JP4299731B2 - Reception device and reception reproduction method - Google Patents

Description

  The present invention relates to a digital data communication system used, for example, in a building broadcasting facility.

  In a digital data communication system on a network such as a LAN, communication is performed by packetizing not only text and still images but also data such as voice and moving images. One of the digital communication and reproduction methods is a “streaming reproduction method” in which reproduction is performed without waiting for reception until the end of data. The streaming reproduction method is a method for transmitting audio and moving images in real time, and a transmission buffer and a reception buffer are provided in the transmission device and the reception device, respectively. The processing of the transmitting device and the receiving device in this method is as follows.

  For example, in the case of voice communication, an analog voice signal input to a transmission device is converted into digital voice data by an analog / digital converter (ADC) and stored in a transmission buffer. The receiving device receives and disassembles the packet, and stores the digital audio data in the receiving buffer. When the amount of data stored in the reception buffer exceeds a certain level, the digital audio data is read from the reception buffer, converted into analog audio data by a digital / analog converter (DAC), and output.

In the communication of voice and moving images in this communication system, the following two things are required.
1. The receiving device receives continuous data packets at the same time interval as the packet transmission interval in the transmitting device.
2. The sampling frequency (hereinafter referred to as “transmission-side sampling frequency”) fs in the transmission device and the sampling frequency (hereinafter referred to as “reception-side sampling frequency”) fr in the reception device must match.

  However, depending on the network traffic conditions, there are problems such as fluctuations in packet transmission time, packet loss due to collision, and failure to arrive in order. Further, since the transmission device and the reception device use sampling frequencies based on different internal clocks, an error occurs between the transmission-side sampling frequency fs and the reception-side sampling frequency fr. Therefore, there is a difference between the amount of data included in packets transmitted per same time and the amount of data to be reproduced. For this reason, the amount of data stored in the reception buffer becomes zero during reproduction, which may cause an underflow that stops reproduction. When an underflow occurs, the sound is cut off and the sound quality is deteriorated. Conversely, an overflow may occur in which the received packet cannot be absorbed by the reception buffer. When the overflow occurs, the received packet is discarded, so that the voice data becomes discontinuous and causes noise.

In view of this, Patent Document 1 discloses a receiving apparatus having correction means for correcting the sampling frequency in the receiving apparatus in order to improve the quality of data such as voice. The receiving device described in Patent Document 1 sequentially receives the first and second packets including the packet transmission time together with the digital data from the transmitting device, and acquires the packet transmission time and the reception time when the packet is received. Here, the packet transmission time is the time at which the first and second packets are transmitted from the transmission device to the reception device. Then, the transmission interval is calculated as the time required to transmit the predetermined number of packets, and the reception interval is calculated as the time required to receive the predetermined number of packets. The reception-side sampling frequency fr is corrected so that the difference between the transmission interval and the reception interval becomes small. Therefore, underflow and overflow caused by sampling frequency errors in the transmission device and the reception device can be suppressed, and the quality of reproduced data can be improved.
JP 2002-374319 A

  However, although the receiving apparatus described in Patent Document 1 can suppress underflow and overflow, time information such as packet transmission time is transmitted from the transmitting apparatus to the receiving apparatus in order to adjust the receiving-side sampling frequency fr. Must. Therefore, the transmission device needs to incorporate time information such as packet transmission time into the packet to be transmitted. Therefore, the sampling frequency error cannot be corrected unless the transmission device and the reception device are integrated. Also, incorporation of time information increases the amount of packet data, leading to a decrease in transmission efficiency and an increase in bandwidth used.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a receiving apparatus that can improve the sound quality by correcting the receiving side sampling frequency only on the receiving apparatus side, not on the transmitting apparatus side.

In order to solve the above-described problems, the first invention of the present application is a receiving device connected to a transmitting device via a communication network, and receives a plurality of packets sequentially from the transmitting device via the communication network. Means, a reception buffer for temporarily storing digital data included in each of the plurality of packets, a frequency generation means for generating a reception-side sampling frequency fr, and a reception buffer according to the reception-side sampling frequency fr. For each packet group G _n (n is a natural number) composed of a conversion means for reading out digital data and converting it into analog data, and a predetermined number k (k ≧ 2) packets determined on the receiving side A reception interval ΔTr indicating a time required for the means to receive k packets, and a digital signal included in each of the k packets. And a correction unit that obtains a reading interval ΔTo indicating a time required to read the total data from the reception buffer, and calculates a reception-side sampling frequency fr in the packet group G _n based on the following conditions: I will provide a.

For example, 100 packets G _n between the receiving means time Tr0~Tr1 (n is a natural number) are receiving. The digital data included in each packet received by the receiving means is temporarily stored in the reception buffer. Then, 100 packets of digital data are read from the reception buffer between times To0 and To1. Here, the reception interval ΔTr = Tr 1 −Tr 0 is calculated from the reception times of the first and last packets of the 100 packets. Also, the digital data read interval ΔTo = Tr 1 −Tr 0 is calculated from the read times of the first and last packets out of 100. The correcting unit corrects the reception-side sampling frequency fr so that the calculated reception interval ΔTr and readout interval ΔTo satisfy the condition of fr = fr ′ × (ΔTo ÷ ΔTr) . fr ′ is a sampling frequency on the receiving side in the packet group G _(n−1) .
Here, the reception interval ΔTr is responsive to the transmission interval of both packets from the transmission device to the reception device. Therefore, by correcting the reception-side sampling frequency fr as described above, the transmission-side sampling frequency fs and the reception-side sampling frequency fr can be in a predetermined relationship. By performing the above processing, it is possible to control the accumulation amount of the reception buffer, prevent overflow and underflow, and improve the quality of reproduced data. Further, the correction of the reception side sampling frequency fr described above is performed based on the reception time and reading time acquired on the reception device side. Therefore, there is no need to transmit time information related to the transmission side sampling frequency fs such as the packet transmission time and the packet transmission interval from the transmission device side to the reception device side. Therefore, by simplifying the packet configuration of the packet to be transmitted and reducing the data amount of the packet, the packet transmission efficiency can be improved and the bandwidth to be used can be reduced. The predetermined relationship is a relationship in which the reception-side sampling frequency fr and the transmission-side sampling frequency fs are fr: fs = 1: 1, fr: fs = 1: 2, and the like. Reading can be supported.

According to a second aspect of the present invention, in the first aspect of the present invention, a VCO (Voltage Control Oscillator) is used as the frequency generation means, and the correction means corrects the reception-side sampling frequency fr by adjusting an applied voltage to the VCO. A receiving apparatus is provided.
The third invention of the present application is the monitoring means according to the first invention of the present application, which monitors the transmission order of the packets received by the receiving means and the accumulation amount in the reception buffer, and determines whether or not the packets are normally received. The conversion means reads out the stored digital data from the reception buffer according to the determination as to whether or not the monitoring means has received the packet normally, and the reception buffer is in the monitoring means Provided is a receiving device for storing digital data in accordance with determination as to whether or not a packet is normally received and the amount stored in the reception buffer.
A fourth invention of the present application is a reception reproduction method used for a reception device connected to a transmission device via a communication network, wherein a reception step of sequentially receiving a plurality of packets from the transmission device via the communication network; An accumulation step for temporarily accumulating digital data included in each of the plurality of packets in a reception buffer, a frequency generation step for generating a reception-side sampling frequency fr, and an accumulation in the reception buffer according to the reception-side sampling frequency fr For each packet group G _n (n is a natural number) composed of a predetermined number k (k ≧ 2) packets determined on the receiving side, A reception interval ΔTr indicating the time required to receive k packets in the step, and the k packets. A reading interval ΔTo indicating the time required to read the digital data contained in each from the reception buffer is obtained, and the reception-side sampling frequency fr in the packet group _Gn is calculated based on the following conditions: And a receiving / reproducing method including a correcting step for correcting.
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the reception-side sampling frequency in the packet group G (n−1).

A fifth invention of the present application is a reception reproduction program used for a reception device connected to a transmission device via a communication network, the reception means sequentially receiving a plurality of packets from the transmission device via the communication network, A reception buffer for temporarily storing digital data included in each of a plurality of packets, frequency generation means for generating a reception-side sampling frequency fr, and reading out the digital data stored in the reception buffer according to the reception-side sampling frequency fr; For each packet group G _n (n is a natural number) composed of conversion means for converting into analog data and a predetermined number k (k ≧ 2) packets determined on the receiving side, the receiving means includes k pieces A reception interval ΔTr indicating the time required to receive a packet, and a digital included in each of the k packets. The reading interval ΔTo indicating the time required to read the total data from the reception buffer is acquired, and the computer is caused to function as correction means for calculating the reception-side sampling frequency fr in the packet group G _n based on the following conditions: A receiving / reproducing program is provided.
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the receiving-side sampling frequency in the packet group G _(n−1) .

