JPH11163888A - Voice coding transmitter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice coding transmitter by which a 'voiced sound' signal at a speech head is completely and surely sent. SOLUTION: A voice coding section 12 generates and outputs voice data resulting from coding a voice signal. A storage output section 13 stores voice data outputted from the voice coder 12 for a prescribed time and provides an output of them. A cell processing section 14 obtains voice cells resulting from assembling voice data into cells outputted from a storage output section 13 and provides an output. An idle cell generating section 15 generates an idle cell sent for management of a transmission line only. A voice level detection section 16 detects a voice level denoting the strength of the voice signal. A voiced sound discrimination section 17 discriminates whether or not the detected voice level reaches a reference voice level. A selection section 18 selects either the voice cell or the idle cell as a transmission cell according to the discrimination result. A transmission processing section 19 sends the selected transmission cell to a network.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice coded transmission apparatus for transmitting a coded voice signal to a transmission line,
More specifically, the present invention relates to a voice coded transmission device that transmits only a sound part of a given voice signal to a transmission path.

[0002]

2. Description of the Related Art In an ATM network for transmitting a cell of a predetermined size in an asynchronous communication mode (Asynchronous Transfer Mode), when congestion occurs in the network during communication, a part of the cell being transmitted at that time. Is to be discarded to suppress the expansion of congestion.
When a cell is discarded, retransmission of a packet including the cell is performed, thereby increasing the risk of congestion due to the retransmission, and extending the communication time due to the retransmission, resulting in data transmission. It is generally known that transmission efficiency is reduced.

[0003] In recent years, a technique for transmitting an encoded voice signal via an ATM network has been generalized. Therefore, various techniques relating to congestion in a network especially for audio data obtained by encoding an audio signal have been proposed and actually used. One such technique is called silence compression. In silence compression, the strength of a given audio signal is compared with a predetermined criterion, and those that satisfy this criterion are referred to as "voiced" audio signals,
Those that do not satisfy the condition are referred to as “silent” audio signals. And
For a "voiced" voice signal, the entire contents are encoded and transmitted to the network. For a "voiceless" voice signal, a cell is transmitted that only indicates "silence" and transmitted to the network.

[0004] In a communication system to which this silence compression is applied, consider a case where a conversation is made between a first telephone and a second telephone connected via, for example, an ATM network. In this case, in a normal conversation, when the user on the first telephone side is talking, the user on the second telephone side will naturally listen silently. In other words, it is expected that the “voiceless” voice signal is continuously provided on the first telephone side while the “silent” voice signal is provided on the second telephone side. Therefore, for the "voiced" voice signal given by the first telephone, cells in which the entire contents are encoded are transmitted to the network, and the second telephone calls the original "voice" from these cells. Play the sound signal of "sound". At the same time, for the "silent" voice signal given on the second telephone side, a cell is generated to transmit only "silence" and transmitted to the network, and the first telephone responds to this cell. Restore the "silent" audio signal. By performing silence compression processing as described above, ATM
Many of the cells transmitted on the network are cells encoded from the "voiced" voice signal on the side of the telephone where the user is talking, and a small portion of the remaining cells are on the side of the telephone where the user is listening Cell that conveys "silence". Therefore, by applying the silence compression, the traffic in the network is reduced by half, and the user hardly feels any particular inconvenience in terms of convenience.

[0005] As a prior art relating to congestion in a network on the assumption that the above-mentioned silence compression is applied, Japanese Patent Application Laid-Open No. H05-205,051 discloses a conventional technique.
No. 244104, “Speech encoder” (hereinafter referred to as “prior art”), Japanese Unexamined Patent Publication No. 6-69950, “Speech coded transmission device” (hereinafter referred to as “prior art”),
Japanese Unexamined Patent Publication No. 8-8933 discloses a "speech cell coding apparatus" (hereinafter referred to as "prior art").

In the prior art, the degree of audio data compression (including silence compression) is increased or decreased following the cell storage amount of a transmission buffer for storing cells to be transmitted to a network. According to this conventional technique, the degree of compression of audio data is controlled to increase only when the cell storage amount of the transmission buffer is large due to network congestion. Deterioration of voice quality can be suppressed only when congestion occurs in the network, and it is possible to maintain voice quality when transmitting voice data in the ATM network.

