JPH03178252A - Voice adder circuit - Google Patents

Abstract

PURPOSE:To attain voice adder circuit suitable for miniaturization without use of lots of adders in full digital system by outputting a voice data to a digital channel synchronously with a time slot of a talker-being an output destination. CONSTITUTION:An input selection means 3 combines a 1st talker voice data inputted from a voice data input means 1 and a silence data from a silence data storage means 2, outputs the result to a voice adder means 4, which adds the 1st voice data and the silence data and outputs the result of addition to a 2-party talking voice storage means 7, in which the data is stored. A voice data of a 2nd talker is stored similarly in the 2-party talking voice storage means 7. An output selection means 9 outputs a voice data of a 2nd talker stored in the 2-party talking voice storage means 7 when a time slot of a 1st talker comes and sends the 1st talker voice data in a time slot of the 2nd talker. Thus, the voice addition by full digital system is attained without use of lots of adders and miniaturization by LSI is attained.

Description

[Detailed Description of the Invention] [Summary] The voices of two or more speakers are added together and sent to each speaker, and the voice received by each speaker does not include the speaker's own voice. The purpose of this invention is to provide an audio data input means for manually inputting multiple digital audio data, with the aim of providing an audio addition circuit that is all digital and suitable for miniaturization without using a large number of adders, regarding an audio addition circuit that adds audio. and a silent data storage means for storing silent data to be added to the voice data to which the voice of the addition partner is not input.
input selection means for inputting the audio data, the silent data, and the added mixed audio data, selecting and outputting two sets of data to be added, and outputting the selected audio data as individual audio data; and manual selection. audio adding means for adding the two sets of data output from the means and outputting the addition result as mixed audio data; and multi-party call audio holding means for holding the latest mixed audio data output from the audio adding means. and two-party call audio holding means for holding two-party call audio data in which silent data is added to the audio data output from the audio addition means in the case of a two-party call;
Individual voice holding means holds the individual voice data output from the input selection means, inputs the mixed voice data and the individual voice data of the output first speaker, subtracts the individual voice data from the mixed voice data, and outputs the result. a voice subtraction means; and an output selection means for inputting and selecting voice data from the voice subtraction means and the two-party conversation voice holding means, and outputting the voice data to a digital communication path in synchronization with a time slot of an output destination speaker. Configure to prepare.

[Industrial application field]

The present invention provides an audio addition circuit that adds the voices of two or more speakers, sends the result to each speaker, and adds the voices so that the voices received by each speaker do not include the speaker's own voice. Regarding.

In recent years, the digitalization of switching equipment has progressed rapidly.
The voice addition circuit used when making a multiparty call in a digital exchange uses a system in which digital voice is first converted into analog voice, added, and then converted back into digital voice. However, since the above method includes an analog part, it is difficult to integrate it into an LSI, which is a bottleneck in downsizing the audio adder circuit.

For this reason, an all-digital audio addition circuit is required, but a method that replaces the analog system's digital/analog converter with a gate circuit and performs digital audio addition with a configuration similar to the analog system requires only a small number of adders. However, there is a strong possibility that sufficient effects cannot be obtained in terms of miniaturization and economicalization, and an efficient digital audio adding circuit has not yet appeared.

From the above, there is a need for a digital audio adder that is suitable for LSI implementation and does not use a large number of adders.

[Conventional technology]

Fig. 3 is a block diagram of an analog audio adder circuit according to the prior art, and Fig. 4 is a block diagram of a digital audio adder circuit constructed according to the prior art. , shows the logical configuration of an audio addition circuit that adds and outputs the audio data of the three speakers so that their own audio is not included in the audio that each speaker hears. 21a to 21d in the figure are digital/analog conversion circuits (
D/A), 22a to 22d are audio adders (ADD),
23a to 23d are analog/digital conversion circuits (A/
D), and regarding the human input to the voice adders 22a to 22d, the input and output voices of speaker A are shown by solid lines, and the other voices are shown by dotted lines.