A sixth invention of the present application is a computer-readable recording medium in which a reception / playback program used in a reception device connected to a transmission device via a communication network is recorded, wherein a plurality of packets are transmitted via the communication network. A reception step for sequentially receiving from the transmission device; a storage step for temporarily storing digital data contained in each of the plurality of packets in a reception buffer; a frequency generation step for generating a reception-side sampling frequency fr; and the reception-side sampling A packet group G _n composed of a conversion step of reading digital data stored in the reception buffer according to the frequency fr and converting it into analog data, and a predetermined number k (k ≧ 2) packets determined on the reception side (n is a natural number) for each, to receive the k packets in the receiving step A reception interval ΔTr indicating the time that the k pieces of digital data packets are included in each acquired a read interval ΔTo indicating the time required to read from the receive buffer, the based on the following conditions recording the received reproducing program for executing a correction step of correcting for calculating the reception side sampling frequency fr in the packet group G _n, and a computer-readable recording medium.
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the receiving-side sampling frequency in the packet group G _(n−1) .

The seventh invention of the present application is a broadcasting system, which is connected to the voice input means and the voice input means, and converts the input voice into digital voice data by sampling based on the transmission side sampling frequency fs, Voice transmission means for sequentially outputting packets including digital voice data to the transmission path, and connected to the voice transmission means via the transmission path, and converting the digital voice data into analog voice data based on the sampling frequency fr on the receiving side And a voice output unit connected to the voice receiving unit and outputting the analog voice data. Here, the voice receiving means includes a receiving means for sequentially receiving a plurality of packets from the transmitting device, a receiving buffer for temporarily storing digital data included in each of the plurality of packets, and a receiving-side sampling frequency fr. A frequency generation means for generating, a conversion means for reading out digital data stored in the reception buffer according to the reception side sampling frequency fr and converting it into analog data, and a predetermined number k (k ≧ 2) determined on the reception side For each packet group G _n (n is a natural number) composed of packets , the reception means includes a reception interval ΔTr indicating the time required for receiving k packets, and each of the k packets. get a read interval ΔTo indicating the time required for digital data to read from the receive buffer being, Based on the following conditions and has a correction means for calculating a reception side sampling frequency fr in the packet group G _n.
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the receiving-side sampling frequency in the packet group G _(n−1) .

  By using the present invention, the reception side sampling frequency fr can be corrected without transmitting time information related to the transmission side sampling frequency fs such as packet transmission time and packet transmission interval from the transmission side to the reception side. it can.

<Outline of the invention>
In the present invention, the sampling on the receiving side is performed based on the reception interval when the packet transmitted from the transmission device is received by the reception unit of the reception device and the reading interval when the digital data included in the packet is read from the reception buffer. The frequency fr is corrected.
FIG. 1 is a conceptual explanatory diagram showing an outline of the present invention. Here, it is assumed that the reception interval and the reading interval are calculated with 100 packets as one packet group. The receiving unit of the receiving apparatus sequentially receives the packets P _{0 to} P ₁₀₀ during the times Tr _{0 to} Tr ₁ . The digital data included in the packets P _{0 to} P ₁₀₀ received by the receiving unit is temporarily stored in the reception buffer, and is read from the reception buffer between times To ₀ and To ₁ . Here, the reception interval ΔTr ₁ = Tr ₁ −Tr ₀ is calculated as the time required to receive 100 packets constituting one packet group from the reception times of the packets P ₀ and P ₁₀₀ . Further, a reading interval ΔTo ₁ = To ₁ −To ₀ is calculated as the time required to read the digital data contained in 100 packets from the reading time of the digital data contained in the packets P ₀ and P ₁₀₀ . Next, the difference between the reception interval ΔTr ₁ and the reading interval ΔTo ₁ is detected, and the reception-side sampling frequency fr is corrected so as to reduce this difference. Similarly, the same processing is performed for the next 100 packets. Here, the reception interval ΔTr ₂ of the next packet group is calculated as ΔTr ₂ = Tr ₂ −Tr ₁ from the difference in reception time between the packet P ₁₀₀ and the packet P ₂₀₀ . Further, the reading interval ΔTo ₂ of the next packet group is calculated as ΔTo ₂ = To ₂ −To ₁ from the difference in reading time of the digital data included in the packet P ₁₀₀ and the packet P ₂₀₀ . Then, the reception side sampling frequency fr is corrected so that the difference between the reception interval ΔTr ₁ and the reading interval ΔTo ₁ becomes small. As described above, by repeating the same processing in units of 100 packets, the overflow and underflow of the reception buffer due to the sampling frequency error between the transmission device and the reception device can be prevented, and the quality of the reproduced data can be improved. Can be increased.

As described above, the correction of the reception-side sampling frequency fr is performed based on the reception interval and reading interval acquired by the reception device without receiving time information such as a packet transmission interval from the transmission device. Therefore, it is not necessary to transmit / receive time information between the transmission device and the reception device, and the reception-side sampling frequency fr can be corrected only by the reception device.
<First embodiment>
Next, a voice communication system to which the receiving apparatus of the present invention is applied will be described in detail with reference to the first embodiment. In the following, it is assumed that a predetermined number k (k is a natural number and k ≧ 2) packets are used as one packet group to calculate the reception interval and the reading interval.
[Constitution]
FIG. 2 is a block diagram showing the configuration of the voice communication system according to the first embodiment. The voice communication system 10 is configured by connecting a transmission device 1 and a reception device 2 via a network 3 such as a LAN. In FIG. 2, only one transmitting device 1 and one receiving device 2 connected to the network 3 are shown, but normally a plurality of transmitting devices 1 and receiving devices 2 are connected to the network 3. Further, the transmission device 1 and the reception device 2 may be configured integrally.

  The transmission apparatus 1 includes an input unit 11, an amplifier 13, an ADC 15, an input buffer 17, an encoding unit 19, a transmission buffer 21, a transmission unit 23, and a VCO 25. The sound input to the input unit 11 such as a microphone is converted into an analog sound signal. The analog audio signal is amplified by the amplifier 13, converted to digital audio data by the ADC 15 based on the sampling frequency fs generated by the VCO 25, and stored in the input buffer 17. The encoding unit 19 encodes the digital audio data stored in the input buffer 17 into compressed audio encoding information and stores it in the transmission buffer 21. Note that the VCO 25 can change the sampling frequency fs generated by the applied voltage. The transmission unit 23 assembles the compressed audio coding information of the transmission buffer 21 into a packet, and transmits the packet to the network 3 at regular time intervals.

  The reception device 2 includes a reception unit 31, a reception time acquisition unit 33, a reception buffer 35, a reception buffer monitoring unit 37, a reading unit 39, a decoding unit 41, a reproduction buffer 43, a reading time acquisition unit 45, a DAC 47, an amplifier 49, and an output. Part 51, correction part 53, and VCO 55. A packet sent to the network 3 is received and disassembled by the receiving unit 31. The decomposed compressed audio coding information is stored in the reception buffer 35. The reading unit 39 is activated every time a later-described reading time acquisition unit 45 counts digital voice data for one packet, and outputs compressed audio coding information for one packet from the reception buffer 35 to the decoding unit 41. The compressed audio coding information is decoded and expanded by the decoding unit 41 to become digital audio data. This digital audio data is stored in the reproduction buffer 43. Then, the reading time acquisition unit 45 reads the digital audio data from the reproduction buffer 43 at the reception side sampling frequency fr. The read digital audio data is converted into an analog audio signal by the DAC 47 at the reception side sampling frequency fr. This reception side sampling frequency fr is generated by the VCO 55. The analog audio signal is amplified by the amplifier 49, and is amplified into the air from the output unit 51 such as a speaker. The reception buffer monitoring unit 37 monitors the amount of data in the reception buffer 35. Here, when the encoding unit 19 on the transmission device 1 side is omitted, the decoding unit 41 of the reception device 2 can be omitted. The output unit 51 such as a speaker may be provided independently of the receiving device 2.

