KR100457751B1 - SPEECH CODEC MODE ASSIGNMENT METHOD FOR VoIP(VOICE OVER IP) WITH AMR SPEECH CODEC - Google Patents

Abstract

The present invention relates to an Internet telephony system. In particular, in order to reduce packet loss and improve call quality by applying a voice codec having a variable bit rate to obtain good sound quality in the operation of an Internet telephony service, Internet telephony using a voice codec with variable transmission rate to apply an AMR voice codec with a transmission mode to an Internet telephony system (VoIP), and to monitor the network status and determine the mode of the AMR voice codec accordingly. The purpose of the present invention is to provide a voice codec mode allocation method in a system. To this end, the present invention provides a method for allocating a voice codec mode in an Internet telephony system to which a voice codec having a variable bit rate is applied, by defining jitter, which is a channel monitoring parameter, dividing the range of the parameter into each step. A first step of allocating a voice codec mode and calculating a jitter value in the received packet; A second step of selecting a voice codec mode allocated to a step corresponding to the jitter value among the voice codec modes allocated to each step; And a third step of transmitting a packet to a sender using the selected voice codec mode.

Description

SPECCH CODEC MODE ASSIGNMENT METHOD FOR VoIP (VOICE OVER IP) WITH AMR SPEECH CODEC

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an internet telephone system, and more particularly, to a method for allocating a codec mode in an internet telephone system using a voice codec having a variable transmission rate.

The basic premise of the Voice over IP (VoIP) service in the Internet Protocol (IP) network (hereinafter simply referred to as "IP") is that the voice signal is converted into a packet form and included in the IP packet. It can be routed between devices with functionality. These packets reach the destination via a suitable routing path and can be reproduced with fairly good sound quality.

The VoIP communication method can be classified into three types such as PC-to-PC, PC-to-phone, and phone-to-phone, and various technical elements are required to provide services.

Techniques for providing VoIP services include user terminal technology including voice / digital signal conversion and compression technology, user terminal information processing technology, server technology related to communication setting and control, and mutual interoperability and compatibility between terminals. Core technologies include gateway technology, call channel setup and management technology, IP address and phone number mapping technology. These technologies have been developed by several standardization bodies or consortiums and selected as standards.

To date, the development of Internet telephony standards has been based on ITU-T's H.323, which has been completed by active contributions from ETSI and IMTC organizations.

H.323-based Internet telephony using RTP / RTCP protocols is becoming inefficient in combination with several existing ITU-T standards, and Internet Protocol-oriented standards are being developed at the IETF. Internet telephony standard based on Initiation Protocol. Until now, ITU-T's H.323-based standards and product development have been mainstream, but in the future, active competition with SIP-based VoIP technology is expected.

H.323, on the other hand, is one of ITU-T's H.32x series of recommendations for providing multimedia communication services in various types of networks, and is the basic technology for real-time transmission of audio, video, and data in packet-based networks. to be. H.323 terminals used for real-time two-way communication include audio / video compression protocols for compressing multimedia signals, protocols for call setup and control, real-time transmission and control protocols, protocol specification configuration, user equipment interface, and system. The control and linkage with the LAN are shown in Figs. 1A and 1B, respectively.

That is, FIG. 1A is a configuration diagram of an embodiment of a general H.323 terminal protocol, and FIG. 1B is a configuration diagram of an embodiment of a general H.323 terminal.

Since the basic service provided by the H.323 terminal in the Internet telephony system is voice communication, at least one of the voice codecs supported by H.323 shown in [Table 1] must be supported for efficient information transmission.

ITU-T standard Coding algorithm Transmission speed G.711 PCM 64 kbits / s G.722 SB-ADPCM 48,56,64 kbits / s G.723.1 MP-MLQ / ACELP 5.3,6.3 kbits / s G.728 LD-CELP 16 kbits / s G.729 CS-ACELP 8 kbits / s

On the other hand, the proposal on the VoIP technology is mainly focused on the development of large equipment for network traffic processing or the Internet protocol, and there is no clear proposal for the voice signal processing technology in the terminal for improving the voice quality. Therefore, the development of technology for providing high quality voice service has been relatively neglected, and the operators are facing difficulties due to the user's dissatisfaction with the call quality.