The digital voices of the four speakers A, B, C, and D are input to digital/analog conversion circuits 21a to 21d, respectively, and are converted from digital to analog voices. In the case of a four-party call, the analogized audio is sent to three audio adders 22b to 22d, the audio of speaker B is sent to audio adders 22a, 22c, and 2'2d, and the audio of speaker C is sent to audio adders 22a, 22c, and 2'2d. The voices are input to the voice adders 22a, 22b, 22d, etc., and are added to the voice adders other than the voice adder that adds the voices sent to each speaker. In order to prevent the own audio from being added in this way, the output destination audio adder is instructed by a control section (not shown) to the SEL terminals of the digital/analog conversion circuits 21a to 21d.

The voices added in the voice adders 22a to 22d are outputted to corresponding analog/digital conversion circuits 23a to 23d and digitized again. Then, the output of the same 23b is sent to speaker B as follows. This allows each speaker to simultaneously listen to the mixed voices of three other speakers, excluding his or her own voice.

However, in the audio adder circuit as shown in FIG. 3, since the audio adders 22a to 22d are analog adders, the LS
There is a problem in that it is difficult to downsize by integrating.

FIG. 4 shows the basic structure of an audio adder circuit which is completely digitalized according to the prior art in order to eliminate the drawbacks of the analog adder described above.

If the audio adder circuit is configured digitally, a digital/analog conversion circuit is not required and input can be performed using a gate circuit. However, if the input audio is, for example, 8-bit A-law PCM audio, it can be converted to, for example, 13-bit linear ( It is necessary to convert it into digital audio using the straight line rule and add it. Therefore, if the adder is a 4-bit IC, four adders 31a to 31d are needed to add the voices of speakers A and B, which divide the 13 bits into four digits and add them. (Lowercase letters in the figure indicate bit numbers).

Therefore, in order to perform a four-party call A-D, adders 32a to 32d are needed to add the voices of speakers C and D in addition to the above.
Adders 33a to 33d are required to add the addition results of A and B and the addition results of C and D.

Furthermore, a circuit is required to subtract the output destination audio from the added audio, but the subtraction circuit is an adder that uses complements, and as shown in Figure 4, there are four adders each consisting of four adders.
A set of adder circuits 34-37 is required.

In Fig. 4, the example of a four-party call of A-D is shown with the process of carrying the addition result omitted, but the process of carry can also be applied to free combinations such as addition of A and C. If , the number of adders becomes extremely large and the configuration becomes even more complicated.

As a digital audio adding circuit, a configuration method other than that shown in FIG. 4 can be used, for example, by adding A and B, then adding C, and then
A configuration method such as adding D to the addition result of B and C can be considered, but as long as the same principle as in Fig. 4 is used, it is inevitable that the number of adders will increase, and for this reason, the configuration as shown in Fig. 4 This digital audio adder circuit has not yet been put to practical use.

[Problems to be Solved by the Invention] As described above, since the audio adder circuit according to the prior art uses an analog audio adder, it cannot be converted into a digital LSI and has the problem of being difficult to miniaturize. Furthermore, even if the audio adder circuit is digitized using the conventional technology, a large number of adders are required, making it difficult to achieve the goal of miniaturization.

SUMMARY OF THE INVENTION An object of the present invention is to provide an all-digital audio adding circuit that does not use a large number of adders and is suitable for miniaturization.

[Means to solve the problem]

FIG. 1 is a diagram explaining the principle of the present invention.