The reception time acquisition unit 33 receives a packet every time a predetermined number k (k is a natural number, k ≧ 2) constituting one packet group is received by the reception unit 31, that is, every predetermined number k. Get the time. Then, a reception interval is calculated as a time required to receive a predetermined number k packets constituting one packet group from the difference between the reception times.
The read time acquisition unit 45 counts one packet of digital audio data (hereinafter referred to as reproduction data) to be reproduced after being read from the reproduction buffer 43 by the DAC 47 and converted to analog. The reproduction data amount is counted, for example, by counting interrupts that occur every time digital audio data for one packet is reproduced. In addition, the reading time acquisition unit 45 calls the reading unit 39 every time the reproduction data is counted for one packet, and causes the reading unit 39 to read the compressed audio coding information for the next one packet from the reception buffer 35. Further, the read time acquisition unit 45 counts a predetermined number k of reproduced data every time compressed audio coding information included in a predetermined number k packets constituting one packet group is read from the reception buffer 35. Read time is acquired every time. Then, the readout interval is calculated as the time required to read out the compressed speech encoded information included in the predetermined number k packets constituting one packet group from the difference in the obtained readout times.

The correction unit 53 acquires the reception interval and the read interval from the reception time acquisition unit 33 and the read time acquisition unit 45, and corrects the reception-side sampling frequency fr so that the error between the reception interval and the read interval becomes small.
The receiving device receives a packet in response to a transmission interval of packets from the transmitting device to the receiving device. That is, the packet reception interval responds to the packet transmission interval. Further, the transmission side sampling frequency fs is inversely proportional to the packet transmission interval. Similarly, the receiving-side sampling frequency fr is inversely proportional to the packet reading interval. Therefore, the reception-side sampling frequency fr can be brought close to the transmission-side sampling frequency fs by correcting the reception-side sampling frequency fr so as to reduce the error between the reception interval and the reading interval. Therefore, underflow and overflow caused by sampling frequency errors on the transmission side and the reception side can be suppressed, and the quality of reproduced data can be improved.

  The correction unit 53 may correct the reception-side sampling frequency fr so that the reception interval and the reading interval have a predetermined relationship. That is, the reception-side sampling frequency fr is corrected so that the reception interval and the read-out interval are not limited to a 1: 1 relationship, for example, a predetermined relationship such as 1: 2, 1: 3. By correcting in this way, the reception-side sampling frequency fr and the transmission-side sampling frequency fs can be in a predetermined relationship. Therefore, for example, underflow and overflow can be suppressed while supporting reading at double speed, triple speed, and the like.

In the following, a case where the reception-side sampling frequency fr is corrected so that the reception interval and the reading interval have a 1: 1 relationship will be described as an example. This will be described in more detail.
(1) Calculation Method for Reception Interval and Read Interval First, a method for obtaining a reception interval and a read interval will be described in more detail with reference to FIG. Here, the reception unit 31 sequentially packet P ₁ from the first packet P _0, the packet P ₂ ... have received the packet group G ₁ by a predetermined number k packets, packet group G ₂ ... packet group G _n Is configured. Here, the subscript of the packet P is a number indicating the transmission order of the packets, and the subscript of the packet group _Gn is a number indicating the transmission order of the packet groups. Further, the reception time Tr _n, read intervals the To _n, reception interval [Delta] Tr _n, each subscript of the read interval .DELTA.To _n corresponds to the subscript of the packet group G _n. The subscript n is a natural number of 0 or more. Hereinafter, a method for calculating the reception time and the readout time for each packet group will be described.
a) Regarding the _first packet group G _{1 The} reception time acquisition unit 33 acquires the reception time Tr ₀ of the first packet P ₀ included in the first packet group G ₁ at the reception unit 31. Next, the reception time Tr ₁ at the reception unit 31 of the _kth packet P _k from the first packet P ₀ is acquired. Then, the reception time acquisition unit 33 determines the reception interval ΔTr ₁ as the time required to receive the predetermined number k packets constituting the _first packet group G ₁ from the difference between the reception times Tr ₁ and Tr _0. = Tr ₁ −Tr ₀ is calculated.

On the other hand, the read time acquisition unit 45 acquires the read time To _{0 at} which the compressed audio coding information included in the first packet P ₀ is read from the reception buffer 35. Next, the read time To ₁ of the compressed audio coding information included in the _kth packet P _k from the first packet P ₀ is acquired. Then, the read time acquisition unit 45 required to read the compressed speech encoded information included in the predetermined number k packets constituting the _first packet group G ₁ from the difference between the read times To ₁ and To ₀ . As the time, the reading interval ΔTo ₁ = To ₁ −To ₀ is calculated.
b) for the second packet group as the reception time of the first packet included in the second packet group G _2, reception time Tr ₁ of the first of the last packet P _k packet group G ₁ it is used. Next, the reception time acquisition unit 33 acquires the reception time Tr ₂ at the reception unit 31 of the kth packet P _2k from the packet P _k . Then, the reception time acquisition unit 33 calculates the reception interval ΔTr ₂ = Tr ₂ −Tr ₁ of the second packet group G ₂ from the difference between the reception times Tr ₂ and Tr ₁ .

Similarly, the read time acquisition unit 45, a read time To ₁ compression speech coding information included in the packet P _k, read interval from the difference between the read time To ₂ of the compressed audio coded information contained in the packet P _2k ΔTo ₂ = To ₂ −To ₁ is calculated.
c) About the n-th packet group The reception time acquisition unit 33 obtains the reception time Tr _(n-1) of the first packet P _{(n-1) k} included in the n-th packet group G _n at the reception unit 31. get. Next, the reception time Tr _n of the _k-th packet P _{nk at} the receiving unit 31 is acquired from the first packet P _{(n−1) k} . Then, the reception time acquisition unit 33 calculates the time required to receive the predetermined number k packets constituting the _nth packet group G _n from the difference between the reception times Tr _n and Tr _(n−1) . The reception interval ΔTr _n = Tr _n −Tr _(n−1) is calculated. Here, P _{(n-1) k} is the last packet included in the packet group G _(n-1) .

On the other hand, the read time acquisition unit 45 reads the read time To _{(n−1) of the} compressed speech encoded information included in the packet P _{(n−1) k} and the compression included in the packet P _nk. calculating a read interval _{ΔTo n = to n -To (n} -1) from the difference between the read time the to _n speech coding information.
(2) Correction of Receiving-Sampling Frequency fr The correction unit 53 receives as follows so that the error between the reception interval acquired from the reception time acquisition unit 33 and the read interval acquired from the read time acquisition unit 45 is reduced. The side sampling frequency fr is corrected.
a) the receiving side the sampling frequency fr ₁ corrected for the first packet group G ₁ in the first packet group G ₁ is calculated by the following equation.

fr ₁ = fr ₀ × (ΔTo ₁ ÷ ΔTr ₁ )
Here, fr _{0 is} the reception-side sampling frequency before correction.
b) the receiving side the sampling frequency fr ₂ after correction for the second packet group G ₂ in the second packet group G ₂ is calculated by the following equation.

fr ₂ = fr ₁ × (ΔTo ₂ ÷ ΔTr ₂ )
Here, fr _{1 is} the reception-side sampling frequency after correction in the _first packet group G ₁ .
c) About the n-th packet group G ₂ The corrected reception-side sampling frequency fr _n in the n-th packet group G _n is calculated as follows.

fr _n = fr _(n−1) × (ΔTo _n ÷ ΔTr _n )
Here, fr _{(n- 1)} is the corrected reception-side sampling frequency in the _(n-1) th packet group G _(n-1) .
[Packet structure]
FIG. 4A shows a configuration example of a packet to be transmitted / received. The packet includes an IP header, a UDP header, and voice digital data. In the IP header, a source IP address, a destination IP address, and the like are described. In the UDP header, a transmission source port number, a transmission destination port number, a message length, and the like are described. Note that a TCP packet can be used instead of a UDP packet in the data portion of the IP packet. In this case, a TCP header is used instead of the UDP header.

FIG. 4B shows another configuration example of the transmitted / received packet. The packet shown in FIG. 4B further includes an RTP header. The RTP header includes the following information, for example.
a) Version b) Padding c) CSRC: Indicates how many CSRC identifiers are included.

d) Payload type e) Sequential number: Indicates the order of packets.
g) SSRC: transmitting device identifier h) CSRC: original sender identifier list.
When the packet is configured as shown in FIG. 4B, the reception time acquisition unit 33 and the read time acquisition unit 45 acquire the reception time and the read time based on the sequential number.
[processing]
Next, the flow of processing performed by the receiving apparatus 2 having the above configuration will be described in detail. The receiving device 2 executes five processes: a reception process, a reproduction process, a reception time acquisition process, a read time acquisition process, and a correction process. Hereinafter, these processes will be specifically described with reference to the drawings.
(1) Reception Processing FIG. 5 is a flowchart illustrating an example of the flow of reception processing performed by the reception device 2. The receiving device 2 starts the following process by receiving a command indicating the start of communication from the transmitting device 1.

