JPS6019178B2 - Simultaneous connection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a simultaneous connection system that connects multiple channels simultaneously, such as in a conference call, by processing call information from multiple channels.

Call information from multiple channels is collected from subscriber voice samples (PC
M), that is, in the case of digital transmission, the configuration shown in FIG. 1 is known in the conventional simultaneous connection system for a plurality of channels.

In FIG. 1, reference numeral 2 denotes an expander EXP, which expands the voice samples from the subscriber, which have been non-linearly compressed to expand the dynamic range, into linear samples to simplify multiplication, addition, etc. 9 is an OMP compressor, which after simultaneous connection processing compresses the voice sample again so as to match the voice sample in the communication path.

3 and 8 are multipliers, 3 multiplies the size of the subscriber's voice sample by a certain coefficient (1/4 in this example) to reduce it, and 8
are connected simultaneously, and then multiplied again by a certain coefficient (4 in this example) to return to the original size.

A shift register SRI of 5 and an adder of 4 add the multiplied and reduced subscriber voice samples.

Gate 6 GI and shift register SR3 7 send the summed audio samples to each subscriber's channel.

17 shift registers SR2 and 16 adders subtract the voice samples of the subscribers of that channel from the voice samples sent to each channel.

The conventional simultaneous connection system with the above configuration will be explained with reference to FIGS. 5 and 7.

FIG. 5 shows the flow of signals in each part, and FIG. 7 shows the magnitude of the signals.

That is, the decompressor 2 receives voice samples (16
Bit configuration) Appears in serials sequentially from cancer.

Each of these is multiplied by a constant coefficient 1/4 by a multiplier 3.
As a result, a signal as shown in FIG. 7b is input to the adder 4. This adder 4 adds the output of the shift register 5 and the output of the multiplier 3.
For 6 bits, the output of the shift register 5 is not added due to the signal Adderl bit. Therefore, the information accumulated in the shift register 5 is the SRI input in FIG.
n=4).

In this manner, when the addition process for all conference participants is completed, the gate 6 is opened by the clock, the contents of the shift register 5 are transferred to the shift register 7, and the gate 6 is closed again.

The output of this shift register 7 is SR3 in FIG.
The output and the signal shown in FIG. 7c appear. Incidentally, R: is also ten al treasure a2 ten a3. On the other hand, shift register 17
As a result, the signal shown at the SRI input in FIG. 5 is delayed and appears in synchronization with the SR3 output.

Therefore, the adder 16 subtracts the output of the shift register 17 from the output of the shift register 7.

The output of this adder 16 is shown as the output in FIG. 7d and FIG. 516. This is to prevent so-called sidetone, which prevents the sound emitted by each subscriber from being returned to that subscriber. The outputs of the adders 16 are each quadrupled by the multipliers 8. The output of this multiplier 8 is shown in FIG. 7e. The COMP input shown in FIG. 5 is for explaining one embodiment of the present invention shown in FIG. 2, and will be described later. Here, the output from the multiplier 8 is nonlinearly compressed again by the compressor 9 and distributed to each subscriber, that is, the signal e is distributed to the subscriber Ao, the signal A is distributed to the subscriber A, and the signal is distributed to the subscriber A. be done. In such a simultaneous connection where the voice sample is constantly multiplied by a constant coefficient (1/4 in Figure 1), the voice sample is reduced by one degree even if there is only one subscriber in the call, and the signal-to-noise power ratio is reduced due to rounding, etc. becomes smaller, and as a result of subsequent addition processing, noise is also added, and even if the signal is restored to its original size, the signal-to-noise power ratio is not improved, resulting in poor speech quality.

Also, although Fig. 8 a to e correspond to Fig. 7 a to e, they show cases where the audio level is high, and as shown in Fig. 8, there are many subscribers talking, In addition, when each audio sample is just before overflow, it is once reduced and added up and sent to multiple subscribers who are trying to connect at the same time using multiplier 8. However, when returning to the original size, the result of adding up the voice samples of a large number of talking subscribers exceeds the overload level as shown in FIG. 8e, causing overflow and overload noise. Therefore, there is a drawback that the call quality deteriorates. The present invention has improved such drawbacks, and provides an apparatus for simultaneously connecting multiple subscribers by performing arithmetic processing such as multiplication and addition on digital voice samples of multiple subscribers. The subscriber who is on a call is determined based on the voice samples from multiple subscribers, and the operating function of the arithmetic processing of the voice samples of the multiple subscribers or the subscriber who is on a call is changed according to the number of the subscribers, and a certain number of the subscribers are determined. The objective is to control deterioration in call quality by establishing simultaneous connections between subscribers. In other words, in a device that connects multiple channels simultaneously, all or part of the call information from the multiple channels is transmitted depending on the number of active channels of the multiple channels that are simultaneously communicating. By changing the arithmetic processing method, the voice Fujimoto is always multiplied by a constant coefficient, resulting in an increase in noise due to rounding, etc., and a deterioration in call quality due to the addition of noise due to addition. The goal is to reduce the deterioration of call quality due to overload noise resulting from overflow. FIG. 2 is a block diagram of an embodiment of the present invention, and the same reference numerals as in FIG. 1 have the same functions.