In the prior art, a priority is added to each cell to be transmitted to the network, and when congestion occurs in the network, some cells have to be discarded so that the cells are discarded according to the priority. It was made. According to this prior art, the priority when discarded is determined by a cell composed of auxiliary data, a cell composed of data representing silence, a cell other than these,
Thus, the influence on the audio quality of the audio signal reproduced on the receiving side when the cell is discarded can be reduced.

In the prior art, when a part of cells constituting voice data is discarded and lost during transmission on the ATM network, information of the lost cell is interpolated from cells before and after normal arrival. is there. According to the prior art, on the transmitting side, a selection signal indicating an optimal interpolation method for interpolating information of each cell is multiplexed and transmitted to a corresponding cell, and when a cell is missing on the receiving side, Since the information of the lost cell can be interpolated by the optimal interpolation method according to the selection signal, when the information is lost due to the discard of the cell, the quality deterioration of the audio signal reproduced on the receiving side can be suppressed. .

[0009]

When the above-described silent compression is realized, for example, the strength of a given audio signal is compared with the strength of a predetermined reference, and the strength of the reference signal exceeds the strength of this reference. Must be detected. Then, according to the detection result, select one of the cell that encodes the audio signal and the cell that transmits “silence”,
Must be sent to the network. Therefore, depending on the accuracy of the above-described comparison and detection, the timing of cell selection according to the detection result, and the like, the very beginning of the speech, that is, a very short period during which the speech signal given at the beginning of the speech transitions from “silence” to “voiced” However, there is a problem that an introduction portion of a voice signal that should be "voiced" is "silent" and may not be transmitted to the network. When such a phenomenon occurs, for example, if the voice given at the beginning of the speech has a short and sharp consonant such as "Sashisuse Soso" or "Hahifueho", the way of transmitting this voice to the receiving side becomes unnatural. And the intended words are not clearly communicated,
There was a problem that unintended words could be transmitted by mistake.

On the other hand, the above-mentioned prior arts are all based on silent compression, and aim at improving the quality of an audio signal reproduced from a cell transmitted to a receiving side. According to the prior art, it is possible to prevent disconnection of the head or tail by preserving as much as possible the cell that transmits “silence”. However, this prior art is intended to prevent head-off or tail-cutting caused by cell discarding due to network congestion. It does not prevent the beginning of the talk or the end of the talk from being interrupted when the speaker begins to speak. In addition, there is no description in the prior art and the speech head.

SUMMARY OF THE INVENTION An object of the present invention is to provide a speech coded transmission apparatus in which a "voiced" speech signal at the beginning of a speech is completely and reliably transmitted to a network.

[0012]

According to the present invention, (a) a delay processing section for outputting a second audio signal obtained by delaying the first audio signal, and (b) this second audio signal An audio encoding unit that generates audio data obtained by encoding the signal;
(C) a cell processing section for outputting a voice cell obtained by converting the voice data generated by the voice coding section into cells, and (iv) a cell processing when the first voice signal satisfies a predetermined criterion. The speech coded transmission device is provided with a selection unit that selects a speech cell output from the processing unit as a transmission cell.

That is, according to the first aspect of the present invention, audio data obtained by encoding the second audio signal is generated from the second audio signal obtained by delaying the first audio signal.
A voice cell is obtained by converting the voice data into a cell. Then, when the first audio signal satisfies the predetermined criterion, the audio cell obtained from the second audio signal is selected as a transmission cell. For this reason, when the first audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, the first audio signal after the time corresponding to the above-described delay from the time when this change is actually detected is detected. A voice cell corresponding to the content of the signal is selected as a transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as the transmission cell, the "voiced" voice signal at the beginning of the talk can be transmitted completely and reliably.

According to the second aspect of the present invention, (a) a delay processing section for outputting a second audio signal obtained by delaying the first audio signal, and (b) an audio obtained by encoding the second audio signal A voice coding unit for generating data; (c) a cell processing unit for outputting voice cells obtained by converting the voice data generated by the voice coding unit into cells; and (d) a voice signal transmitted only for managing a transmission path. (E) selecting a voice cell output from the cell processing unit as a transmission cell when the first voice signal satisfies a predetermined criterion; When the first criterion signal does not satisfy the criterion, the speech coded transmission device is provided with a selection unit for selecting a vacant cell generated by the vacant cell generation unit as a transmission cell.