Step S1: The receiving unit 31 performs the following steps S2 to S6 until a command indicating the end of communication is received from the transmission device 1. When a command indicating the end of communication is received, the process ends.
Step S2: The receiving unit 31 waits for a packet, and when receiving a packet, proceeds to step S3.

Steps S3 and S4: The reception buffer monitoring unit 37 determines whether or not there is an empty memory in which new compressed audio coding information is written in the reception buffer 35 (S3). If there is free memory, the reception unit 31 stores the received packet in the reception buffer 35 (S4).
Steps S5 and S6: When there is no free memory in the reception buffer 35, the reception unit 31 discards the old compressed audio coding information stored in the reception buffer 35 (S5), and the compressed audio of the newly received packet The encoded information is written in the reception buffer 35 (S6).

Through the above processing, the compressed audio coding information is accumulated in the reception buffer 35 one after another. In step S1, the case where a command indicating the end of communication is received from the transmission device 1 is described. However, the communication may be ended based on the command in the reception device 2. In steps S5 and S6, the old compressed speech coding information may be held in the reception buffer 35, and the compressed speech coding information of the newly received packet may be discarded.
(2) Reproduction Process FIG. 6 is a flowchart illustrating an example of the flow of the reproduction process performed by the reception device 2. The receiving device 2 starts the following process by receiving a command indicating the start of communication from the transmitting device 1.

Step S11: The reception buffer monitoring unit 37 determines whether or not the reproduction start condition is satisfied. As an example of the reproduction start condition, “data amount of compressed audio coding information accumulated in the reception buffer 35 ≧ threshold value D1” can be cited. As the threshold value D1, an experience value suitable for the communication system is obtained in consideration of the network load and the reproduction speed.
Step S12: The reception buffer monitoring unit 37 transmits a reproduction start signal to the reading unit 39 when the reproduction start condition is satisfied. Upon receiving the reproduction start signal, the reading unit 39 starts the audio reproduction process.

Step S13: Until the reception device 2 receives a command indicating the end of communication from the transmission device 1, the following reproduction processing from steps S14 to S17 is performed. When a command indicating the end of communication is received, the reproduction process is terminated.
Step S14: When the reproduction is continued, the reading unit 39 reads the compressed audio coding information for each packet from the reception buffer 35 in response to the call of the reading time acquisition unit 45. Next, the decoding unit 41 sequentially decompresses and decodes the compressed audio coding information included in one packet, and stores it in the reproduction buffer 43. The digital audio data stored in the reproduction buffer 43 is converted into an analog audio signal based on the corrected reception side sampling frequency fr by the DAC 47, amplified by the amplifier 49, and amplified in the air from the output unit 51. Reproduction is performed until all the compressed audio coding information included in one packet is converted into analog data and output.

Steps S15 and S16: When the reproduction of one packet is completed, the reception buffer monitoring unit 37 determines whether there is compressed audio coding information stored in the reception buffer 35 (S15). When the compressed audio coding information is stored in the reception buffer 35, the reading unit 39 reads the compressed audio coding information for one packet from the reception buffer 35 (S16).
Step S17: When the compressed audio coding information is not stored in the reception buffer 35, the reading unit 39 inserts the silence data for one packet into the reproduction buffer 43 according to the instruction from the reception buffer monitoring unit 37.

In the above description, the case where a command indicating the start or end of communication is received from the transmission device 1 has been described. However, communication may be started or ended based on a command in the reception device 2.
(3) Reception Time Acquisition Process FIG. 7 is a flowchart illustrating an example of the flow of the reception time acquisition process performed by the reception device 2. As described above, the reception time acquisition unit 33 acquires the reception time at the reception unit 31 for each predetermined number k packets constituting one packet group.

Step S21: n = 0 is set as the initial value of the subscript n of the packet group G _n indicating the transmission order of the packet group.
Steps S22 and S23: The receiving unit 31 receives the first packet P ₀ included in the first packet group G ₁ (S22). The reception time acquisition unit 33 acquires the reception time Tr ₀ of the first packet P ₀ included in the first packet group G ₁ at the reception unit 31 (S23).

Step S24: The value of n is added.
Step S25: Until the reception device 2 receives a command indicating the end of communication from the transmission device 1, the reception time acquisition processing from Steps S24 to S30 is performed. When a command indicating the end of communication is received, the reception time acquisition process is terminated.
Steps S26 and S27: The receiving unit 31 sequentially receives packets (S26). The reception time acquisition unit 33 counts the number of packets every time the reception unit 31 receives a packet (S27).

Step S28: The reception time acquisition unit 33 determines whether a predetermined number k of packets constituting one packet group has been counted. If the predetermined number k has not been counted, the process returns to step S25, and the receiving unit 31 further receives a packet.
Step S29: If the predetermined number k is counted, the reception time acquisition unit 33 acquires the reception time Trn of the packet received by the reception unit 31 after counting the predetermined number k. Here, the reception time acquisition unit 33 acquires Tr1 as the reception time after counting a predetermined number k in the first packet group G1. The reception time Tr1 is the reception time of the first packet in the second packet group G2. That is, Tr (n−1) is acquired as the reception time of the first packet in the nth packet group G n , and Trn is acquired as the reception time after counting a predetermined number k.

Step S30: The reception time acquisition unit 33 calculates a reception interval from the difference between the reception time of the first packet and the reception time of the packet after counting a predetermined number k. ΔTr _n = Tr _n −Tr _(n−1) is calculated as a reception interval in the n-th packet group G _n .
In the above description, a case where a command indicating communication start or communication end is received from the transmission device 1 is described. However, communication may be started or ended based on a command in the reception device 2.
(4) Reading Time Acquisition Process FIG. 8 is a flowchart illustrating an example of a reading time acquisition process performed by the receiving device 2. Similar to the reception time acquisition unit 33, the read time acquisition unit 45 responds to reading out the compressed audio coding information included in a predetermined number k packets constituting one packet group from the reception buffer 35. Get the reading time. That is, the reading time acquisition unit 45 acquires the reading time every time a predetermined number k of reproduction data is counted.

Step S31: n = 0 is set as the initial value of the subscript n of the packet group G _n indicating the transmission order of the packet group.
Steps S32 and S33: The reading time acquisition unit 45 calls the reading unit 39 after counting the first reproduction data for one packet. In response to the call, the reading unit 39 reads the compressed audio coding information included in the first packet P ₀ included in the first packet group G ₁ from the reception buffer 35 (S32). Then, the read time acquisition unit 45 acquires the read time Tr ₀ of the compressed speech encoded information included in the first packet P ₀ included in the first packet group G ₁ (S33).

Step S34: Next, the value of n is added.
Step S35: Read time acquisition processing from steps S34 to S40 is performed until the reception device 2 receives a command indicating the end of communication from the transmission device 1. When a command indicating the end of communication is received, the read time acquisition process is terminated.
Steps S36 and S37: The read time acquisition unit 45 counts the reproduction data to be reproduced after being read from the reproduction buffer 43 by the DAC 47 and converted to analog (S36). The reading time acquisition unit 45 calls the reading unit 39 every time the reproduction data is counted for one packet, and the reading unit 39 sequentially receives the compressed audio coding information from the reception buffer 35 in response to the call of the reading time acquisition unit 45. Is read (S37). Here, when the compressed audio coding information is not stored in the reception buffer 35, silence data is inserted in the reproduction buffer 43, and the read time acquisition unit 45 performs the silence data in the same manner as the reproduction data count. Count.

Step S38: The read time acquisition unit 45 determines whether or not the reproduction data amount has been counted by a predetermined number k of packets constituting one packet group. If the predetermined number k has not been counted, the process returns to step S35, and the compressed speech encoded information is further read by the reading unit 49.
Step S39: If the predetermined number k is counted, the reading time acquisition unit 45 acquires the reading time Ton of the compressed speech encoded information read from the reception buffer after counting the predetermined number k. If the compressed audio coding information is not accumulated in the reception buffer 35 and silence data is inserted into the reproduction buffer 43, the time when the reading unit 49 inserts the silence data is set as the compressed audio coding information. Acquired as read time. Alternatively, the time when the reading time acquisition unit 45 calls the reading unit 39 is acquired as the reading time. Here, the read time acquisition unit 45 acquires To1 as the read time of the compressed speech encoded information after a predetermined number k counts in the first packet group G1. The readout time To1 is the readout time of the compressed speech encoded information included in the first packet in the second packet group G2. That is, To (n-1) is acquired as the read time of the first packet in the nth packet group Gn , and Ton is acquired as the read time after counting a predetermined number k.

