the buffer memories 13 and 14. The selection circuit 32 selects the audio codes of these sound blocks inputted through the gate 34 and outputs them to the terminal 10.

Note that when the flag detection circuit 30 detects two fixed patterns in succession, it is not a flag but a voice code, as described above. Therefore, terminal 10 is used as a voice code.
output, and then repeat the above operation.

In this way, it is possible to reproduce the entire audio while correctly inserting silence blocks. In the above explanation, we have described a method of storing the number of silent blocks preceding a sound block as information regarding silent blocks, but instead, it is possible to store the time position of the sound block, that is, the serial number of the block from the beginning of the message. It is also possible to do this. The difference between the two is whether to send the block serial number or the difference.

As can be seen from the above, according to the present invention, the audio block size, which is the unit for determining whether there is a sound or no sound, is completely independent of the characteristics of the hardware such as the magnetic disk and the buffer memory in the computer. It can be determined. Therefore, the optimum block size can be set based on the characteristics of the voice, and band compression can be smoothly introduced.

In addition, information about markers indicating the beginning of blocks and silent blocks are embedded in the audio code, and the computer control unit and magnetic disk only treat them as a set of data. There is no need to worry about identification etc. Therefore, the management of voice compression/expansion and the management of the entire message can be clearly separated, and a highly flexible and simple system can be constructed, which is extremely effective. Also, the added circuits are very simple except for the voice detection circuit. With advances in digital signal processing technology and LSI technology, audio detection circuits can be implemented relatively easily, and above all, it is much simpler than implementing it in a control section within a storage device.
(IEICE Communication Method Study Group Material: CS78-69
(Refer to "Echo Suppressor for DSI") As described in the embodiments above, according to the present invention, by adding a small-scale circuit, silent section compression can be achieved autonomously without depending on the hardware of the storage device. An audio storage method that performs decompression can be obtained, and its effects are remarkable.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an audio storage method according to the prior art, FIGS. 2 and 3 are block diagrams showing an embodiment of the present invention, and FIG. 4 is a block diagram showing the data stored in FIGS. FIG. 2 is an explanatory diagram showing the data structure of In the figure, 13, 14, 21 are buffer memories, 17 is a magnetic disk, 22 is an audio detection circuit,
24 is a DMA clock control circuit, 25 and 32 are selection circuits, 23 and 31 are block counters, and 30
is a flag detection circuit.

Claims (1)

[Claims] 1. In an audio storage method in which audio is digitized and stored, during storage, the code sequence of the digitized audio signal is divided into blocks consisting of a fixed number of codes, and each block is divided into only silent blocks or sound blocks. , a flag and information regarding the time position of the block or the number of audio blocks preceding the block are added and stored at the beginning of the block, and during playback, the flag is detected from the stored code sequence and stored. The present invention is characterized in that it finds the beginning of a sound block, inserts a silent block based on the time position attached to the flag or information regarding the number of preceding silent blocks, and then extracts and reproduces the accumulated sound blocks. Audio storage method.