That is, according to the second aspect of the present invention, audio data obtained by encoding the second audio signal is generated from the second audio signal obtained by delaying the first audio signal.
A voice cell is obtained by converting the voice data into a cell. In addition, an empty cell generation unit generates an empty cell to be transmitted only for managing the transmission path. Then, when the first audio signal satisfies the predetermined criterion, the audio cell obtained from the second audio signal is selected as a transmission cell. A cell is selected as a transmission cell. For this reason, when the first audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, the first sound signal is not changed from the time when the change is actually detected to the time after the time corresponding to the delay described above. A voice cell corresponding to the content of the voice signal is selected as a transmission cell, and before that, an empty cell is selected as a transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as the transmission cell, the "voiced" voice signal at the beginning of the talk can be transmitted completely and reliably.

According to the third aspect of the present invention, (a) audio data corresponding to an audio signal is stored at a first time, and the audio data is output at a second time after the first time. An accumulating output unit, (b) a cell processing unit for obtaining and outputting an audio cell obtained by converting the audio data output by the accumulating output unit into cells, and (c) when the audio signal satisfies a predetermined criterion, The speech coded transmission device is provided with a selection unit for selecting a speech cell output from the cell processing unit as a transmission cell.

That is, according to the third aspect of the present invention, after the audio data corresponding to the audio signal is stored at the first time,
At the second time point after the first time point, the cell is formed and a corresponding voice cell is obtained. Then, when the voice signal satisfies the predetermined criterion, the voice cell obtained at the second time is selected as a transmission cell. Therefore, when the audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, the content of the audio signal after the time corresponding to the time corresponding to the difference between the first and second time points is The corresponding voice cell is selected as a transmission cell.
Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as the transmission cell, the "voiced" voice signal at the beginning of the talk can be transmitted completely and reliably.

According to the fourth aspect of the present invention, (a) audio data corresponding to an audio signal is stored at a first time, and the audio data is output at a second time after the first time. An accumulator / output unit; (b) a cell processing unit for obtaining and outputting a cell of the audio data output from the accumulator / output unit; and (c) an empty cell transmitted only for managing the transmission path. An empty cell generation unit to be generated; and (d) when the audio signal satisfies a predetermined criterion, an audio cell output from the cell processing unit is selected as a transmission cell, and the audio signal satisfies the criterion. Is not satisfied, the speech encoding transmission apparatus is provided with a selection unit for selecting the empty cell generated by the empty cell generation unit as a transmission cell.

That is, after the audio data corresponding to the audio signal is stored at the first point in time,
At the second time point after the first time point, the cell is formed and a corresponding voice cell is obtained. In addition, an empty cell generation unit generates an empty cell to be transmitted only for managing the transmission path. When the voice signal satisfies the predetermined criterion, the voice cell determined at the second time is selected as a transmission cell, while when the voice signal does not satisfy the criterion, an empty cell is selected as a transmission cell. You. Therefore, when the audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, after the time corresponding to the time corresponding to the difference between the first and second time points, Is selected as a transmission cell, and an empty cell before that is selected as a transmission cell. Therefore, for example, from "silence" to "voiced"
When detecting the change to the above, even if the time when the above change is actually detected is not exactly accurate, a voice cell corresponding to the content of the voice signal after a time immediately before that is selected as a transmission cell. Can be completely and reliably transmitted.

According to the fifth aspect of the present invention, (a) an audio encoding section for generating and outputting audio data obtained by encoding an audio signal, and (b) the audio data output from the audio encoding section is converted into a first audio data. A storage output unit that stores the audio data at a time point and outputs the audio data at a second time point after the first time point, and (c) obtains a voice cell obtained by converting the audio data output by the storage output unit into cells. A cell processing unit for outputting
(D) When the voice signal satisfies a predetermined criterion, the voice coded transmission device is provided with a selection unit for selecting a voice cell output from the cell processing unit as a transmission cell.

That is, according to the fifth aspect of the present invention, after the audio data obtained by the encoding of the audio signal by the audio encoding unit is stored at the first time, the second data after the first time is stored.
At the point of time, a corresponding voice cell is obtained. Then, when the voice signal satisfies the predetermined criterion, the voice cell obtained at the second time is selected as a transmission cell. Therefore, when the audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion,
The voice cell corresponding to the content of the voice signal after the time point advanced by the time corresponding to the difference between the first and second time points is selected as the transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as the transmission cell, the "voiced" voice signal at the beginning of the talk can be transmitted completely and reliably.

