JP2000151694A - Method for controlling correction for voice fluctuation, voice reproduction device and voice repeater - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize voice reproduction with excellent quality by controlling setting of a reproduction wait time optimum dynamically in response to a change in an arrival time even when the arrival time of voice data is fluctuated. SOLUTION: Upon the receipt of a notice, a fluctuation correction control section 105 acquires transmission time information from received data and based on the information, calculates an optimum reproduction wait time by the voice fluctuation correction control method that is a feature of this invention. The control section 105, based on the calculated wait time, writes digital voice data to an address of the reproduction buffer 101 that is corrected from a reception section 100 to apply fluctuation correction control by waiting an optimum reproduction wait time.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice communication system, and more particularly to a voice fluctuation correction control method for receiving and reproducing voice data in a system for transmitting and receiving voice data using a packet communication network such as the Internet. ,
Further, the present invention relates to an audio reproducing device and an audio relay device according to the audio fluctuation correction control method.

[0002]

2. Description of the Related Art In recent years, voice telephones have been switched to local area networks (LANs).
At present, the construction and services of a communication system using the Internet or an intranet for telephone calls between telephones are being gradually implemented. By the way, in conventional private communication, voice telephone is exclusively provided by PBX (Private Branch Exchange: private branch exchange), while data communication is performed by LAN.
The two communication networks are independent of each other. Therefore, if voice is converted to digital data adapted to communication data according to the LAN communication protocol so that voice can be transmitted and received as LAN data, voice communication using the LAN becomes possible. The integration of the voice telephone network and the data communication network will be realized.

However, a communication protocol used in a LAN is an Internet protocol (IP: Internet Protocol).
t Protocol) is standard. On the other hand, in the Internet, TCP (Transmission Control Protocol) is used as a transport layer protocol of the OSI reference model.
In fact, UDP (User Datagram Protocol) is used. TCP performs connection-type communication connection and has functions such as packet order control, flow control, error detection / recovery, retransmission, and congestion control. A certain communication service is provided.
On the other hand, UDP is a connectionless protocol, and is used for communication that requires real-time performance because it does not have a function like TCP. For example, in order to realize a voice call, it is of low importance to retransmit even if a part of voice data is lost, and voice data generated continuously must be sequentially transmitted. That is, UDP such as voice is used for communication that requires real-time properties.
The problem is that while Internet communications are good for sending and receiving discrete traffic, like voice,
Continuously generated data consists of UDP data packets.
Therefore, after high-speed transmission is performed sequentially and continuously, transmission is not always performed via the same transmission path. In other words, the time from when each data packet is successively transmitted at high speed by UDP until it is received at the receiving side via various transmission paths is not constant, and the time until it is received varies. Therefore, the sound is not always received in the order of transmission, and unless some control is performed, even if the sound is sequentially reproduced and output in the order of reception, the sound cannot always be reproduced as a continuous sound.

In order to prevent such discontinuous reproduction of audio, a reception buffer has been provided on the reception / reproduction side, and each audio data packet sequentially received is stored in the reproduction buffer. Is temporarily stored in a state of being rearranged in the order of transmission, and is reproduced and output for the first time after waiting for a time when continuous reproduction is possible. FIG. 5 shows an audio reproducing apparatus according to the prior art, which will be briefly described as follows.

That is, each voice data packet from the packet communication network is sequentially received by the receiving unit 500, and digital voice data (encoded and compressed state) in the received voice data packet is transmitted to the reproduction buffer 501 by packet transmission. The information is temporarily stored in a state of being rearranged in order. On the other hand, in parallel with such temporary accumulation,
Digital audio data that can be continuously reproduced is stored in a reproduction buffer 5
After a certain period of time temporarily stored in the storage buffer 01, digital audio data is sequentially read from the reproduction buffer 500 by the reproduction data reading unit 502 at a constant period. The sequentially read digital audio data is then sequentially converted into an analog audio signal by a D / A conversion unit 503 in a decoded and expanded state, and then reproduced as audio from an audio output unit (such as a speaker or a headphone) 504. This is what has been output. As described above, on a packet communication network (asynchronous communication network) such as the Internet, a LAN, an ATM network, a frame relay network, etc., each voice data packet is not always received by the receiving side in the transmission order, and the received voice data packet is not always received. Each of the data packets is temporarily stored in a state where the data packets are continuously rearranged in the transmission order, and a certain waiting time is required until the data packets are read and reproduced.