The biggest deterioration of voice quality of service in VoIP is the loss of transmission of voice packets, delay of transmission of voice data, delay of transmission of voice data, end-to-end echo phenomenon, and load of voice codec. The barrier is the biggest cause.

In particular, as listed in Table 1, all voice codecs of VoIP systems currently supported have a fixed transmission rate, and thus have a problem in that they cannot flexibly cope with packet loss in a network with high traffic fluctuations.

In addition, the sound quality of the existing voice codec as described above has a problem that the sound quality of the AMR voice codec having a variable transmission rate recently selected as the standard of 3GPP.

In other words, the voice codec in the currently used Internet telephone system uses a voice codec having a fixed transmission rate, which not only can not flexibly cope with packet loss in a high-traffic Internet network, but also has a poor sound quality. In addition, there is a problem that the sound quality of the newly developed AMR voice codec is lower than, and thus there is a problem that can not provide high-quality Internet phone service to users.

The present invention is to solve the above-described problems, in the operation of the Internet telephony service, by applying a voice codec having a variable bit rate that can obtain a good sound quality to reduce packet loss and improve the call quality In order to apply an AMR voice codec having eight transmission modes to an Internet telephone system (VoIP), a voice having a variable transmission rate for monitoring the state of the network and thus determining the mode of the AMR voice codec accordingly An object of the present invention is to provide a voice codec mode allocation method in an Internet telephone system using a codec.

1A is a diagram illustrating an embodiment of a general H.323 terminal-side protocol.

1B is a block diagram of an embodiment of a general H.323 terminal.

2 is a configuration diagram of an embodiment of an Internet telephone system to which the present invention is applied.

3 is a diagram illustrating an embodiment of an Internet telephone (VoIP) terminal to which an AMR voice codec applied to the present invention is applied.

4 is an exemplary view illustrating a voice codec mode allocation method using Jit _i parameters applied to the present invention.

5 is an exemplary diagram of an adaptive codec mode allocation algorithm applied to the present invention.

6A to 6C are diagrams illustrating an example of a voice codec mode allocation method in an Internet telephone system to which a voice codec having a variable data rate is applied according to the present invention.

7A and 7B show an embodiment execution screen of a PC program for experimenting with the present invention.

8 is an example of an experimental environment for experimenting with voice transmission over the Internet with respect to the present invention.

9A to 9F are exemplary diagrams showing the results of experiments with respect to the present invention.

In order to achieve the above object, the present invention provides a voice codec mode allocation method in an Internet telephone system to which a voice codec having a variable transmission rate is applied. Dividing, assigning a voice codec mode to each step, and calculating a jitter value in the received packet; A second step of selecting a voice codec mode allocated to a step corresponding to the jitter value among the voice codec modes allocated to each step; And a third step of transmitting a packet to a sender using the selected voice codec mode.

In addition, the present invention, in the voice codec mode allocation method in the Internet telephone system to which the voice codec having a variable data rate is applied, by defining the channel monitoring parameters Jitter (Jitter) and CumLoss to divide the range of the parameters in each step, A first step of defining a voice codec mode table by allocating a voice codec mode for each step, and receiving a packet transmitted from a transmitter; A second step of modifying the voice codec mode table defined in the first step by calculating jitter and CumLoss in the received packet and modifying the range of each step; A third step of selecting a voice codec mode assigned to a step corresponding to the calculated jitter value in a modified voice codec mode table; And a fourth step of transmitting a packet to a transmitter using the selected voice codec mode.

In addition, the present invention, in the voice codec mode allocation method in the Internet telephone system to which the voice codec having a variable data rate is applied, by defining the channel monitoring parameters Jitter (Jitter) and CumLoss to divide the range of the parameters in each step, A first step of defining a voice codec mode table by allocating a voice codec mode for each step, and receiving a packet transmitted from a transmitter; A second step of modifying the voice codec mode table defined in the first step by calculating jitter and CumLoss in the received packet and modifying the range of each step; A third step of selecting a voice codec mode assigned to a step corresponding to the calculated jitter value in a modified voice codec mode table; And a fourth step of transmitting the packet to the transmitter using the selected voice codec mode, and transmitting information on the voice codec mode selected in the third step in the packet.