According to the sixth aspect of the present invention, (a) an audio encoding section for generating and outputting audio data obtained by encoding an audio signal, and (b) the audio data output by the audio encoding section as the first audio data. A storage output unit that stores the audio data at a time point and outputs the audio data at a second time point after the first time point, and (c) obtains a voice cell obtained by converting the audio data output by the storage output unit into cells. A cell processing unit for outputting
(D) an empty cell generator for generating an empty cell transmitted only for transmission path management, and (e) audio output from the cell processor when the audio signal satisfies a predetermined criterion. The voice coded transmission device is provided with a selection unit that selects a cell as a transmission cell and selects the empty cell generated by the empty cell generation unit as a transmission cell when the above-mentioned criterion is not satisfied by the above-mentioned audio signal.

That is, in the invention according to claim 6, after the audio data obtained by encoding the audio signal by the audio encoding unit is stored at the first time, the second data after the first time is stored.
At the point of time, a corresponding voice cell is obtained. In addition, an empty cell generation unit generates an empty cell to be transmitted only for managing the transmission path. When the voice signal satisfies the predetermined criterion, the voice cell determined at the second time is selected as a transmission cell, while when the voice signal does not satisfy the criterion, an empty cell is selected as a transmission cell. You. Therefore, when the audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, after the time corresponding to the time corresponding to the difference between the first and second time points, Is selected as a transmission cell, and an empty cell before that is selected as a transmission cell. Therefore,
For example, when detecting a change from “silence” to “voiced”,
Even if the time at which the above change is actually detected is not exactly accurate, a voice cell corresponding to the content of the voice signal after a time immediately before that is selected as a transmission cell, so that a "voiced" voice signal at the beginning of the speech Can be completely and reliably transmitted.

According to the seventh aspect of the present invention, (a) an audio encoding section for generating and outputting audio data obtained by encoding an audio signal, and (b) audio data immediately after the audio encoding section outputs the audio data. A storage output unit that stores the audio data previously stored in the storage location and reads out and outputs the previously stored audio data from the second storage location; and (c) an audio cell that is a cell of the audio data output by the storage output unit. And (d) when the audio signal satisfies a predetermined criterion, selects a voice cell output from the cell processing unit as a transmission cell. A transmission device is provided.

That is, according to the seventh aspect of the present invention, the audio data immediately after being obtained by the encoding of the audio signal by the audio encoding unit is stored in the first storage position in the storage output unit and previously stored. The read audio data is read from the second storage position in the accumulation output unit and is cellized, and a corresponding audio cell is obtained. Then, when the audio signal satisfies the predetermined criterion, the audio cell obtained from the audio data stored in the storage output unit is selected as the transmission cell. Therefore, when the audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, the difference between the times at which the audio data is stored in the first and second storage positions in the storage output unit is determined. An audio cell corresponding to the content of the audio signal after the point in time corresponding to the corresponding time is selected as a transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as the transmission cell, the "voiced" voice signal at the beginning of the talk can be transmitted completely and reliably.

According to the invention of claim 8, (a) an audio encoding section for generating and outputting audio data obtained by encoding an audio signal, and (b) audio data immediately after the audio encoding section outputs the audio data. A storage output unit that stores the audio data previously stored in the storage location and reads out and outputs the previously stored audio data from the second storage location; and (c) an audio cell that is a cell of the audio data output by the storage output unit. (D) a vacant cell generator for generating a vacant cell transmitted only for transmission path management; and (e) when the audio signal satisfies a predetermined criterion. The voice cell output from the cell processing unit is selected as a transmission cell, and the vacant cell generated by the vacant cell generation unit is transmitted when the voice signal does not satisfy the criterion. And it is provided to the speech coding transmission equipment selection unit configured to select it.