Step S40: The read time acquisition unit 45 reads the read interval from the difference between the read time of the compressed speech encoded information included in the first packet and the read time of the compressed speech encoded information included in the packet after a predetermined number k counts. Is calculated. _{Therefore, ΔTo n = To n -To (} n-1) as a read interval in the n-th packet group G _n is calculated.
In the above description, a case where a command indicating communication start or communication end is received from the transmission device 1 is described. However, communication may be started or ended based on a command in the reception device 2.
(5) Correction Processing FIG. 9 is a flowchart showing an example of the flow of correction processing.

Step S41: correction unit 53, a receiving interval [Delta] Tr _n from the receiving time acquiring unit 33 determines whether it has acquired the read interval .DELTA.To _n from the read time acquisition unit 45.
Step S42: When the receiving interval and the reading interval are acquired, the receiving side sampling frequency fr is corrected so that the error between the receiving interval and the reading interval is reduced. Here, the corrected receiving side sampling frequency fr _n in the n-th packet group G _n is calculated as follows.

fr _n = fr _(n−1) × (ΔTo _n ÷ ΔTr _n )
Here, fr _{(n- 1)} is the corrected reception-side sampling frequency in the _(n-1) th packet group G _(n-1) .
More specifically, the correction unit 53 corrects the reception side sampling frequency fr _n by correcting the voltage value Vc applied to the VCO 55. For example, the new voltage value Vc can be calculated as follows.

Vc (V) = Vcp + D · Vmax / 2F
Where Vcp: voltage value before correction
Vmax: Upper limit of VCO control voltage
F: Amplitude of VCO control voltage (+ direction and-direction)
D = A · (ΔTr _n −ΔTo _n ) / ΔTo _n
A: Conversion coefficient
It is.

Then, the correction unit 53 outputs the calculated voltage value Vc to the VCO 55. As a result, the reception-side sampling frequency fr _n is generated based on the new voltage Vc.
Step S43: When the correction unit 53 has not received a communication end signal, the correction unit 53 performs the processes of steps S41 to S43 for the next packet group.
Here, the correction of the reception-side sampling frequency fr _n can be performed using the coefficient a as in the following equation.

fr _n = fr _(n-1) +
a × {fr _(n−1) × (ΔTo _n ÷ ΔTr _n ) −fr _(n−1) }
By giving the coefficient a in this way, a change in the reception-side sampling frequency fr due to the correction is set gently. For example, the reception-side sampling frequency fr is gradually brought closer to the transmission-side sampling frequency fs by several corrections. Therefore, a sudden change in sound quality due to correction can be mitigated, and a decrease in sound quality can be reduced. Further, when there is a measurement error, it is possible to reduce malfunctions such as when the reception-side sampling frequency fr is corrected in response to the measurement error immediately.

Further, as described above, the correction is not limited to the correction so that the reception interval and the reading interval have a relationship of 1: 1, but the reception-side sampling frequency fr is set so as to have a predetermined relationship such as 1: 2. It may be corrected.
[effect]
As described above, the reception-side sampling frequency fr can be corrected based on the reception interval and reading interval acquired by the reception device 2. Therefore, the amount of data stored in the reception buffer can be controlled, overflow and underflow can be prevented, and the quality of reproduced data can be improved. Further, the reception time and the read time are acquired on the receiving device side. Therefore, it is not necessary to transmit the time information related to the packet transmission time and the packet transmission interval from the transmission device side to the reception device side. Therefore, by simplifying the packet configuration of the packet to be transmitted and reducing the time of incorporation of the time information into the packet and the amount of packet data, the packet transmission efficiency can be improved and the bandwidth used can be reduced.

The packets need not be transmitted from the transmission apparatus 1 at regular intervals, and the first embodiment can be applied even when packets are transmitted at irregular intervals. However, when packets are transmitted from the transmission apparatus 1 at regular intervals, the reception intervals and readout intervals of individual packets can be matched by correcting the reception-side sampling frequency fr. That is, in the above, since the correction of the reception side sampling frequency fr is performed for each packet group, it is difficult to correct the difference between the reception interval and the reading interval for each individual packet constituting the packet group. However, when packets are transmitted from the transmission apparatus 1 at regular intervals, the reception intervals of the individual packets and the digital data reading intervals can be matched by correcting the reception-side sampling frequency fr. Therefore, the reception buffer amount is stabilized, and underflow and overflow can be suppressed more precisely.
<Second Embodiment>
In the second embodiment, countermeasures are taken against network failures such as packet loss, packet arrival delay and fluctuation while suppressing underflow and overflow.
[Constitution]
The configuration of the voice communication system according to the second embodiment is the same as that of the first embodiment shown in FIG. Further, the transmitted packet includes the RTP header shown in FIG. 4B of the first embodiment. Therefore, the receiving unit 31 acquires the packet sequential number SN _n from the received packet. The reception buffer monitoring unit 37 monitors the sequential number SN _{n of the} packet. Furthermore, the reception buffer monitoring unit 37 monitors the amount of data stored in the reception buffer 35 in order to detect network failures such as packet loss, delay in packet arrival at the receiving device, and fluctuations.

In the following, it is assumed that a predetermined number k (k is a natural number and k ≧ 2) packets are used as one packet group to calculate the reception interval and the reading interval. In addition, an average of reception intervals required to receive a predetermined number k of packets is ΔTrav.
Next, the flow of processing performed by the receiving device 2 will be described in detail together with countermeasures against network failures. First, a network failure will be described.
[Network failure]
A packet received from the transmission device 1 is lost due to a network failure, or the arrival time of the received packet is delayed. Network failures are classified into three types, for example, (a) packet lost, (b) burst lost, and (c) fluctuation, and each will be described with reference to FIG. FIG. 10 is an explanatory diagram showing a packet transmitted from the transmission apparatus 1 and received by the reception unit 31 together with a transition of time.
a) Packet Loss FIG. 10A is an explanatory diagram of packet loss. The packets P ₉₈ and P _{99 with} sequential numbers SN ₉₈ and SN ₉₉ are normally received by the receiving unit 31. However, a loss occurs in the packet transmitted from the transmission apparatus 1, and the packets P ₁₀₀ and P _{101 of} the sequential numbers SN ₁₀₀ and SN ₁₀₁ cannot be received by the reception unit 31. Then, the receiving unit 31 receives the packet P ₁₀₂ sequential numbers SN ₁₀₂ to the next packet P _99. Therefore, the valid compressed speech coding information included in the packets P ₁₀₀ and P ₁₀₁ is not stored in the reception buffer 35, but invalid compressed speech coding information corresponding to the packets P ₁₀₀ and P ₁₀₁ is stored. Reading from the buffer 35 to the reproduction buffer 43 is performed. Here, a case where the amount of data stored in the reception buffer 35 does not become “0” is defined as packet lost. That is, in the case of packet lost, the lost packet data can be absorbed by the reception buffer 35.
b) Burst Lost FIG. 10B is an explanatory diagram of burst lost. Up to the packet P _{11 with the} sequential number SN ₁₁ is normally received by the receiving unit 31. However, a loss occurs in the packet transmitted from the transmission device 1, and packets P ₁₂ to P ₅₁ with sequential numbers SN _{12 to} SN ₅₁ cannot be received by the receiving unit 31. Then, the receiving unit 31 receives the packet P ₅₂ with the sequential number SN ₅₂ next to the packet P ₁₁ . Defined herein, by the packet P ₁₂ ~ packet P ₅₁ to the sequential number SN ₁₂ to SN ₅₁ is not stored in the reception buffer 35, a case where the amount of accumulated data in the receive buffer 35 becomes "0" and burst Lost To do. It differs from the packet lost described above in that the amount of data stored in the reception buffer 35 becomes “0” when a packet loss occurs.

When burst loss occurs, when a packet received after burst loss is received at a time when it is originally received unless there is a network failure, the received packet is written in the reception buffer 35. For example, packets up to the packet P _{51 with} the sequential number SN ₅₁ are lost, but if the packet P ₅₂ is received at the time when it is originally received, the packet P ₅₂ is written in the reception buffer 35. On the other hand, if it is received at a time different from the original time, it is discarded as an old packet is received.
c) Fluctuation FIG. 10C is an explanatory diagram of fluctuation. Up to the packet P _{11 with the} sequential number SN ₁₁ is normally received by the receiving unit 31. However, there is a delay until the receiving unit 31 receives the next packet P ₁₂ of the sequential number SN ₁₁ due to network fluctuations. Here, a case where the amount of data stored in the reception buffer 35 becomes “0” due to a delay in reception of the packet P _{12 by} the reception unit 31 is defined as fluctuation.