[0006]

As described above, up to now, a reproduction buffer having a fixed capacity is prepared on the receiving and reproducing side, and each of the audio data packets sequentially received is transmitted to the reproducing buffer. While the data is temporarily stored in a sequentially sorted state, the data is sequentially read out and reproduced from the reproduction buffer after a predetermined reproduction waiting time. However, such a method may cause a problem depending on the case. I have. This is because, even in a good communication state in which the packet communication network does not require a constant reproduction wait time, it is necessary to wait for a certain time before reading and reproducing, and conversely, it exceeds the constant reproduction wait time. This is because if such a reception delay occurs, a blank time will occur during reproduction.

[0007] An object of the present invention is to provide a transmission voice data packet which is received in a state where the reception order is different from the transmission order, that is, when received as a scattered state. Therefore, the present invention provides a sound fluctuation correction control method in which an optimum reproduction waiting time is dynamically set at any time, and each received sound data packet can be reproduced and output as good sound. Also, when relaying to a telephone network by a voice relay device, high quality voice transmission is realized by dynamically reproducing and transmitting voice with an optimum reproduction waiting time according to the state of the communication network.

[0008]

In order to solve the above-mentioned problems, the present invention firstly receives a voice packet containing encoded digital voice data via a communication network and extracts the digital voice data. In the case where digital audio data is temporarily stored in the reproduction buffer and reproduced and output, the difference between the current reproduction position and the write position on the reproduction buffer calculated from the transmission time information stamped at the time of transmission included in the received audio packet. That is, the reproduction waiting interval is monitored, and the audio data writing position is corrected based on the reproduction waiting interval.

Second, in the first solution procedure, the audio data writing is performed by comparing the reproduction waiting interval calculated at the time of receiving the audio packet with a preset minimum or maximum allowable reproduction waiting interval. The position is corrected.

Third, in the second solution procedure, the audio packet reproduction is performed by comparing the reproduction wait interval calculated at the time of receiving the audio packet with a preset minimum or maximum allowable reproduction wait interval. When the waiting interval is out of the range between the minimum value and the maximum value, the voice data writing position is corrected.

Fourth, in the third solution procedure, a comparison is made between the reproduction waiting interval calculated at the time of receiving the voice packet and the preset minimum and maximum allowable reproduction waiting intervals.
The audio data writing position is corrected based on the frequency of occurrence of an event in which the audio packet reproduction waiting interval is out of the range between the minimum value and the maximum value.

Fifth, in the second solution procedure, the reproduction waiting interval calculated at the time of receiving the audio packet is compared with a preset minimum allowable reproduction waiting interval, and the minimum value or the minimum value is determined. The audio data writing position is corrected in a direction to shorten the reproduction waiting interval based on the frequency of occurrence of a small event.

Sixthly, an audio reproducing apparatus according to the present invention comprises: a data receiving section for receiving an audio data packet from a packet communication network; a fluctuation correction control section using the above-described control method for controlling audio fluctuation; A reproduction buffer for storing digital audio data in the data packet, a reproduction data read unit for reading the digital audio data to be reproduced from the reproduction buffer, and converting the digital audio data output by the reproduction data read unit into audio data for output The configuration includes a D / A conversion unit and an audio output unit that reproduces and outputs the audio data output from the D / A conversion unit.

Seventh, the packet communication network of the audio reproducing apparatus according to the sixth aspect may be the Internet, a local area network, an ATM network or a frame relay.

Eighth, an audio relay apparatus according to the present invention is provided with a line corresponding section for outputting audio to a switched line network in place of the audio output section of the above-described audio reproducing apparatus. .