In the figure, 1 is an audio data input means for inputting a plurality of digital audio data to be added, and 2 is a silent data storage for storing silent data to be added to the audio data in a state where no audio to be added is input. Means 3 inputs a plurality of voice data from the voice data input means 1, silent data from the silent data storage means 2, and added mixed voice data from the voice adding means 4, and selects two sets of data to be added. 4 adds the two sets of data outputted from the input selection means 3, and converts the addition result into mixed audio data. 5 is a multi-party call audio holding means for holding the latest mixed audio data outputted from the audio addition means 4; 6 is an audio addition means 4 outputted from the audio addition means 4 in the case of a two-party call. 2-party call audio holding means for holding 2-party call audio data in which audio data and silent data are added; 7 is an individual audio holding unit for holding the individual audio data output from the input selection means 3; 8 is an individual audio holding means for holding the individual audio data output from the input selection means 3 audio subtraction means for inputting mixed audio data from the multi-party call audio holding means 5 and individual audio data from the individual audio holding means 7, subtracting individual audio data of a destination speaker from the mixed audio data and outputting the result; 9 manually selects the mixed voice data and the voice data for the two-party conversation from the voice subtraction means 8 and the two-party conversation voice holding means 6, selects the voice data to be sent to the output destination speaker, and outputs the voice data. This is an output selection means that outputs the output to the digital channel in synchronization with the time slot of the previous speaker.

[For production]

In FIG. 1, a plurality of pieces of audio data to be added are input to the audio data input means 1, and are converted into straight line code. The plurality of voice data are manually input into the manual selection means 3 in serial order, but the first voice data that is input first is input from the silent data storage means 4 because the voice data of the addition partner has not been input yet. It is combined with the silent data and output to the audio adding means 4, where it is added and output as mixed audio data. However, at this time, the matching audio data is the same as only the i-th audio data. At this time, the first audio data outputted by the input selection means 3 for addition is outputted as individual audio data to the individual audio holding means 6 and held therein.

The mixed voice data added by the voice adding means 4 is held in the multiparty voice holding means 5 and outputted to the input selection means 3. Next, when the second audio data is input to the input selection means 3, the second audio data is added to the audio addition means 4 together with the mixed audio data added in the audio addition means 4, that is, the first audio data. is output to and added to. The addition result in the voice adding means 4 is output to the multiparty voice holding means 5, and the latest voice data is held in a form overlaid on the previously held mixed voice data. Further, the second voice data is also held as individual voice data by the human selection means 3 in the individual voice holding means 6 together with the first voice data.

Furthermore, when adding the third audio data, similarly to the above,
The output from the voice addition means 4 is input again to the input selection means 3, added to the third voice data and held in the multiparty call voice holding means 5, and the third voice data is also sent to the individual voice holding means 6. Retained.

When all the additions are completed in this way, for example, when outputting the mixed voice data to the first speaker, the multiparty voice holding means 5
The final mixed voice data held in the final mixed voice data and the individual voice data of the first speaker held in the individual voice holding means 6 are input to the voice subtraction means 8, and the individual voice data is subtracted from the mixed voice data. It outputs to the output selection means 9. The output selection means 9 outputs the subtracted mixed voice data for the time slot for the first speaker of the digital communication path.

Therefore, mixed audio data excluding the i-th speaker's audio data is sent to the first speaker.

Above, the voice addition of three or more parties has been described, but an example of a call between two parties will be described.

Audio adder circuits are used for multiparty calls such as conference calls, and are also used for monitoring at switchboards. For example, when two parties make a call via a switchboard, an operator at the switchboard interrupts the three-party call and monitors the call between the two parties to confirm whether the call is being conducted without any problems. (Monitoring is also a type of three-party call) After the operation, a two-party call is performed through the voice addition circuit in a state where the operator no longer intervenes. In a two-party call as described above, each party only needs to receive the other party's voice, so there is no need to add multiple pieces of voice data, and there is no need to subtract voice data. It is more efficient to separate the processing means.

In FIG. 1, when information indicating a two-party call is input manually by a means not shown, the input selection means 3 selects the voice data of the first speaker inputted from the voice data input means 1 to the silent data storage means. The voice adding means 4 adds the first voice data and the silent data and outputs the addition result to the two-party conversation voice holding means 7 for holding. . The voice data of the second speaker is similarly held in the two-party conversation voice holding means 7.

The output selection means 9 outputs the second speaker's voice data held in the two-party conversation voice holding means 7 when the first speaker's time slot arrives, Then, the first speaker's voice data is transmitted. As a result, the first and second speakers hear only the other party's voice. Note that the output selection means 9 also selects the output from the voice subtraction means 8 and the output from the two-party voice holding means 7 in the case of the multi-party conversation.