When fluctuations occur, packets received after the fluctuations are received at short time intervals, delayed from the time at which the reception unit 31 originally receives the packets. Therefore, the received packet is discarded until the packet received after the fluctuation is received at the original reception time. When the packet is received at the original reception time, the received packet is written into the reception buffer 35.
[processing]
Next, processing performed in each unit will be described.
(1) State determination processing First, state determination processing for whether a packet is received normally without packet loss or arrival time delay, or whether it is in a packet lost, burst lost, or fluctuation state due to a network failure will be described. To do. This state determination process is performed by the reception buffer monitoring unit 37 during a reproduction process described later. Hereinafter, it is assumed that the received packet includes a sequential number SN _n as shown in FIG. The reception buffer monitoring unit 37 monitors the sequential number SN _n of the packet received by the reception unit 31 and monitors the reception buffer amount in the reception buffer 35.
a) Whether the packets are normally received in the order of transmission without delay Whether the packets are received in the order of transmission is determined by monitoring the sequential number SN _n of the received packet by the reception buffer monitoring unit 37 and before and after the sequential number SN _n . This is done by determining whether there are any relationships or omissions. Further, whether or not there is a delay in packet reception is performed by monitoring the reception buffer amount in the reception buffer 35 by the reception buffer monitoring unit 37. For example, when a packet is lost or when reception of a packet is delayed, the compressed speech encoded information is read without storing the compressed speech encoded information in the reception buffer 35, and the reception buffer The amount decreases. It is judged normal by seeing that there is no decrease.
b) Whether the packet is lost or not is determined by determining whether the sequential number SN _n is missing by the reception buffer monitoring unit 37. Here, in the packet lost, the packet loss is such that it can be absorbed by the reception buffer 35. Therefore, the reception buffer amount in the reception buffer 35 decreases, but the reception buffer amount does not become “0”.
c) Whether the data is in the burst lost or fluctuation state or the burst lost or fluctuation state is determined by determining whether the reception buffer amount is “0”. In this case, since more packets than can be absorbed by the reception buffer 35 are lost, the reception buffer amount in the reception buffer 35 becomes “0”. When it is determined that the state is the burst lost or the fluctuation, the reception buffer monitoring unit 37 sets a “lost flag” indicating that the burst lost or the fluctuation is present.

Further, whether it is a burst lost or a fluctuation is performed, for example, by comparing sequential numbers SN _n as follows. As shown in FIGS. 10B and 10C, the sequential number of the packet received by the receiving unit 31 immediately before burst loss or fluctuation occurs is SN _last . Here, when burst lost or fluctuation occurs, the reception buffer monitoring unit 37 monitors that the reception buffer amount of the reception buffer 35 is “0”, and in response to an instruction from the reception buffer monitoring unit 37. Silence data is inserted into the reproduction buffer 43. If the silent data insertion count is Csd, the sequential number of the received packet that is supposed to be received if the burst lost or fluctuation does not occur is estimated. The estimated sequential number SN _e to be estimated can be calculated from SN _e = SN _last + Csd. Here, the sequential number of the received packet received immediately after the occurrence of burst lost or fluctuation is SN _a . The sequential number SN _a is compared with the estimated sequential number SN _e . When the difference between the estimated sequential number SN _e and the sequential number SN _a is within an allowable range, it is determined that the burst is lost. That is, the packets up to the estimated sequential number SN _e are lost, but the packets received thereafter are received at the time when they are originally received. On the other hand, when it is not within the allowable range, there is a gap between the estimated sequential number SN _e and the sequential number SN _a, and it is determined that there is a delay in the reception time of the packet, and that the state is a fluctuation.
(2) Reception processing Next, reception processing will be described. FIG. 11 is a flowchart illustrating an example of a flow of reception processing performed by the reception device 2. The receiving device 2 starts the following process by receiving a command indicating the start of communication from the transmitting device 1.

Step S51: The receiving unit 31 performs the following steps S52 to S61 until a command indicating the end of communication is received from the transmission device 1. When a command indicating the end of communication is received, the process ends.
Step S52: The receiving unit 31 waits for a packet, and when receiving a packet, proceeds to step S53.

Step S53: The reception buffer monitoring unit 37 checks whether or not the “lost flag” indicating that the burst lost or fluctuation state is set. The “lost flag” indicates a result of a state determination process performed during a reproduction process described later.
Step S54: If the “lost flag” is not set, that is, if the packet is normally received or the packet is lost, it is determined whether or not writing to the reception buffer 35 is possible. This is because the reception of a packet has a delay that exceeds an allowable range, and an old packet is received. If it is determined that writing is possible, the process proceeds to step S55. If it is determined that writing is not possible, the process proceeds to step S61.

Step S55: The reception buffer monitoring unit 37 determines whether or not there is an empty memory in which new compressed audio coding information is written in the reception buffer 35.
Step S56: If there is free memory, the receiving unit 31 writes the received packet in the reception buffer 35.
Steps S57 and S58: When there is no free memory in the reception buffer 35, the reception unit 31 discards the old compressed audio coding information stored in the reception buffer 35 (S57), and the compressed audio of the newly received packet. The encoded information is written in the reception buffer 35 (S58).

Step S59: If the “lost flag” is set, that is, if it is in a burst lost or fluctuation state, it is determined whether or not writing to the reception buffer 35 is possible. This is because in the case of burst lost, the received packet is originally received at the time it is received, but in the case of fluctuation, the packet reception is delayed more than the allowable range and an old packet is received. is there. Therefore, in the case of fluctuation, it is necessary to discard the old packet received after a delay exceeding the allowable range. Therefore, the reception buffer monitoring unit 37 makes a comparison between the sequential number SN _a received packet currently being received and the estimated sequential number SN _e. Then, based on the comparison result, it is determined whether or not to write the compressed audio coding information included in the currently received packet to the reception buffer 35. If the received packet is received at the time it is originally received, it is determined to be written. On the other hand, if the received packet has been received after a delay of an allowable range or more than the time when it was originally received, it is determined not to be written.

Step S60: If it is determined in step S59 that writing is to be performed, a “write restart flag” indicating that the compressed audio coding information included in the currently received packet is to be written to the reception buffer 35 is set. Next, the process proceeds to step S55, and writing to the reception buffer 35 is performed.
Step S61: On the other hand, if it is determined in step S59 that writing is not performed, the currently received packet is discarded.

Through the above processing, the compressed audio coding information is accumulated in the reception buffer 35 one after another. In step S51, the case where a command indicating the end of communication is received from the transmission device 1 is described. However, the communication may be ended based on the command in the reception device 2. In steps S57 and S58, the old compressed speech coding information may be held in the reception buffer 35, and the compressed speech coding information of the newly received packet may be discarded.
(2) Reproduction Process Next, the reproduction process will be described. FIG. 12 is a flowchart illustrating an example of the flow of reproduction processing performed by the reception device 2. The receiving device 2 starts the following process by receiving a command indicating the start of communication from the transmitting device 1.

Steps S62 and S63: The reception buffer monitoring unit 37 determines whether or not the reproduction start condition is satisfied, and if satisfied, transmits a reproduction start signal to the reading unit 39 (S62). The reading unit 39 that has received the reproduction start signal starts the audio reproduction process (S63).
Step S64: Until the reception device 2 receives a command indicating the end of communication from the transmission device 1, the reproduction processing from steps S64 to S77 is performed. When a command indicating the end of communication is received, the reproduction process is terminated.

  Step S65: When the reproduction is continued, the reading unit 39 reads the compressed audio coding information for each packet from the reception buffer 35 in response to the call of the reading time acquisition unit 45. Next, until the decoding unit 41, the reproduction buffer 43, the reading time acquisition unit 45, the DAC 47, the amplifier 49, and the output unit 51 convert all the compressed audio coding information contained in one packet into analog data and output it, Playback is performed.

  Step S66: When the reproduction for one packet is completed, the reception buffer monitoring unit 37 checks whether or not the “lost flag” indicating that the burst lost or fluctuation state is set. The “lost flag” at this time indicates a state before one packet is reproduced in step S65. If the “lost flag” is not set, that is, if the packet is normally received or the packet is lost before reproducing one packet in step S65, the process proceeds to step S67. On the other hand, if the “lost flag” is set, that is, if it is in a burst lost or fluctuation state, the process proceeds to step S74.