Ninth, in the voice relay apparatus according to the eighth aspect, the switched network is a public analog telephone network, a public ISDN.
Alternatively, it may be an extension telephone network.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention is applied to a voice reproducing apparatus for receiving voice data and reproducing voice via an asynchronous communication network such as a packet communication network and a voice relay apparatus for transmitting reproduced voice to a switching network. In particular, the present invention is applied to an Internet telephone that communicates voice data using the Internet.

Example 1 Hereinafter, a first embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram of the audio reproducing apparatus according to the present embodiment.
A receiving unit 100 receives a data packet from a packet communication network, and notifies 105 that data has been received. Reference numeral 105 denotes a fluctuation correction control unit which is a feature of the present invention.
01 is controlled. A playback buffer 101 temporarily stores digital audio data output from the playback buffer 105. Reference numeral 102 denotes a reproduction data readout unit, which reads digital audio data after the digital audio data that can be continuously reproduced is stored in 101, and reads out the digital audio data.
Output to 3. Reference numeral 103 denotes a D / A conversion unit that converts digital audio data into audio data and outputs the data to 104.
Reference numeral 104 denotes an audio output unit such as a speaker or headphones.

The playback buffer 101, playback data readout unit 102, and fluctuation correction control unit 105 of the audio playback apparatus configured as described above will be described below in detail with reference to FIGS.

FIG. 2 is a diagram for explaining a buffer control process of the fluctuation correction control unit 105. The arrangement of the blocks in the figure represents the reproduction buffer 101 in FIG. 1. The fluctuation correction control unit 105 in FIG. 1 writes the audio data at the optimum position, and the reproduction data reading unit 102 starts from the left block in FIG. It means reading and reproducing.
B0 is a unit block for sound reproduction. This value is the size of audio data that has been converted into digital audio data at every predetermined sampling time, for example, 8 bits 8 kHz at a sampling time of 30 ms.
When the μ-law of z is used, digital audio data of 240 bytes is obtained, which is the size of an audio reproduction unit block. B1 indicates the block number of the reproduction block currently being reproduced, and B2 indicates the block number of the reproduction block in which the received audio data is written. d
Is a value indicating the interval between the write block number B2 and the reproduction block number B1, that is, the number of audio data blocks that wait until the received audio data is written and then reproduced. The hatched portions in FIG. 2 indicate the audio data waiting from reception of the audio data to reproduction.

The following relationship exists between the number of blocks waiting to be reproduced and audio reproduction. That is, as the number of waiting blocks increases, the time from when the audio data is received to when the audio data is reproduced becomes longer. However, even if the arrival time greatly fluctuates depending on the communication state, the audio can be reproduced continuously without interruption. . On the other hand, the smaller the number of blocks waiting to be played back, the shorter the time from when the audio data is received to when it is played back, and it can be played back with a small delay time, but with a delay longer than the time waiting for playback. When audio data arrives, the audio is interrupted and played. The fluctuation correction control unit 105 performs the calculation of the waiting interval from the reception of the audio data to the reproduction, that is, the process of determining the write block number B2 for the reproduction block number B1. Note that Tmin represents the minimum reproduction waiting time, and this value represents the minimum value of the interval d between the write block number B2 and the reproduction block number B1. Tmax
Represents the maximum reproduction waiting time, and represents the maximum value of the interval d. The values of Tmin and Tmax are values that are referred to for the fluctuation correction control processing, and are values that are set before receiving audio data. For example, if the reproduction unit block B0 is 30
If digital audio data sampled at m-second intervals is used, a minimum playback wait time of 60 ms and a maximum playback delay time of 300 ms are set to Tmin.
= 2, Tmax = 10.

FIG. 3 is a flowchart for explaining the operation of the fluctuation correction control unit 105. In FIG. 3, each time a voice arrives, the flow is executed, an interval d between the current playback block number and the write block number in the playback buffer is obtained, and this interval d is compared with the maximum value Tmax and the minimum value Tmin. The operation of correcting the position to be written to the reproduction buffer according to the frequency of occurrence in the specific time range is described.