That is, in the eighth aspect of the present invention, the audio data immediately after being obtained by the encoding of the audio signal by the audio encoding unit is stored in the first storage position in the storage output unit and previously stored. The read audio data is read from the second storage position in the accumulation output unit and is cellized, and a corresponding audio cell is obtained. In addition, an empty cell generation unit generates an empty cell to be transmitted only for managing the transmission path. Then, when the audio signal satisfies the predetermined criterion, the audio cell obtained from the audio data stored in the storage output unit is selected as a transmission cell. Selected as a transmission cell. Therefore, when the audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, the difference between the times at which the audio data is stored in the first and second storage positions in the storage output unit is determined. After the point in time corresponding to the corresponding time, a voice cell corresponding to the content of the voice signal is selected as a transmission cell, and before that, an empty cell is selected as a transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as the transmission cell,
The "voiced" voice signal at the beginning of the speech can be transmitted completely and reliably.

[0028]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below with reference to embodiments.

FIG. 1 shows the configuration of a speech coded transmission apparatus according to an embodiment of the present invention. In FIG. 1, an input device 11 inputs a voice signal, and is connected to a device such as a private branch exchange (PBX). Voice encoding unit 1
2 generates and outputs audio data obtained by encoding an audio signal provided from the input device 11. Storage output unit 13
Is output after accumulating the audio data output by the audio encoding device 12 for a certain period of time. The cell processing section 14 obtains and outputs a voice cell obtained by converting the voice data output from the accumulation output section 13 into cells. The empty cell generation unit 15 generates an empty cell transmitted only for managing the transmission path. The audio level detection section 16 detects an audio level indicating the strength of the audio signal provided from the input device 11. Sound existence judgment unit 17
Indicates whether or not the audio level detected by the audio level detection unit 16 satisfies a predetermined criterion, that is, “sound” when the audio signal reaches the reference audio level.
This is to determine whether the audio signal is an audio signal of “no sound” or a “silent” audio signal that does not reach the reference audio level. The selection unit 18 selects one of a voice cell and an empty cell as a transmission cell according to the determination result of the sound determination unit 17. The transmission processing unit 19 performs a process of transmitting the transmission cell selected by the selection unit 18 to the network, and includes, for example, an ATM switch.

FIG. 2 shows the configuration of the accumulation output section shown in FIG. In the figure, a storage unit 21 stores given audio data. The writing unit 22 causes the storage unit 21 to store the audio data immediately after output by the audio encoding device 12. The first storage location 22a is a location in the storage unit 21 where the audio data is written by the writing unit 22. Reading means 23
Reads out and outputs the audio data previously stored by the storage unit 21. The second storage location 23 a
3 is a position in the storage unit 21 from which audio data is read.

FIG. 3 specifically shows processing of an audio signal according to an embodiment of the present invention.
FIG. 3A shows a change in the sound level detected by the sound level detection unit 16, and FIG. 3B in the lower part shows a change in the sound level of the sound signal restored from the transmission cell transmitted to the network by the transmission processing unit 19. , Respectively.
In FIG. 3A, the horizontal axis indicates the time at which the audio level is detected by the audio level detection unit 16, and the vertical axis indicates the magnitude of the audio level. FIG.
In (b), the horizontal axis corresponds to the time at which the transmission cell was transmitted by the transmission processing unit 19, and the vertical axis corresponds to the audio level of the audio signal that can be restored from the transmission cell. 3 (a) and 3 (b) are drawn such that the times shown on the horizontal axis coincide with each other.

In FIG. 3A, L ₁ is the input device 11
L ₂ , which is the minimum sound level necessary for distinguishing a speech signal intentionally input from the CDMA from unrelated noise.
Indicates that the voice signal input from the input device 11
7 is the sound level when it is determined that the sound is present. Sa ₁ and Sa ₂ drawn by solid lines indicate changes in the audio signal intentionally input from the input device 11. Also, t ₁ is the time when the audio signal Sa ₁ reaches the audio level L ₁ , t ₂ is the time when the audio signal Sa ₁ reaches the audio level L ₂ , and t ₃ is the time when the audio signal Sa ₁ interrupts the audio level L _1. , t ₄ is the time the audio signal Sa ₂ reaches a sound level L _1, t ₅ the time the audio signal Sa ₂ reaches a sound level L _2, t ₆ is at the time the audio signal Sa ₂ is interrupted the audio level L ₁ The difference between the times t ₁ and t _{2 and} the difference between the times t ₄ and t ₅ are set to be the same time Δt. On the other hand, in FIG. ₃ (b), L 3 is the minimum sound level required to discern the sound signal to be restored from a transmission cell from which the extraneous noise, Sb ₁ and Sb ₂ drawn by a solid line The above-described audio signal Sa ₁
And shows changes of the audio signal to be restored from the voice cell respectively Sa ₂ is delivered from the delivery unit 19 correspondingly. Time Further, in FIG. 3 (a) and 3 (b), the time that the audio signal Sb ₁ reaches a sound level L ₃ is the time the time t ₂ described above, the audio signal Sb ₂ reaches a sound level L ₃ is as described above have been in such a way that t _5.