  Steps S67 and S68: If the packet is normally received or is in a packet lost state, the reception buffer monitoring unit 37 reproduces one packet in step S65, and then stores the compressed audio code in the reception buffer 35. It is determined whether or not the digitized information is accumulated (S67). When the compressed audio coding information is stored in the reception buffer 35, the reading unit 39 reads the compressed audio coding information for one packet from the reception buffer 35 (S68).

Step S69: If compressed audio coding information is not accumulated in the reception buffer 35, the reception buffer monitoring unit 37 performs a state determination process after reproducing one packet in step S65.
Steps S70 and S71: If it is determined as a result of the state determination process in step S69 that the burst lost or fluctuation state is not detected, the reading unit 39 reproduces silence data for one packet according to an instruction from the reception buffer monitoring unit 37. Insert into buffer 43.

  Steps S72 and 73: If it is determined in step S70 that the burst lost or fluctuation state exists, the reception buffer monitoring unit 37 sets a “lost flag” (S72). At this time, since the reception unit 31 has not received the packet normally, the reception time acquisition unit 33 cannot acquire the reception time of the packet. Therefore, the reception time acquisition unit 33 sets the average reception interval ΔTav required to receive a predetermined number k of packets constituting one packet group as the reception interval of the packet group that lost the packet (S73). Then, it progresses to step S71 and silence data is inserted.

  Step S74: If it is determined in step S66 that the state before reproducing one packet in step S65 is a burst lost or fluctuation state, it is determined whether or not a “write resume flag” is set. The “write restart flag” is a flag set in step S60 of the reception process, and indicates whether or not the currently received reception packet is to be written in the reception buffer 35.

  Steps S75 and S76: When the “write resume flag” is set, that is, when the currently received reception packet is written to the reception buffer 35, the “lost flag” is terminated (S75). In this case, the writing of the currently received packet to the reception buffer 35 is resumed, and the packet is lost from the burst lost or fluctuation state. Then, the reading unit 39 reads the compressed audio coding information for one packet from the reception buffer 35 (S76).

Step S77: If the “write resume flag” is not set, that is, if the currently received packet is discarded, the reading unit 39 reproduces the silence data for one packet according to the instruction of the reception buffer monitoring unit 37. 43.
In the above description, a case where a command indicating communication start or communication end is received from the transmission device 1 is described. However, communication may be started or ended based on a command in the reception device 2.
(3) Reception Time Acquisition Processing FIG. 13 is a flowchart illustrating an example of the flow of reception time acquisition processing performed by the reception device 2. As described above, the reception time acquisition unit 33 acquires the reception time of a packet every time a predetermined number k packets constituting one packet group are received by the reception unit 31, that is, every predetermined number k.

Hereinafter, a sequential number serving as a reference for monitoring packet loss and reception time delay is referred to as a reference sequential number SN _n . Here, the subscript n of the reference sequential number SN _n corresponds to the packet group G _n .
Step S81: At the start of the reception time acquisition process, the reception buffer monitoring unit 37 sets _n = as the initial value of the reference sequential number SN _{n and} the initial value of the subscript n of the packet group G _n indicating the transmission order of the packet group. Set to 0.

Steps S82 and S83: The receiving unit 31 receives the first packet P ₀ included in the first packet group G ₁ (S82). Then, the reception buffer monitoring unit 37 acquires the sequential number SN _r of the first packet P ₀ included in the first packet group G ₁ from the RTP header (S83).
Step S84: Next, the reception buffer monitoring unit 37 sets the acquired first sequential number SN _r to the reference sequential number SN ₀ of the first packet P ₀ (SN ₀ = SN _r ). For example, when the sequential number of the acquired packet P ₀ is “3”, SN ₀ = SN ₃ .

Step S85: The reception time acquisition unit 33 acquires the reception time Tr ₀ of the first packet P _{0 at} the reception unit 31.
Step S86: Next, n is added. By adding 1 to n, the reference sequential numbers are SN ₁ , SN ₂ , SN ₃ .
Step S87: A predetermined number k, which is the number of packets constituting one packet group, is added to the reference sequential number SN _n-1 before addition in step S86 to calculate the reference sequential number SN _n . For example, reference sequential numbers SN ₁ = SN ₀ + k, SN ₂ = SN ₁ + k,. The reception interval is calculated based on the time when the packet having the reference sequential number SN _n is received.

Step S88: Reception time acquisition processing is performed until the reception device 2 receives a command indicating the end of communication from the transmission device 1. When a command indicating the end of communication is received, the reception time acquisition process is terminated.
Steps S89 and S90: The receiving unit 31 sequentially receives packets (S89). The reception buffer monitoring unit 37 acquires the sequential number SN _r of the received packet from the RTP header (S90).

Step S91: The reception buffer monitoring unit 37 determines whether it is in a burst lost or fluctuation state.
Step S92: If it is not the state of the burst lost or wobble, sequential number SN _r packet acquired in step S90 is compared whether the reference sequential number SN _n or calculated in step _{S87 (SN r ≧ SN n)} . If SN _r <SN _n , the process returns to step S88 to receive the packet. For example, when the reference sequential number SN ₀ = SN ₃ , the next reference sequential number SN ₁ is calculated as SN ₁ = SN _{(3 + k)} in step S87. Therefore, steps S88 to S92 are repeated until the sequential number SN _r of the received packet becomes SN _r = SN _{(3 + k)} .

Step S93: If SNr ≧ SNn, it is determined whether or not the reference sequential number SNn is equal to the sequential number SNr of the received packet (SNr = SNn).
Step S94: When SNr = SNn, that is, when the packet is normally received without packet loss, the reception time acquisition unit 33 acquires Trn as the reception time of the packet with the sequential number SNr (SNr = SNn). To do. Here, when SNr = SN1, the reception time acquisition unit 33 acquires Tr1 as the reception time after receiving a predetermined number k packets in the first packet group G1. The reception time Tr1 is the reception time of the first packet in the second packet group G2. That is, Tr (n−1) is acquired as the reception time of the first packet in the n-th packet group G n , and Trn is acquired as the reception time after receiving a predetermined number k packets.

Step S95: If it is determined in step S93 that the reference sequential number SNn is not equal to the sequential number SNr, that is, if SNr> SNn and the packet is lost, the packet of the sequential number SNr currently received is firstly received. The reception time RTr of is acquired.
Step S96: Next, the number of lost packets LC lost due to packet loss is calculated from LC = SNr−SNn. Note that the number of missing packets can also be calculated by monitoring the reception buffer amount by the reception buffer monitoring unit 37. That is, since the packet is not stored in the reception buffer 35 due to packet loss, the packet loss number LC can be obtained by monitoring the decrease in the reception buffer amount by the reception buffer monitoring unit 37.

Step S97: Time Tc at which the packet could not be received is calculated using the packet loss number LC. Here, the average of reception intervals required to receive a predetermined number k of packets is ΔTrav, and the reception interval per packet is ΔTrav / k. Therefore, the time when the packet cannot be received is calculated as Tc = LC × ΔTrav / k.
Step S98: Next, the reception time when the last packet of one packet group is received without loss is estimated. The estimated reception time Te is calculated by subtracting the time Tc = LC × ΔTrav / k at which the packet could not be received from the current reception time RTr (Te = RTr−Tc).

This will be described in more detail with reference to FIG. If there is no packet loss, the receiving unit 31 receives the packet P100 having the sequential number SN100. That is, it is assumed that the reference sequential number is SN100. At this time, in FIG. 10A, the packets of P100 and the next packet P101 are lost, and the sequential numbers SN100 and SN101 cannot be received. Further, after the packet loss has received the packet of the sequential number SN102, the sequential number is S N102. Therefore, the number of lost packets LC = SN102−SN100 = 2 is calculated. Further, the reception time of the sequential number S N102 is RTr102.
Therefore, the time Tc = 2 × ΔTrav / 100 when the packet could not be received is calculated, and the estimated reception time Te = RTr102-2 × ΔTrav / 100 is calculated.