The working variables used for the processing in FIG. 3 will be described below. The correction value AJ is a value for correcting the block number in which the audio data is written, and takes an integer value. When the state where the arrival of the audio data is continuously delayed continues, this value increases and is corrected so that the audio data is written to a reproduction buffer temporally separated from the current reproduction block. On the other hand, if audio data continues to arrive early,
This value decreases, and is corrected so as to write to a reproduction buffer that is temporally closer to the current reproduction block.

The transmission time information T is information indicating a time depending on the data amount of voice encoded by the voice transmitting device, and is transmitted together with the voice data. When transmitting and receiving real-time traffic such as audio and video on the Internet, RTP (Real Time Protocol) is used. RTP is a protocol mainly defined in the upper layer of UDP, and defines information such as a time stamp and a sequence number for the purpose of controlling real-time data. The time information at the time of voice transmission is transmitted using the time information part of this protocol. For example, when the 8-bit 8 kHz μ-law is used at a sampling time of 30 ms, 240 bytes of audio data are generated every 30 ms, and the transmission time information T at this time is added by 240 for each packet. The set value is set.

The minimum value comparison counter CMin is a value that counts the number of times that the interval d is continuously equal to or longer than the minimum reproduction waiting time Tmin.
When Min is continuously counted over a preset comparison value, the audio data is constantly reproduced with a reproduction waiting time longer than Tmin. Is changed so that is closer to the reproduction block position side. The comparison value set in advance for comparison with CMin is the corrected maximum value AJWmax.
When setting to change the correction value AJ when Min is counted, AJWmax = 10.

The correction value increase counter AJP is a value for counting the number of times that the interval d is continuously smaller than the minimum reproduction waiting time Tmin, and the counter value is larger than a preset comparison value. When the error occurs, the correction value AJ is changed so that the audio data is written in the direction temporally away from the reproduction block position. The comparison value set in advance for comparison with AJP is a correction value increase change counter AJPCh. For example, when the correction value AJ is changed when AJP is counted twice consecutively,
AJPCh = 2.

The correction value decrease counter AJM is a value for counting the number of consecutive occurrences when the difference d is greater than the maximum reproduction waiting time Tmax, and the counter value is continuously larger than a preset comparison value. When the error occurs, the correction value AJ is changed so that the audio data is written to the reproduction buffer temporally close to the reproduction block position. A comparison value set in advance for comparison with AJM is a correction value decrease change counter AJMCh. For example, when the correction value AJ is changed when AJM is counted twice consecutively, AJMCh = 2. Becomes

The audio reproducing apparatus executes the flow of FIG. 3 every time a new audio packet arrives, and determines the write position of the audio data to be written into the reproduction buffer based on the correction value AJ and the transmission time information T. And transmission time information T
The correction value A is calculated based on the difference between the write position on the reproduction buffer and the current reproduction position, that is, the reproduction wait interval.
Calculate J. The transmission time information T is, for example, a value obtained by engraving the time at which the audio was sampled with a value depending on the data amount obtained by encoding the audio data on the audio transmitting device side as a reference value. It is transmitted together with the audio data.

Hereinafter, the operation will be described in detail with reference to FIG. When new voice data is received (S1), transmission time information T is obtained from the received data (S2). Next, the current playback block number B1 is obtained (S3). Then, based on the transmission time information T and the correction value AJ, the write block number B
2 is calculated (S4). The correction value AJ has been updated by the immediately preceding received data processing (the initial value is 0). The transmission time information T is 8 at 30 ms interval sampling.
In the case of the μ-law of the bit 8 kHz, the value is an integral multiple of 240. By dividing T by 240 and adding the correction value AJ,
The corrected write block number B2 is calculated.

Next, the value of the current reproduction block number B1 is subtracted from the write block number B2 to calculate an interval d (S5). Next, the minimum value comparison counter CMin is incremented (increment: +1 updated) (S6). This CMin is a value used to correct the write block number so that the sound is reproduced earlier when the case where the interval d is larger than Tmin occurs continuously.
Next, it is determined whether the writing / reproducing position block interval d is equal to or less than the minimum reproduction waiting time Tmin (S7). If the determination is false, then the interval d is branched to compare with the maximum reproduction wait time Tmax, and if true, the interval d is set to Tmi.
Since the case where the value is larger than n is not continuous, the minimum value comparison counter CMin is initialized (S8).