That is, the present invention is to solve the problem of improving the call quality, which is difficult to overcome by a voice codec having a fixed transmission rate in a conventional Internet Telephony (VoIP) system as described above, a variable that can obtain a good sound quality A voice codec (AMR voice codec) having a transmission rate (hereinafter, simply referred to as an "AMR voice codec") is applied to an Internet telephone system in which a network is changed, thereby reducing packet loss and improving call quality.

To this end, first, AMR voice codec with 8 transmission modes is applied to the Internet phone system, define the parameters that can monitor the channel status of the Internet network according to traffic fluctuations, and then monitor the network status using the parameters. We propose a method of determining the mode of the voice codec accordingly.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

2 is a configuration diagram of an embodiment of an Internet telephone system to which the present invention is applied.

As shown in the figure, the Internet telephone system to which the present invention is applied uses an Internet telephone (VoIP) terminal (hereinafter simply referred to as "Internet telephone") to which an AMR voice codec is applied, and an Internet protocol. It consists of an internet network (network) 30, and consists of a router (20, 20 ') for connecting the Internet phone and the network.

That is, the present invention uses the terminal (10, 10 ') to which the AMR voice codec is applied, while the state of the Internet network 30 using a parameter defined to monitor the channel state of the Internet network 30 according to the traffic change We will propose a method that can improve the call quality of the Internet phone by monitoring and determining the mode of the voice codec accordingly.

3 is a configuration diagram of an embodiment of an Internet telephony (VoIP) terminal to which an AMR voice codec applied to the present invention is applied, and FIG. 4 is an example of a codec mode allocation method using Jit _i parameters applied to the present invention. 5 is an exemplary diagram of an adaptive codec mode allocation algorithm applied to the present invention.

That is, the Internet phone terminal shown in FIG. 3 is identical in structure to the Internet phone terminal shown in FIG. 1B, but has an additional feature of using an AMR voice codec in addition to the voice codec used in the existing Internet phone terminal. Therefore, it can be seen that the function of the system controller is different from the existing terminal.

On the other hand, the main parameters affecting the voice quality of VoIP systems such as packet loss, jitter, and transmission delay are due to the lack of network bandwidth, mainly through queuing or routing at each transport node in the packet network. Will occur. In this case, in order to ensure good quality of service (QoS) even in a bad network state, it is desirable to adjust the size of a transmitted packet according to the network state. Adjusting the size of a packet changes the transmission rate of a voice codec in units of packets. By adaptively adjusting the packet size according to the channel condition, it is possible to reduce packet loss and improve sound quality.

Voice codecs currently supported by the H.323 protocol for VoIP systems include G.711 (PCM), G.723.1 (MP-MLQ ACELP), G.729 (CS-ACELP), and GSM-FR. As described above, these voice codecs have a fixed transmission rate and thus cannot adapt to a rapidly changing channel condition.

AMR voice codecs, on the other hand, are voice codecs with variable rates that can vary from 12.2 kbits / s to 4.75 kbits / s in every analysis frame. Since the AMR voice codec is adopted as a standard for IMT-2000, when the IMT-2000 wireless network and the IP network's voice traffic are required to be interworked, the transcoding process performed at the gateway between the two networks can be eliminated. have. This reduces overall transmission delay and complexity in the switching network, thus improving overall sound quality. Accordingly, the present invention intends to use a new VoIP terminal to which the AMR voice codec is applied as shown in FIG.

Meanwhile, in order to select one of eight codec modes of an AMR voice codec on a packet basis, a channel state must be estimated. For this, timestamp information included in an RTP header is used.

In addition, to determine the transmission rate of the AMR voice codec to be applied to the VoIP system, the jitter parameter defined in Equation 1 below is used as a measurement parameter for monitoring channel conditions.

Jit _i = | (R _i -S _i )-(R _i-1 -S _i-1 ) |

In this case, R _i and S _i represent the arrival time of the i-th packet and the timestamp included in the RTP header, respectively, and by calculating [Equation 1] at the receiving side, interarrival jitter of the packet data, that is, channel state Can be estimated.

Assuming that the initial value of the de-jitter buffer size is set to 40 msec, the range of Jit_i can be linearly divided into eight states as shown in [Table 2]. The interarrival jitter of the received packet calculated by [Equation 1] and the AMR codec mode of the transmitted packet are allocated using [Table 2].