As described above, in the configuration shown in FIG. 1, voice addition is performed using an all-digital system, and two sets of addition means, namely, voice addition means and voice subtraction means (binary number addition means using complements), are used to perform multi-digit addition. It is possible to make phone calls and two-party calls, and there is no need to use multiple adders. Therefore, miniaturization by LSI can be easily achieved.

〔Example〕

FIG. 2 shows the constituent countries of one embodiment of the present invention.

In the figure, 11 is a RAM that manually stores voice data temporarily.
(random access memory), 12 is an 8-hit PC
A code conversion circuit converts M audio data into 13-bit linear law audio data, ■3 is a silent data memory that stores silent data, and ■4 to 16 and 25 are selectors (SELI).
~5EL3,5EL4), 17 is an adder (ADD+
), 18 to 20 are latch circuits, 2I is a negative logic circuit (NOT), and 22 is an adder (AD) for audio data subtraction.
D2 ), 23 and 24 are bit conversion circuits that convert the number of bits, and 26 is a code conversion circuit that converts the straight line rule into PCM.

FIG. 2 will be explained using a four-party call as an example.

The PCM voice data of each of the four speakers is input to the RAM 11 for one frame of the digital communication path, and -
time is remembered. The audio data is serially read out under the control of a control section (not shown), input to the code conversion circuit 12, and converted from 8-bit PCM audio data to 13-bit linear audio data. Note that the code conversion circuit 12 is constituted by a ROM (read-only memory).

The voice data of the i-th speaker outputted from the code conversion circuit 12 is inputted to the adder 7 via the selector 14, and at the same time, silent data from the silent data memory 13 is inputted to the adder 17 via the selector 15. and are added in the adder 17. The addition result is held in the latch circuit 18 and is again input to the adder 17 via the selector 14. Also, at this time, the first audio data is selected by the selector I6.
The signal is held in the latch circuit 19 via the latch circuit 19.

Next, when the second speaker's voice data is output from the code conversion circuit 12, it is input to the adder 17 via the selector 15, and the first voice data (silent data) that was previously input to the selector 14 is (mixed audio data) with adder 1
7 is added. At this time, the second audio data is held in the latch circuit 19 via the selector 6, and the addition result is held in the latch circuit 18 as an overlay. Below, 3rd. The voice data of the first speaker is added in the same way, and when the voice data for one frame of the digital communication channel has been input, the mixed voice data of the four speakers is held in the latch circuit 18. Further, the first to fourth audio data are individually held in the latch circuit 19.

In this state, when the time slot of, for example, the third speaker arrives on the digital communication path on the output side, the mixed voice data held in the latch circuit 18 is input to the adder 22 under the control of a control section (not shown). be done. Further, the voice data of the first speaker held in the latch circuit I9 is converted into logic data "i" and "0" in the negative logic circuit, and is input to the bit conversion circuit 23. Latch circuit 19
The audio data held in the 18 bits is 13 bits as mentioned above, but the audio data of the 4 parties is added and the latch circuit 18
Since the mixed audio data held in the latch circuit 19 has a maximum of 16 bits, the audio data held in the latch circuit 19 is converted from 13 bits to 16 bits in the bit conversion circuit 23.

As a result of the above, the logically converted data of the mixed voice data held in the latch circuit 18 and the first speaker's voice data held in the latch circuit 19 are input to the adder circuit, and both data are combined. The addition is followed by subtraction using a well-known binary number, which is sent to the code conversion circuit 26 via the selector 25, where it is converted into an 8-bit PC from the 16-bit linear law data.
It is converted into M data and output to a digital communication channel.

That is, mixed audio data excluding the first speaker's voice is sent to the first speaker.

Next, a two-party call will be explained.