Next, the operation of the speech coded transmission apparatus according to this embodiment will be described.

The input device 11 supplies the input audio signal to the audio encoding unit 12 and the audio level detection unit 16, respectively. The audio encoding unit 12 generates and outputs audio data obtained by digitally encoding an audio signal supplied from the input device 11. The accumulation output unit 13 includes the audio encoding device 1
After accumulating the audio data output by the audio data 2 for a fixed time, that is, the time Δt shown in FIG. 3, the audio data is output in order from the audio data that has been stored first. That is, as shown in FIG.
The audio data immediately after output from the audio encoding device 12 is written into the first storage location 2 in the storage unit 21 by the writing unit 22.
2a and the first storage location 22a
Is incremented. After the elapse of the time Δt, the value of the second storage location 23a is incremented and coincides with the value of the first storage location 22a, and the voice data previously stored by the reading means 23 is stored in the second storage location 23a. It is read from the position 23a and output.

The cell processing section 14 is provided with the storage output section 13
The ATM cell is converted into an ATM cell to output the voice data. The empty cell generation unit 15 generates an empty cell that is transmitted only for managing the transmission path and represents “silence”. The audio level detection section 16 detects an audio level indicating the strength of the audio signal provided from the input device 11. The sound determination section 17 determines whether or not the audio level detected by the audio level detection section 16 reaches a predetermined reference audio level, and outputs a determination signal. The selecting unit 18 selects a voice cell as a transmission cell according to the determination signal output from the sound determination unit 17 if the detected voice level is a determination signal indicating that the detected voice level reaches a reference voice level. If the determined voice level does not reach the reference voice level, a vacant cell is selected as a transmission cell. Transmission processing unit 19
Sends out the transmission cell selected by the selection unit 18, that is, one of a voice cell and an empty cell to the network.

According to the above-described speech coded transmission apparatus in the present embodiment, when detecting a change in the sound signal from "silence" to "speech", the time when the change is actually detected is strictly accurate. At least, a voice cell corresponding to the content of the voice signal after a point immediately before that is selected as a transmission cell. For example, as shown in FIG.
The audio level L ₂ when the audio signal input from 1 is determined to be “voiced” by the voiced sound determination unit 17 is the minimum audio level required to distinguish the voice signal from noise unrelated to this audio signal. for greater than L ₁ is usually,
Continued selecting a transmission cell according to the determination signal of "voiced" and "silence", FIGS. 3 (a) audio data in the area P ₁ and the region P ₂ shown by hatching in is ignored, empty cell transmission It is selected as a cell. On the other hand, in the present embodiment, as shown in FIG. 3B, since the audio data is delayed by the time Δt with respect to the determination signal, “silence”
From the time that is the time change was retroactive only time Δt than time t ₂ or time t ₅ has been detected in the reality of the "voiced" t ₁
Alternatively the time t ₄ subsequent voice cell is selected as the transmission cell, the time t ₁ or time t ₄ the empty cell for the previous is selected as the transmission cell. Therefore, if the value of the time Δt is appropriately determined, the “voiced” voice signal at the beginning of the speech can be transmitted completely and reliably, and the reduction of traffic by silence compression and the transmission of natural speech without the beginning of speech can be achieved. Can be compatible.

In the voice coded transmission apparatus according to the present embodiment, the storage output unit 13 for storing voice data after storing it for a certain period of time is used as means for delaying voice data by the time Δt. Other means may be used as long as the signal can be delayed. Δt
Does not need to be a fixed value, and the value of Δt may be made variable according to, for example, the degree of increase in the audio level.

[0038]

As described above, according to the first aspect of the present invention, when the first audio signal is changing from a state not satisfying the criterion to a state satisfying the criterion, this change is actually detected. The voice cell corresponding to the content of the first voice signal after the time point corresponding to the delay corresponding to the delay time is selected as the transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as the transmission cell, the "voiced" voice signal at the beginning of the speech can be transmitted completely and reliably,
It is possible to achieve both the reduction of traffic by silence compression and the transmission of natural voice without truncation.