These processes are performed for every transport packet. For example, when Jit _i is 4 msec, the transmitted packet is assigned an AMR codec mode of 12.2 kbits / s, and when Jit _i is 14 msec, 7.95 kbits / s is allocated and transmitted. The process for allocation is shown in FIG. Meanwhile, Table 2 below shows an initial state for AMR voice codec mode allocation.

Range of Jit _i (msec) AMR codec mode 0 (range [1]) <Jit _i <5 (range [2]) 12.2 kbits / s 5 (range [2]) <Jit _i <10 (range [3]) 10.2 kbits / s 10 (range [3]) <Jit _i <15 (range [4]) 7.95 kbits / s 15 (range [4]) <Jit _i <20 (range [5]) 7.40 kbits / s 20 (range [5]) <Jit _i <25 (range [6]) 6.70 kbits / s 25 (range [6]) <Jit _i <30 (range [7]) 5.90 kbits / s 30 (range [7]) <Jit _i <40 (range [8]) 5.15 kbits / s 40 (range [8]) <Jit _i 4.75 kbits / s

That is, like the voice codec mode allocation method described above, a VoIP system to which a voice codec mode allocation method having a fixed Jit _i state range is defined as "AMR_1" as given in [Table 2]. This will be described with reference to the flowchart shown in FIG. 6A.

At this time, if the Jit _i range of each state for the codec mode allocation shown in [Table 2] is fixed, packet loss may increase when a bad channel state lasts for a long time. In addition, when the network conditions are very good, the allocated codec modes can be concentrated around the high data rate, which can increase the load on the network.

In order to consider these cases, it is reasonable in the long term to adaptively change the Jit _i range of each state according to the network state. To this end, the present invention proposes an adaptive codec mode allocation method for applying an AMR voice codec to VoIP in order to more efficiently adapt to channel conditions.

That is, to monitor channel conditions in the long term, two parameters are used.

First, use the "cumulativenumber of packets lost" parameter included in the header of the RTCP packet transmitted every 5 seconds, and also use the codec modes assigned during the period in which the RTCP is transmitted. Use distribution characteristics as parameters.

In this case, by using the parameters described above, the Jit _i range of each state is adaptively changed according to the network state, thereby implementing a voice codec mode allocation method in an Internet telephone system to which a voice codec having a variable bit rate according to the present invention is applied. There are two ways to do this. That is, the proposed method is named "reverse channel state monitoring method (AMR_2)" and "forward channel state monitoring method (AMR_3)", respectively. The proposed codec mode allocation algorithm proposed to apply the AMR voice codec is shown in FIG. Shown in Meanwhile, the two methods will be described in detail with reference to FIGS. 6B and 6C.

First, the following describes a specific method of modifying the Jit _i range (hereinafter, simply referred to as "state range") of each state and the advantages resulting from modifying the state range.

The specific method to modify the status range is as follows.

That is, CumLoss is obtained by extracting the number of packets lost for 5 seconds from the RTCP header information. If the currently measured CumLoss is larger than the previously measured CumLoss, you can clearly think that the channel is in a bad state, and adjust the range values in [Table 2] used to determine the voice codec mode based on this. Eight range values are needed to determine the voice codec mode, and each range value is increased or decreased by 2 with reference to the CumLoss comparison result. Algorithms for implementing the above process are shown in Table 3 below.

FOR i = 1 to 8IF CumLoss _j > CumLoss _j-1 range [i] + = step ELSE IF CumLoss _j <CumLoss _j-1 range [i]-= stepENDwhere, step = 2 Cumuls: Cumulative number of lost packets j: count of RTCP packets

The second way to change the range is to refer to the codec mode distribution used during the same five seconds. In other words, it analyzes how many eight modes are allocated in 5 seconds and finds the most allocated mode (M). The value will be a value between 1 (original number mode with the highest bit rate -12.2) to 8 (original number mode with the lowest bit rate -4.75) and compared with the lower range value needed to determine the codec mode. To change the range. An algorithm for implementing the above process is shown in Table 4 below.

Calculate M _max FOR k = 1 to 8IF k> M _max range [k]-= step ELSE IF k <M _max range [k] + = stepENDM _max : the codec mode which have been assignedmost frequently during a RTCP intervalk: number assigned to each codec mode, k = 1 (12.2 kbps) to k = 8 (4.75 kbps)

On the other hand, the benefits of modifying the state range are as follows.