Step S99: When the reception time Tr _n is acquired in step S94, the reception time acquisition unit 33 calculates ΔTr _n = Tr _n −Tr _(n−1) as the reception interval in the n-th packet group G _n . . If the estimated reception time Te is calculated in step S98, ΔTr _n = Te−Tr _(n−1) is calculated as the reception interval in the _nth packet group G _n .
If it is determined in step S91 that the state is a burst lost or a fluctuation, the reception time acquisition unit 33 acquires the average ΔTrav of the reception intervals set in step S72 of the reproduction process as the reception interval.
(4) Reading time acquisition process and correction process The reading time acquisition process and the correction process are the same as those in the first embodiment, and a description thereof will be omitted.
[effect]
As described above, even when a network failure occurs, the reception-side sampling frequency fr can be corrected based on the reception interval ΔTr and the reading interval ΔTo acquired in the reception device 2. Therefore, the reception side sampling frequency fr can be corrected without transmitting the time information related to the transmission side sampling frequency fs such as the packet transmission time and the packet transmission interval from the transmission device side to the reception device side.

The packets need not be transmitted from the transmission device 1 at regular intervals, and the second embodiment can be applied even when packets are transmitted at irregular intervals. If the packet at regular intervals is transmitted, it is possible to more accurately calculate the estimated reception time Te from the foregoing DerutaTrav.
<Other embodiment examples>
(A) In the first and second embodiments described above, the voice communication system has been described. However, the receiving apparatus according to the present invention is preferably used not only for voice communication but also for moving picture communication. Further, it is also suitable as a receiving device for still images, text data, and the like.

  (B) A program that causes a computer to execute the above-described processing and a computer-readable recording medium that records the program are included in the present invention. Here, examples of the recording medium include a computer readable / writable flexible disk, hard disk, semiconductor memory, CD-ROM, DVD, magneto-optical disk (MO), and others.

  By using the present invention, it can be expected to improve the quality of audio reproduced by a receiving apparatus of a communication system that employs a stream reproduction method.

The concept explanatory drawing of this invention. 1 is a configuration diagram showing a configuration of a communication system according to a first embodiment. Explanatory drawing of the acquisition method of a receiving interval and a reading interval. (A) Explanatory drawing which shows an example of the packet structure in 1st Embodiment. (B) Explanatory drawing which shows another example of the packet structure in 1st Embodiment. The flowchart which shows the flow of the reception process in 1st Embodiment. The flowchart which shows the flow of the reproduction | regeneration processing in 1st Embodiment. The flowchart which shows the flow of the reception time acquisition process in 1st Embodiment. The flowchart which shows the flow of the reading time acquisition process in 1st Embodiment. The flowchart which shows the flow of the correction process in 1st Embodiment. (A) Explanatory drawing of packet lost. (B) Explanatory drawing of burst lost. (C) Explanatory drawing of fluctuation. The flowchart which shows the flow of the reception process in 2nd Embodiment. The flowchart which shows the flow of the reproduction | regeneration processing in 2nd Embodiment. The flowchart which shows the flow of the reception time acquisition process in 2nd Embodiment.

Explanation of symbols

1: Transmission device 2: Reception device 3: Network 31: Reception unit
33: Reception time acquisition unit
35: Receive buffer
37: Receive buffer monitoring unit
39: Reading unit 43: Playback buffer 45: Reading time acquisition unit 53: Correction unit

Claims (7)

A receiving device connected to a transmitting device via a communication network,
Receiving means for sequentially receiving a plurality of packets from the transmitting device via the communication network;
A reception buffer for temporarily storing digital data included in each of the plurality of packets;
A frequency generating means for generating a receiving side sampling frequency fr;
Conversion means for reading out digital data stored in the reception buffer according to the reception-side sampling frequency fr and converting it into analog data;
For each packet group G _n (n is a natural number) composed of a predetermined number k (k ≧ 2) packets determined on the receiving side, the time required for the receiving means to receive k packets is calculated. And a reading interval ΔTo indicating a time required for reading the digital data contained in each of the k packets from the receiving buffer, and obtaining the packet group G based on the following conditions: and correcting means for calculating a reception side sampling frequency fr in _n,
A receiving device.
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the receiving-side sampling frequency in the packet group G _(n−1) . As the frequency generation means, a VCO (Voltage Control Oscillator) is used,
The receiving apparatus according to claim 1, wherein the correction unit corrects the reception-side sampling frequency fr by adjusting a voltage applied to the VCO . Monitoring means for monitoring the transmission order of the packets received by the receiving means and the amount stored in the reception buffer, and further determining whether or not the packets are normally received;
The converting means includes
In response to the determination as to whether or not the monitoring means has received the packet normally, the stored digital data is read from the reception buffer,
The receive buffer is
The receiving apparatus according to claim 1 , wherein digital data is accumulated in accordance with a determination as to whether or not a packet is normally received by the monitoring unit and an accumulation amount in the reception buffer . A reception reproduction method used for a reception device connected to a transmission device via a communication network,
A reception step of sequentially receiving a plurality of packets from the transmission device via the communication network;
An accumulation step of temporarily accumulating digital data included in each of the plurality of packets in a reception buffer;
A frequency generation step of generating a reception-side sampling frequency fr;
A conversion step of reading the digital data stored in the reception buffer according to the reception-side sampling frequency fr and converting it into analog data;
Packet group G composed of a predetermined number k (k ≧ 2) packets determined on the receiving side _n For each (n is a natural number), the reception interval ΔTr indicating the time taken to receive k packets in the reception step and the digital data included in each of the k packets are read from the reception buffer. And a reading interval ΔTo indicating the time required for the packet group G based on the following conditions: _n A correction step for correcting the reception-side sampling frequency fr in
Receiving and playing method including:
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the packet group G _(N-1) Receiver sampling frequency at. A reception reproduction program used for a reception device connected to a transmission device via a communication network,
Receiving means for sequentially receiving a plurality of packets from the transmitting device via the communication network ;
Reception buffer for temporarily storing the digital data included in each of the plurality of packets,
A frequency generating means for generating a receiving side sampling frequency fr ;
Conversion means for reading digital data stored in the reception buffer according to the reception-side sampling frequency fr and converting it into analog data ;
For each packet group G _n (n is a natural number) composed of a predetermined number k (k ≧ 2) packets determined on the receiving side, the time required for the receiving means to receive k packets is calculated. And a reading interval ΔTo indicating a time required for reading the digital data contained in each of the k packets from the receiving buffer, and obtaining the packet group G based on the following conditions: correcting means for calculating a reception side sampling frequency fr in _n,
As a reception / reproduction program for causing a computer to function .
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the receiving-side sampling frequency in the packet group G _(n−1) . A computer-readable recording medium on which a reception / playback program used for a reception device connected to a transmission device via a communication network is recorded ,
A reception step of sequentially receiving a plurality of packets from the transmission device via the communication network;
An accumulation step of temporarily accumulating digital data included in each of the plurality of packets in a reception buffer;
A frequency generation step of generating a reception-side sampling frequency fr;
A conversion step of reading the digital data stored in the reception buffer according to the reception-side sampling frequency fr and converting it into analog data;
For each packet group G _n (n is a natural number) composed of a predetermined number k (k ≧ 2) packets determined on the receiving side, the time required to receive k packets in the reception step is determined. And a reading interval ΔTo indicating a time required for reading the digital data contained in each of the k packets from the receiving buffer, and obtaining the packet group G based on the following conditions: a correction step of correcting for calculating the reception side sampling frequency fr in _n,
Recording the received playback program order to execute the computer-readable recording medium.
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the receiving-side sampling frequency in the packet group G _(n−1) . A broadcasting system,
Voice input means;
A voice transmission unit connected to the voice input unit, which converts the input voice into digital voice data by sampling based on the transmission-side sampling frequency fs, and sequentially outputs packets including the digital voice data to the transmission path; ,
Voice receiving means connected to the voice transmitting means via the transmission line, and converting the digital voice data into analog voice data based on a receiving-side sampling frequency fr;
  Voice output means connected to the voice receiving means and outputting the analog voice data;
The voice receiving means is
Receiving means for sequentially receiving a plurality of packets from the transmitting device;
A reception buffer for temporarily storing digital data included in each of the plurality of packets;
A frequency generating means for generating a receiving side sampling frequency fr;
Conversion means for reading out digital data stored in the reception buffer according to the reception-side sampling frequency fr and converting it into analog data;
Packet group G composed of a predetermined number k (k ≧ 2) packets determined on the receiving side _n For each (n is a natural number), the reception means ΔTr indicating the time required for the reception means to receive k packets and the digital data contained in each of the k packets are read from the reception buffer. And a reading interval ΔTo indicating the time required for the packet group G based on the following conditions: _n Correction means for calculating the receiving-side sampling frequency fr in
Broadcast system equipped with.
fr = fr ′ × (ΔTo ÷ ΔTr)
Here, fr ′ is the packet group G _(N-1) Receiver sampling frequency at.

Images