According to the second aspect of the present invention, the first
Is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, the content of the first audio signal is not changed after the time corresponding to the above-mentioned delay from the time when this change is actually detected. Is selected as a transmission cell, and an empty cell before that is selected as a transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as a transmission cell, the "voiced" voice signal at the head of the speech can be transmitted completely and reliably, and both the reduction of traffic due to silence compression and the transmission of natural voice without truncation of the head can be achieved. be able to.

Further, according to the third aspect of the present invention, when the audio signal is changing from a state not satisfying the criterion to a state satisfying the criterion, it corresponds to the difference between the first and second time points. The voice cell corresponding to the content of the voice signal after the point in time when the time goes back is selected as the transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as a transmission cell, the "voiced" voice signal at the head of the speech can be transmitted completely and reliably, and both the reduction of traffic due to silence compression and the transmission of natural voice without truncation of the head can be achieved. be able to.

According to the fourth aspect of the invention, when the audio signal is changing from a state not satisfying the criterion to a state satisfying the criterion, it corresponds to the difference between the first and second time points. After the point in time when the time goes back, a voice cell corresponding to the content of the voice signal is selected as a transmission cell,
Before this, empty cells are selected as transmission cells. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as a transmission cell, the "voiced" voice signal at the head of the speech can be transmitted completely and reliably, and both the reduction of traffic due to silence compression and the transmission of natural voice without truncation of the head can be achieved. be able to.

Further, according to the invention described in claim 5, when the audio signal is changing from a state not satisfying the criterion to a state satisfying the criterion, it corresponds to the difference between the first and second time points. The voice cell corresponding to the content of the voice signal after the point in time when the time goes back is selected as the transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as a transmission cell, the "voiced" voice signal at the head of the speech can be transmitted completely and reliably, and both the reduction of traffic due to silence compression and the transmission of natural voice without truncation of the head can be achieved. be able to.

According to the sixth aspect of the present invention, when the audio signal is changing from a state where the standard is not satisfied to a state where the standard is satisfied, this corresponds to the difference between the first and second time points. After the point in time when the time goes back, a voice cell corresponding to the content of the voice signal is selected as a transmission cell,
Before this, empty cells are selected as transmission cells. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as a transmission cell, the "voiced" voice signal at the head of the speech can be transmitted completely and reliably, and both the reduction of traffic due to silence compression and the transmission of natural voice without truncation of the head can be achieved. be able to.

According to the seventh aspect of the present invention, when the audio signal is changing from a state not satisfying the criterion to a state satisfying the criterion, the first signal in the accumulation output unit is output.
The audio cell corresponding to the content of the audio signal after the point in time when the time corresponding to the difference between the times at which the audio data is stored in the second storage location and the time when the audio data is stored is selected as the transmission cell. Therefore, for example, when detecting a change from “silence” to “speech”, even if the time when the above change is actually detected is not exactly accurate, the sound corresponding to the content of the sound signal after a time immediately before that time Since the cell is selected as a transmission cell, the "voiced" voice signal at the head of the speech can be transmitted completely and reliably, and both the reduction of traffic due to silence compression and the transmission of natural voice without truncation of the head can be achieved. be able to.

According to the eighth aspect of the present invention, when the audio signal is changing from a state that does not satisfy the criterion to a state that satisfies the criterion, the first and second storage positions in the accumulation output unit are set. After the time corresponding to the time corresponding to the difference between the times at which the voice data is stored, the voice cell corresponding to the content of the voice signal is selected as the transmission cell, and before that, the empty cell is selected as the transmission cell. You. Therefore, for example, from "silence" to "voiced"
In the case of detecting a change to the above, even if the time when the above change is actually detected is not exactly accurate, a voice cell corresponding to the content of the voice signal after a time immediately before that is selected as a transmission cell. In this case, the "voiced" voice signal can be completely and reliably transmitted, and both the reduction of traffic due to silence compression and the transmission of natural voice without truncation can be achieved.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a speech coded transmission device according to an embodiment of the present invention.

FIG. 2 is a conceptual diagram showing a configuration of an accumulation output unit shown in FIG.