Next, if the correction value increase counter AJP is equal to or larger than the correction value increase change counter AJPCh (S9), the correction value AJ is added (incremented) (S10), and the correction value increase counter AJP is initialized (S11). . afterwards,
The correction value increase counter AJP is added (incremented) and the correction value decrease counter AJM is initialized (S12).
Thereafter, the flow branches to (2), and the received data is written to the write block number B2 (S22). When the interval d is equal to or smaller than Tmin, the processing of the received data is completed here.
And waits for the next received data to arrive.

When the determination in step (S7) is false, the writing / reproducing position block interval d is equal to the maximum reproduction waiting time Tma.
It is determined whether the value is equal to or more than x (S13). When the result is false, that is, when the interval d is between Tmin and Tmax, the process branches to write the received data to the write block number B2 (S22), and returns without changing the correction value AJ. If true, it is determined whether the correction value decrease counter AJM is equal to or greater than the correction value decrease change counter AJMCh (S14), and the correction value A
J is decremented (decrement: -1 updated) (S15), and a correction value decrease counter AJM is initialized (S16). Thereafter, the addition (increment) of the correction value decrease counter AJM and the correction value increase counter AJP are initialized (S1).
7).

When the determinations in steps (S7) and (S13) are both false, that is, when the interval d is between Tmin and Tmax, it is determined whether the minimum value comparison counter CMin is equal to or greater than the corrected maximum value AJWmax. (S18). If true, the correction value AJ is subtracted (decremented) (S19), the correction value decrease counter AJM is initialized (S20), and the minimum value comparison counter CMin is initialized (S21). The correction maximum value AJWmax is a value set in advance, and is determined in step (S1).
In the determination of 8), when the interval d is longer than the minimum reproduction waiting time Tmin, if the number of times of the maximum correction value AJWmax occurs continuously, the correction value AJ is subtracted (decremented).
That is, when the communication network is stable such that the arrival time of the received data is constantly received within a fixed time width, the write block number is corrected so as to reduce the reproduction waiting time. The value AJ is subtracted (decremented).

Thereafter, the received data is written to the write block number B2 (S22), and the flow returns to the standby state for the next received data.

By the processing control of the arrival time fluctuation correction control unit described above, even if the arrival time of voice data fluctuates depending on the state of the communication network, the voice reproduction waiting time is dynamically corrected and controlled, so that the communication network has a large reproduction time. The reproduction wait time is changed to the minimum time when the communication state is good such that no wait time is required, and the reproduction wait time is changed to a long reproduction wait time when the communication state delays the arrival of audio data.

Now, the operation of the audio reproducing apparatus including the fluctuation correction control unit will be described based on the operation of the fluctuation correction control unit described above.

When receiving the digital audio data addressed to the own audio reproducing apparatus from the packet communication network, the receiving section 100 notifies the fluctuation correction control section 105 of the reception. 10
5 receives the notification, obtains the transmission time information T from the received data, and calculates the optimum reproduction waiting time from the value by the fluctuation correction control method described above. Based on the calculated optimal reproduction waiting time, 105
1. Write digital audio data to the optimum position. By dynamically optimizing the writing position, adjustment is made so as to wait for a reproduction waiting time according to the communication state. 102 sequentially reads out digital audio data from the reproduction buffer 101,
Output to 103. The D / A conversion unit 103 sequentially performs D / A conversion on the digital audio data received from the D / A converter 102 and outputs the digital audio data to the audio output unit 104. 104 reproduces and outputs the voice.

According to the above-described first embodiment of the present invention, the asynchronous communication in which the data arrival time typified by the Internet fluctuates by the audio reproduction apparatus including the above-described audio data arrival time fluctuation correction control unit. When performing voice communication on a network, voice reproduction is realized with an optimum reproduction waiting time corresponding to the communication network state.