When determining the bit rate, if you use a fixed range value, only 4.75 kbps mode will be assigned forever at jitter values larger than the initial maximum range value of 40 ms. In other words, the purpose of this study is to adjust the bit rate of the voice encoder according to the variable channel condition. If the fixed range value is used, it may be wrong to get the original research purpose. For example, consider a picture like Table 5 below.

In other words, assuming that the jitter value is in the range of 50 to 110, the fixed bit rate is used regardless of the channel change characteristic (when the jitter varies between 50 and 110) using the initial fixed range. (4.75) will be assigned, but if the range value is changed depending on the channel state, eight bit rates can be allocated evenly over a larger range than the initial value. There is an advantage to use.

FIG. 6A is a flowchart illustrating a method for allocating a voice codec mode in an Internet telephony system to which a voice codec having a variable data rate is applied according to the present invention, and using a voice codec mode having a fixed Jit _i state range.

First, an Internet telephone terminal (hereinafter referred to simply as a "receiver terminal") 10 'on the receiving side is connected to an Internet telephone terminal (hereinafter simply referred to as a "calling terminal") 10 on the calling side through the Internet network. The i-th transmitted packet is received (602).

The controller of the receiving terminal 10 ′ calculates a Jit _i value as described above in order to estimate the state of the channel through the received packet (604).

A voice codec mode corresponding to the calculated Jit _i value is selected from the AMR codec modes shown in [Table 2] (606).

Using the selected voice codec mode, the receiving terminal 10 'transmits a packet to the transmitting terminal 10 (608).

FIG. 6B is a flowchart illustrating a method for allocating a voice codec mode in an Internet telephony system to which a voice codec having a variable bit rate according to the present invention is applied and shows a method for allocating a voice codec mode using a reverse channel state monitoring method. In this case, the reverse channel state monitoring method is a method of adaptively allocating a voice codec mode by monitoring the channel state at the receiving side.

First, the receiving terminal 10 'receives the i th packet transmitted from the calling terminal 10 through the Internet network (642).

The control unit of the receiving terminal 10 'changes the codec mode table as a whole by adjusting each state range of the codec modes shown in [Table 2] using each parameter (Jit _i , CumLoss) in the received packet ( 644).

That is, the 'cumulative number of packets lost (CumLoss)' parameter in the header of the RTCP packet transmitted every 5 seconds indicates the total number of packets lost in the RTP data packets transmitted from the sender since the call started. Adjust the range of each state of the codec mode given in Table 2 according to the CumLoss parameter received at each packet interval.

If the value of the CumLoss parameter of the current RTCP packet is higher than the value of the previous RTCP packet, the channel state is regarded as bad in the long term and the status range of each codec mode is increased. In the opposite case, it is assumed that the channel state is good and the respective state range is reduced.

The distribution characteristic of the assigned codec mode is used to prevent the AMR codec mode from rushing toward the highest or lowest codec mode. To this end, we define the codec mode, Mmax, which is the most allocated during one RTCP packet interval, and based on this codec mode, the status range of the lower codec mode is increased by the step interval, and the status of the higher codec mode is increased. The range is reduced by each step interval so that the distribution of assigned codec modes can be spread evenly across all codec modes.

In operation 644, the voice codec mode corresponding to the Jit _i value calculated through the received packet is selected from the AMR codec mode table changed in operation 646.

Using the selected voice codec mode, the receiving terminal 10 'transmits a packet to the transmitting terminal 10 (648).

Hereinafter, a voice codec mode allocation method using the reverse channel state monitoring method will be described as a specific example. In this case, it is assumed that the value of the current CumLoss parameter is larger than the previous value and the received Jit _i is 17, for example.

That is, the increase in the ComLoss confirmed through the received packet means that the state of the channel is worse, and thus the state range of the voice codec is changed as shown in Table 6 below (642 and 644). At this time, it is assumed that the state range values before the change are the same as the initial values shown in [Table 2].