FIG. 3 is a timing chart specifically illustrating processing of an audio signal according to an embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 12 audio encoding unit 13 accumulation output unit 14 cell processing unit 15 empty cell generation unit 16 audio level detection unit 17 voiced judgment unit 18 selection unit

Claims (8)

[Claims] 1. A delay processing unit that outputs a second audio signal obtained by delaying a first audio signal, an audio encoding unit that generates audio data obtained by encoding the second audio signal, A cell processing unit that outputs a voice cell obtained by converting the voice data generated by the encoding unit into cells; and a voice cell output from the cell processing unit when the first voice signal satisfies a predetermined criterion. And a selecting unit for selecting a transmission cell. 2. A delay processing section for outputting a second audio signal obtained by delaying a first audio signal; an audio encoding section for generating audio data obtained by encoding the second audio signal; A cell processing unit that outputs voice cells obtained by converting the voice data generated by the encoding unit into cells; an empty cell generation unit that generates empty cells transmitted only for transmission path management; and When the first audio signal is satisfied, an audio cell output from the cell processing unit is selected as a transmission cell, and when the first audio signal does not satisfy the criterion, an empty cell generated by the empty cell generation unit. And a selection unit for selecting a transmission cell as a transmission cell. 3. An accumulation output unit for accumulating audio data corresponding to an audio signal at a first point in time and outputting the audio data at a second point in time after the first point in time. A cell processing unit for obtaining and outputting a voice cell obtained by converting the output voice data into cells; and when the voice signal satisfies a predetermined criterion, selecting a voice cell output from the cell processing unit as a transmission cell. A speech coded transmission device comprising: a selection unit. 4. An accumulation output unit for accumulating audio data corresponding to an audio signal at a first point in time and outputting the audio data at a second point in time after the first point in time. A cell processing unit for obtaining and outputting a voice cell obtained by converting the output voice data into a cell; a vacant cell generating unit for generating a vacant cell transmitted only for transmission path management; and A selection unit that selects a voice cell output from the cell processing unit as a transmission cell when satisfies, and selects a vacant cell generated by the vacant cell generation unit as a transmission cell when the criterion is not satisfied by the voice signal. A voice coded transmission device comprising: 5. An audio encoding section for generating and outputting audio data obtained by encoding an audio signal, storing the audio data output from the audio encoding section at a first point in time, and A storage output unit for outputting at a second time point after the time point 1, a cell processing unit for obtaining and outputting voice cells obtained by converting the audio data output by the storage output unit into cells, And a selecting unit for selecting, as a transmission cell, a voice cell output from the cell processing unit when the signal is satisfied. 6. An audio encoding unit for generating and outputting audio data obtained by encoding an audio signal, storing the audio data output by the audio encoding unit at a first point in time, A storage output unit for outputting at a second time point after the time point 1, a cell processing unit for obtaining and outputting a voice cell obtained by converting the voice data output by the storage output unit into cells, A vacant cell generation unit for generating a vacant cell to be transmitted, and when the voice signal satisfies a predetermined criterion, a voice cell output from the cell processing unit is selected as a transmission cell, and the criterion is such that the voice signal is A selecting unit for selecting, as a transmission cell, a vacant cell generated by the vacant cell generating unit when the condition is not satisfied. 7. An audio encoding unit for generating and outputting audio data obtained by encoding an audio signal, and audio data immediately after output by the audio encoding unit are stored in a first storage location and previously stored. A storage output unit for reading and outputting the obtained audio data from the second storage location, a cell processing unit for obtaining and outputting a voice cell obtained by converting the audio data output by the storage output unit into cells, a predetermined reference And a selector for selecting, as a transmission cell, a speech cell output from the cell processing unit when the audio signal satisfies the following. 8. An audio encoding unit for generating and outputting audio data obtained by encoding an audio signal, and audio data immediately after output by the audio encoding unit are stored in a first storage location and previously stored. A storage and output unit for reading and outputting the obtained voice data from the second storage location, a cell processing unit for obtaining and outputting voice cells obtained by converting the voice data output from the storage and output unit into cells, A vacant cell generation unit that generates a vacant cell transmitted only for use, when the voice signal satisfies a predetermined criterion, selects a voice cell output from the cell processing unit as a transmission cell, and sets the criterion to the criterion. And a selecting unit for selecting an empty cell generated by the empty cell generating unit as a transmission cell when the audio signal is not satisfied.