The logic circuit 11 determines the number of subscribers who are talking based on the size of voice samples from a plurality of subscribers, and stores the number of subscribers who are talking in one frame time. Reference numeral 12 is a register that stores coefficients to be multiplied by the audio sample, and reference numeral 13 is a selector that selects the coefficient based on a signal from a logic circuit.

The operation of FIG. 2 will be explained with reference to the time chart of FIG. 5 and FIG. 9 for the case of simultaneous connection when the number of multiple subscribers is four.

To explain the operation of the logic circuit No. 11, it is assumed that Ao subscribers are simultaneously talking in a certain frame, as shown in FIG. 9a.

In Figure 5, this is the size of the subscriber's voice sample, a, = 0
, a2=0. At this time, the logic circuit 11
This identifies that the number of subscribers in the path is 2 and stores it until the next frame. Thereafter, the stored contents are transferred to the register 12, and the number of subscribers is identified again. In FIG. 5, n indicates the number of subscribers in communication, and this corresponds to n being 2. And when time slot 0, select evening 13
Since n is 2, the coefficient by which the voice sample from the subscriber is multiplied, that is, 1/n=1/2, is sent from register 2 to multiplier 3, and the subscriber's voice sample is calculated as shown in Fig. 9b. Make it smaller like . This is added to the audio sample of the next time slot using a shift register 5 and an adder 4. This addition result is stored in the shift register 7 and sent to each subscriber's channel as shown in the SR3 output (however, a,=a2=0) in FIG. 5, that is, as shown in FIG. Figure 16
The voice samples of the subscribers of each channel are subtracted as shown in the output of the compressor 9, and the signal shown in the COMP input of FIG.
The data is compressed to allow simultaneous connection of multiple subscribers. However, R = Sangei San. In this way, the frequency of change in the coefficient multiplied by the voice sample is reduced. As a result, there are many changes in call quality, which reduces the subscriber's discomfort. Furthermore, even if there is a large audio input, the signal will not become overly heavy and will not exceed the bell level, so there will be no quality deterioration. Further, FIG. 3 shows another embodiment of the present invention in which the order of arithmetic processing is different from that in FIG. 2, and the same reference numerals as in FIG. 2 indicate the same functions.