Before change after change AMR codec mode 0 (range [1]) <Jit _i <5 (range [2]) 2 (range [1]) <Jit _i <7 (range [2]) 12.2 kbits / s 5 (range [2]) <Jit _i <10 (range [3]) 7 (range [2]) <Jiti <12 (range [3]) 10.2 kbits / s 10 (range [3]) <Jit _i <15 (range [4]) 12 (range [3]) <Jit _i <17 (range [4]) 7.95 kbits / s 15 (range [4]) <Jit _i <20 (range [5]) 17 (range [4]) <Jit _i <22 (range [5]) 7.40 kbits / s 20 (range [5]) <Jit _i <25 (range [6]) 20 (range [5]) <Jit _i <25 (range [6]) 6.70 kbits / s 25 (range [6]) <Jit _i <30 (range [7]) 27 (range [6]) <Jit _i <32 (range [7]) 5.90 kbits / s 30 (range [7]) <Jit _i <40 (range [8]) 32 (range [7]) <Jit _i <42 (range [8]) 5.15 kbits / s 40 (range [8]) <Jit _i 42 (range [8]) <Jit _i 4.75 kbits / s

In addition, since the received Jit _i value is 17, the receiving terminal 10 'selects a voice codec mode of 7.40 kbits / s from the changed voice codec mode table of Table 6 to the transmitting terminal 10 (646). The packet will be sent.

On the other hand, the channel state estimated by Equation (1) actually represents the quality of data transmitted from the transmitting side to the receiving side, and it can be said that the channel state is estimated from the transmitting side perspective. That is, under the assumption that the channel states of the forward transmission path and the reverse transmission path are similar to each other, the estimated channel state information is used to define the codec mode of the packet transmitted from the receiving side. Therefore, this is defined as a reverse channel state monitoring method.

FIG. 6C is a flowchart illustrating a method for allocating a voice codec mode in an Internet telephony system to which a voice codec having a variable bit rate is applied according to the present invention. FIG. . At this time, in the forward channel state monitoring method, the receiving terminal adaptively allocates the voice codec mode according to the channel state from the transmitting side to the receiving side, and then retransmits the allocated voice codec mode information to the transmitting side to perform another transmission side. In the case of transmitting voice data, the method can be transmitted through the assigned voice codec mode, and can be used when the assumption that the channel states of the forward and reverse paths are similar to each other is not established.

That is, when the assumption that the channel states of the forward and reverse paths are similar to each other does not hold, the jitter information obtained by Equation 1 at the receiving end is used to determine the codec mode of the packet transmitted at the transmitting terminal. It is much more reasonable.

To this end, a method of returning the jitter information obtained from the receiving side to the transmitting side should be devised. In this case, it is much more advantageous to determine and send the codec mode to be used when transmitting the next packet rather than returning the jitter information itself. .

In other words, the channel state obtained from the receiving side, that is, the codec mode using jitter information is returned to the transmitting side by the forward channel state monitoring method, and the transmitting side utilizes this value as the codec mode of the next packet to be transmitted. Therefore, the 'payloadtype' parameter included in the RTP header is used to return the codec mode information allocated by the receiver to the transmitter according to the channel state from the transmitter to the receiver.

First, the receiving terminal 10 ′ receives the i th packet transmitted from the calling terminal 10 through the Internet network (662).

The control unit of the receiving terminal 10 'adjusts each state range of the codec mode shown in [Table 2] by using each parameter (Jiti, CumLoss) in the received packet, thereby making the overall codec mode table [Table 6]. In operation 664, the voice codec mode corresponding to the Jiti value calculated through the received packet is selected from the changed codec mode table (666), and then the receiving terminal 10 'transmits using the selected voice codec mode. The packet is transmitted to the side terminal 10 (668).

In this case, the processes 662 to 668 may be the same as the forward channel state monitoring method described with reference to FIG. 6B. In the reverse channel state monitoring method, the receiving terminal 10 ′ uses the selected voice codec mode. When the packet is transmitted to (10), the method further includes the step of transmitting information on the selected voice codec mode in a transport packet and transmitting the packet together (670).

Meanwhile, in the present invention, in order to return the assigned voice codec mode to the sender, 35 to 42 of the 'payloadtype' parameters which are not used yet are used, and the payload type allocation table used in each mode is shown in [Table 7]. . In this case, the payloadtype number shown in [Table 7] refers to header information indicating what kind of audio codec was used in packet formation in actual packet communication.