In FIG. 2, when information indicating a two-party call is manually entered from a control unit (not shown), the adder 17 is inputted by the selector 14.
Add the voice data of the first speaker inputted via the selector 15 to the silent data inputted from the silent data memo 1 river 3, and add the addition result, that is, the voice data of the first speaker. It is output to the latch circuit 20 and held, and the second speaker's voice data is also held in the latch circuit 20. The selector 25 outputs the second speaker's voice data held in the latch circuit 20 when the first speaker's time slot arrives, and also outputs the second speaker's voice data held in the latch circuit 20 when the first speaker's time slot arrives. Send audio data. Note that the audio data held in the latch circuit 20 is 14 bits by adding 13 bits of audio data and silent data, but when outputting, the bit conversion circuit 24 converts the mixed audio data of the multiparty call into 14 bits. The code conversion circuit 26 converts the straight-line 16 bits into 8-bit PC.
Convert to M.

As described above, the first and second speakers listen only to the other party's voice. In the configuration shown in FIG. 2, two sets of two-party calls can be made at the same time.

In addition to the above, in the embodiment shown in FIG. 2, it is also possible to carry out a process in which, for example, a first speaker and a second speaker communicate, and an operator monitors the conversation. In this case, it would be possible to treat the operator as a third speaker and treat the audio data as simply silence, but the monitor could receive some information and record silence instead of the third speaker's audio data. It is possible to input silent data from the data memory. Thereby, it is possible to completely prevent noise caused by the monitor from being caused to the first and second speakers.

The configuration of the embodiment has been explained above using four-party calls and two-party calls as examples. However, by changing the number of input audio data of RAM II, adders 17, 22, latch circuits 18 to 20, and other processing bit numbers, It is clear that the present invention is applicable to multi-party calls involving more than one person, and the present invention does not exclude these modifications.

〔Effect of the invention〕

As described above, according to the present invention, it is possible to perform audio addition using an all-digital method without using a large number of adders, and this invention has a great effect in realizing miniaturization of such an audio addition circuit by implementing an LSI.

[Brief explanation of drawings]

Fig. 1 is a diagram explaining the principle of the present invention, Fig. 2 is a block diagram of an embodiment of the present invention, Fig. 3 is a block diagram of the conventional technology (analog), and Fig. 4 is a diagram explaining the principle of the present invention.
The figure is a configuration diagram of a conventional technology (digital). In the figure, 1---1... Voice data input means 2---... ・-
---1 Silent data storage means 3-------1--1 Manual selection means 4--1----Audio addition means 5--1--Multiple party voice holding means 6 ・
-・・−・−・・−Individual voice holding means 7−・−−m−・
--One/-Two-party call voice holding means 8 --------Audio subtraction means 9 ---Output selection means 4゜Ih Y':h % spare time 4 points \ゝ“-)l≧〉, Tokomi is poor l! PuEt≦1f treasure 3
ZConventional technology (analog) configuration diagram

Claims (1)

[Scope of Claims] Audio data input means (1) for inputting a plurality of digital audio data to be added; and storing silent data to be added to audio data in a state where no audio to be added is input. a silent data storage means (2); a plurality of audio data from the audio data input means (1);
Input the silent data from the silent data storage means (2) and the added mixed audio data from the audio adding means (4), select and output two sets of data to be added, and output the audio selected for addition. input selection means (3) for outputting data as individual audio data;
an audio addition means (4) for adding the two sets of data output from the above and outputting the addition result as mixed audio data; and a multiplexer for holding the latest mixed audio data output from the audio addition means (4). a two-party call audio holding means (5), and a two-party call audio holding means for holding two-party call audio data in which the audio data output from the audio addition means (4) and silence data are added in the case of a two-party call. means (6); individual voice holding means (7) for holding the individual voice data outputted from the input selection means (3); mixed voice data from the multiparty call voice holding means (5); a voice subtraction means (8) that inputs individual voice data of a destination speaker from the voice holding means (7), subtracts the individual voice data from the mixed voice data, and outputs the result; Two-party call audio retention means (
6) Input the mixed audio data and the audio data for two-party communication, select the audio data to be sent to the output destination speaker, synchronize the audio data with the time slot of the output destination speaker, and connect the digital communication channel. 1. An audio addition circuit comprising: an output selection means (9) for outputting an output to