This operation will be explained with reference to FIGS. 3, 6, and 10. In a certain frame, when subscribers Ao, A,... are talking at the same time as shown in Figure 10, the logic circuit 11
Therefore, since there is no path for subscriber A2, that is, a2=0, the number n of subscribers currently talking is 3, and this is stored. However, in this example, it is necessary that the signals between the subscribers be separated by more than the distance between the conference subscribers, and the shift registers 5' and 7' are also 64 bits, which is equal to the number of subscribers in the case of FIG. It is the composition.
The contents stored in the logic circuit 11 are then stored in the register 12 by sending a signal to the selector 13 so that the coefficient by which the voice sample is multiplied by 1/n, that is, 1'3, at time slot 0'. At that time, each audio sample is added one after another by the shift register 5' and the adder 4 as shown in the SRI input of FIG. 6, and a large signal (R=a.
, 10a3). After this is sent to the shift register 7', the voice samples of each subscriber are subtracted, and the multiplier 8 multiplies each voice sample by the coefficient 1/3 sent from the register 12.
By reducing the size of the audio sample as shown in Figure 10c,
Simultaneous connection operations for multiple subscribers are performed. Here, the 10th
As shown in Figure b, the overload level shown in Figures 1 and 2 is exceeded, but this is because the shift registers 5' and 7' are expanded to 64 bits, which is twice the number of subscribers. The number of subscribers will also be doubled, and the characteristics will not deteriorate. Another simultaneous connection method according to the present invention is the method shown in FIG. That is, in the figure, 18 has a selection function, and 19 has a simultaneous connection function. This method has a logic circuit that selects a busy channel for simultaneous connection processing from among a plurality of channels for simultaneous connection based on its call information, and temporarily stores the selected channel. When there is a maximum of m call channel inputs that can be processed at the same time and are greater than the number of active subscribers (in the figure), when m>sleeve, the simultaneous connection processing is performed by comparing each call information. After selectively limiting the number of active channels that can be connected, simultaneous connection is performed by performing simultaneous connection processing as in the embodiments shown in FIGS. 1, 2, and 3. In other words, if we simply increase the number of calls, when the number of active subscribers n becomes equal in the simultaneous connections shown in Figures 1, 2, and 3, the coefficient 1 that is multiplied by the voice sample of the active subscriber When /n, that is, 1/night, becomes small, the audio sample becomes excessively small. Therefore, by limiting the number of subscribers m′ who are currently talking to more than
Simultaneous connections with good call quality can be made regardless of the situation. An embodiment of this simultaneous connection system is shown in FIG. In the figure, the same symbols as in FIG. 2 indicate the same functions. The logic circuit 14 in this figure selects a subscriber for simultaneous connection processing by comparing the sizes of the voice samples of the subscribers who are talking, and opens and closes the gate 15 in accordance with the signal from the logic circuit 14. In the embodiment shown in FIG. 4, when there is a number m' of active subscribers that is greater than the number of active subscribers that can be processed for simultaneous connection, the active subscribers that can be processed for simultaneous connection in descending order of the size of each audio sample are It has the feature of selecting only the number of subscribers and allowing only the voice samples of the selected subscribers to pass through the gate 15 to perform simultaneous connection processing.

Therefore, even if the number of subscribers who wish to connect at the same time changes, the number of subscribers can be limited to a number that allows simultaneous connection with good speech quality. By adopting the method explained above, that is, by changing the arithmetic processing method according to the change in the number of channels in call, in the above embodiment, when there is one subscriber in call, that is, n=1. Then, the coefficient by which the voice sample is multiplied is 1/n, that is, 1, so that the voice sample is never made smaller, and an increase in noise is suppressed.

When the number of active subscribers is 4, each voice sample is reduced to 1/4 during arithmetic processing, so overflow does not occur. Therefore, overload noise is suppressed and speech quality is improved. Furthermore, in response to a change in the number of active channels of multiple channels to be connected at the same time, by performing simultaneous connection processing according to the number of active calls, the above-mentioned embodiment can reduce the subscriber's voice sample excessively. It is possible to prevent the signal-to-noise power ratio from decreasing due to This allows for simultaneous reinforcement with good call quality.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram of a device using a conventional simultaneous connection method. FIGS. 2 to 4 are configuration diagrams of an embodiment of the present invention. FIG. 5 is a time chart explaining the operation of FIG. 2. FIG. 6 is a time chart explaining the operation of FIG. 3. In FIGS. 5 and 6, Ao, A, , A2, A3 represent each subscriber, ao, a
, , a2, a3 indicate the size of each subscriber's voice sample.
FIG. 7 is an explanatory diagram of the operation of FIG. 1. FIG. 8 is an explanatory diagram of the operation when overload noise occurs in the embodiment of FIG. 1. FIG. 9 is an explanatory diagram of the operation of FIG. 2. FIG. 10 is an explanatory diagram of the operation of FIG. 3. FIG. 11 is a diagram of a space division method equivalent to FIG. 4. In the figure,
2 is an expander, 3 and 4 are multipliers, 4.16 is an adder, 5,
5', 7, 7' and 17 are shift registers, 6 and 15 are gates, 9 is a compressor, and 11 and 11 are logic circuits. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11

Claims (1)

[Scope of Claims] 1. In a device that performs simultaneous connection by performing arithmetic processing on digital voice samples of subscribers of multiple channels, a circuit for determining whether a subscriber is on a call based on voice samples from subscribers of the multiple channels; , a circuit for changing the operating function of the arithmetic processing of the voice sample is provided, and the arithmetic processing method is changed according to the number of active channels among the plurality of channels that are simultaneously communicating. Simultaneous connection method. 2. In a device that performs simultaneous connection by performing arithmetic processing on digital voice samples of subscribers of multiple channels, after selecting a channel to perform simultaneous connection processing from among the multiple channels in use, the selected channel 2. The simultaneous connection system according to claim 1, wherein arithmetic processing for simultaneous connection is performed for the following.