Embodiment 2 Hereinafter, a second embodiment of the present invention will be described with reference to the drawings. FIG. 4 is a block diagram of a voice relay apparatus according to the present embodiment. FIG.
A receiving unit 400 receives a data packet from a packet communication network, and notifies a receiving unit 405 that data has been received. Reference numeral 405 denotes a fluctuation correction control unit which is a feature of the present invention.
1 is controlled to be written. 401 is a playback buffer 4
05 is temporarily stored. Reference numeral 402 denotes a reproduction data readout unit which reads digital audio data after waiting for digital audio data that can be continuously reproduced to be stored in 401;
Output to A D / A converter 403 converts digital audio data into audio data and outputs the audio data to 404. 4
Reference numeral 04 denotes a line corresponding unit for transmitting the voice output from 403 to the telephone line network.

The operation of the voice relay apparatus configured as described above will be described. When receiving the digital voice data addressed to the voice relay apparatus from the packet communication network, the receiving section 400 transmits the digital voice data to the fluctuation correction control section 405. Notify that Upon receiving the notification, the transmission time information 405 obtains transmission time information T from the received data, and calculates an optimum reproduction waiting time from the value by the fluctuation correction control method described above. Based on the calculated optimal reproduction waiting time, 40
5 writes digital audio data to the optimum position of the reproduction buffer 401. By dynamically optimizing the writing position, adjustment is made so as to wait for a reproduction waiting time according to the communication state. 402 sequentially reads out digital audio data from the reproduction buffer 401 and outputs it to 403. The D / A converter 403 sequentially converts the digital audio data received from 402 into a digital signal and outputs the digital audio data to the audio output unit 404.
From 404, it is reproduced and output to the telephone line network.

According to the above operation, according to the second embodiment of the present invention, the voice relay apparatus including the voice data arrival time fluctuation correction control unit causes the data arrival time typified by the Internet to fluctuate. When relaying voice communication in a communication network, voice is reproduced with a reproduction waiting time corresponding to the state of the communication network and transmitted to the telephone line network, whereby an optimum voice relay is realized.

As can be seen from the above description, the writing position of the audio data is changed so that the reproduction waiting time is shortened, and when writing is performed, the result is overwritten on the unreproduced audio data. As a result, the unplayed audio data is invalidated. For example, in the case of sampling at intervals of 30 ms, one writing position correction requires 30
That is, m seconds of audio data are invalidated.
Therefore, if the writing position is continuously changed, as a result of the invalidation of the audio data, the sound is interrupted as perceived. However, according to the present invention, the write position correction process is performed for the first time when an event that reduces the playback wait time by the specified number of times occurs, and as a result, the audio data is in a time-discrete state. Because it has been disabled, it is of little concern for hearing.

On the other hand, contrary to the above, the writing position of the audio data is changed so as to increase the reproduction waiting time, and when writing is performed, the writing position and the audio data have already been written. As a result, there is a large address difference between the current position and the current position, so that when the sound is reproduced during that time, no sound is reproduced (corresponding to a break in sound reproduction), or already reproduced sound data is reproduced. (Corresponding to discontinuous reproduction in audio reproduction). Such inconvenience can be reduced to some extent by an interpolation method based on repeated reproduction of the immediately preceding audio data. Therefore, if the writing position is continuously changed, the continuity of the audio data is lost, and as a result, the unnaturalness is perceived by the interpolation reproduction of the same audio. However, according to the present invention, the write position correction process is performed for the first time when an event that increases the reproduction waiting time by the specified number of times occurs, and as a result, the reproduction waiting time is delayed as a time discrete state. As a result, there is little problem with hearing.

[0044]

As described above, according to the present invention, when voice communication is performed via an asynchronous communication network (such as the Internet) and received digital voice data is reproduced as voice by a voice reproducing apparatus. In addition, the optimal reproduction waiting time is dynamically set according to the state of the communication network, and the high-quality voice communication is realized by reproducing the voice. Also,
When digital voice data received for performing voice communication via an asynchronous communication network is transmitted to a telephone network by a voice relay device, voice playback is dynamically performed at an optimum playback waiting time depending on the state of the communication network. To realize high quality voice transmission.

[Brief description of the drawings]

FIG. 1 is a diagram showing a block configuration of an audio reproducing apparatus according to the present invention.

FIG. 2 is a diagram for explaining fluctuation correction control processing in the audio reproduction device;

FIG. 3 is a diagram showing a flow of the fluctuation correction control processing;

FIG. 4 is a diagram showing a block configuration of a voice relay device according to the present invention;

FIG. 5 is a diagram showing a block configuration of an audio reproducing apparatus according to the related art.

[Explanation of symbols]

100, 400 ... receiving unit, 101, 401 ... reproduction buffer, 102, 402 ... reproduction data reading unit, 103,
403: D / A converter, 104: audio output unit, 105,
405: fluctuation correction control unit, 404: line corresponding unit, B0
... reproduction unit block, B1 ... reproduction position, B2 ... write position

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification code FI Theme coat ゛ (Reference) H04Q 11/04 H04Q 11/04 RF term (Reference) 5K030 GA11 HA08 HB01 JT01 KA03 KA19 5K033 BA14 DB10 DB13 5K069 AA01 BA02 BA10 CA08 FC03 FC13 FC16 5K101 LL05 SS08 9A001 CZ04 CZ08 HH15 JZ25 KK31

Claims (14)

[Claims] When receiving a voice packet including encoded digital voice data via a communication network, extracting the digital voice data, temporarily storing the digital voice data in a reproduction buffer, and reproducing and outputting the data, The difference between the write position on the reproduction buffer calculated from the transmission time information stamped at the time of transmission included in the received audio packet and the current reproduction position, that is, the reproduction wait interval is monitored, and the audio data is A sound fluctuation correction control method, wherein a writing position is corrected. 2. The audio data writing position is corrected by comparing a reproduction waiting interval calculated at the time of receiving an audio packet with a preset minimum or maximum allowable reproduction waiting interval. Item 2. The sound fluctuation correction control method according to Item 1. 3. A range between the minimum value and the maximum value of the audio packet reproduction waiting interval, which is obtained by comparing the reproduction waiting interval calculated at the time of receiving the audio packet with a preset minimum or maximum allowable reproduction waiting interval. 3. The audio fluctuation correction control method according to claim 2, wherein the audio data writing position is corrected when the voice data is outside. 4. A range between the minimum value and the maximum value of the voice packet reproduction waiting interval, which is obtained by comparing the reproduction waiting interval calculated at the time of receiving the voice packet with the preset minimum value and maximum value of the allowable reproduction waiting interval. 4. The sound fluctuation correction control method according to claim 3, wherein the sound data writing position is corrected based on the frequency of occurrence of an outside event. 5. A comparison between a playback waiting interval calculated at the time of receiving a voice packet and a preset minimum value of an allowable playback waiting interval, based on a frequency of occurrence of a minimum value or an event smaller than the minimum value. 3. The audio fluctuation correction control method according to claim 2, wherein the audio data writing position is corrected in a direction to shorten the reproduction waiting interval. 6. A voice fluctuation correction control method according to claim 1, wherein the data receiving unit receives a voice data packet from a packet communication network, and controls the fluctuation of the voice data arrival time. , A reproduction buffer for storing digital audio data in a received data packet, a reproduction data readout unit for reading out digital audio data to be reproduced from the reproduction buffer, and a digital output from the reproduction data readout unit. D / A that converts audio data into audio data and outputs it
An audio reproducing apparatus comprising: a conversion unit; and an audio output unit that reproduces and outputs audio data output from the D / A conversion unit. 7. The audio reproducing apparatus according to claim 6, wherein the packet communication network is the Internet. 8. The audio reproducing apparatus according to claim 6, wherein the packet communication network is a local area network. 9. The audio reproducing apparatus according to claim 6, wherein the packet communication network is an ATM network. 10. The audio reproducing apparatus according to claim 6, wherein the packet communication network is a frame relay network. 11. A voice relay device comprising a line corresponding unit for outputting voice to a circuit switching network in place of the voice output unit according to claim 6. 12. The voice relay apparatus according to claim 11, wherein the circuit switching network is a public analog telephone network. 13. The voice relay apparatus according to claim 11, wherein the circuit switching network is a public ISDN. 14. The voice relay apparatus according to claim 11, wherein the circuit switching network is an extension telephone network.