Payloadtype Codec mode 35 AMR_12.2 36 AMR_10.2 37 AMR_7.95 38 AMR_7.4 39 AMR_6.7 40 AMR_5.95 41 AMR_5.15 42 AMR_4.75

That is, the voice codec mode allocation method using the forward channel state monitoring method is not a matter of how to determine the bit rate of the packet to be sent, but actively requests the transmitter to receive the bit rate that the user wants to receive based on the analyzed jitter value. Way. Therefore, regardless of the bit rate to send, it is required to use the unused payload type number. Will be written.

Hereinafter, a voice codec mode allocation method using the forward channel state monitoring method will be described as a specific example. At this time, assuming that the value of the current CumLoss parameter is larger than the previous value and the received Jit is 17, the example will be described.

In this case, as in the example illustrated in FIG. 6B, the receiving terminal 10 'selects the bit rate of the packet to be transmitted as 7.4 kbps (662 to 668), and transmits the bit rate that the receiving terminal 10 wants to receive. In step 670, the active request is added. At this time, by transmitting payload type 38 as shown in [Table 7], the receiving terminal 10 'side actively requests the transmitting terminal 10 for its desired bit rate.

In other words, if the sender keeps receiving packets in a good environment with no packet loss, the receiver only knows the information from the packet it receives, so the receiver does not know that the bit rate of 7.4 kbps is preferred, and the packet has a high bit rate such as 12.2 kbps. Since there is a possibility of sending a packet, the receiving end actively requests the transmitting end for its desired bit rate.

As described above, the reverse channel state monitoring method is a method in which a receiving terminal adaptively allocates a voice codec mode according to a channel state from a transmitting side to a receiving side and transmits voice data to the transmitting side through the corresponding voice codec mode. This may be used when the channel states of the forward and reverse paths are similar to each other.

On the other hand, it will be described below with reference to the experimental content applying the above method.

That is, in order to evaluate the performance of the method according to the present invention described above, the AMR voice codec distributed by 3GPP was applied to the H.323 protocol to implement a VoIP terminal and then mounted on a PC. In this case, the implemented VoIP terminal utilizes a source open for the public by Equivalnce Pty Ltd., and execution screens of the PC program are as shown in FIGS. 7A and 7B. That is, FIGS. 7A and 7B show an execution screen of an embodiment of a PC program for experimenting with the present invention. FIG. 7A shows an execution screen and FIG. 7B shows a screen of an audio codec selection process among the execution screens.

In addition, to evaluate the performance of the proposed adaptive voice codec mode allocation methods, several voice transmission experiments were performed in full-duplex in a public Internet network. In the experimental environment, as shown in FIG. 8, VoIP terminals applied with respective voice codecs in two different places were mounted on several PCs, and then transmission experiments were simultaneously performed. That is, Figure 8 is an example of an experimental environment for experimenting with the voice transmission over the Internet with respect to the present invention.

In this experiment, G.723.1 and GSM-FR voice codec were used among several voice codecs adopted in the existing VoIP system for performance comparison with VoIP system with AMR voice codec. This is because the highest bit rate mode of the AMR voice codec is similar to the bit rate of GSM-FR, and G.723.1 voice codec is one of the most widely used VoIP systems.

In order to evaluate the performance of the proposed algorithm, the subjective sound quality of the average packet loss rate, the interarrival jitter, and the received voice was evaluated in order to evaluate the performance of the proposed algorithm. The results are shown in FIGS. 9A to 9F. 9A and 9B show average packet loss and average jitter of the first experiment, and FIGS. 9C and 9D show average packet loss and average jitter of the second experiment, and FIGS. 9E and 9F illustrate the third experiment. Average packet loss and average jitter are shown.

That is, the system using the method according to the present invention has a lower average packet loss rate and a smaller jitter value than the conventional audio codec through the diagrams shown in FIGS. 9A to 9F. can see.

On the other hand, the development technology proposed in this patent can be applied not only to the Internet telephone system but also to a communication system in which the characteristics of the transmission channel are not fixed.

That is, the present invention is not limited to the above-described embodiments, and various modifications and changes can be made by those skilled in the art, which are included in the spirit and scope of the present invention as defined in the appended claims.

The present invention has an excellent effect of being able to be less affected by fluid channel characteristics by adaptively adjusting the size of a packet transmitted from a terminal according to the characteristics of a channel in a packet communication environment.

Therefore, there is an excellent effect that can improve the call quality at a low cost in the exploding Internet telephone system.

Claims (5)

A voice codec mode allocation method in an internet telephone system using a voice codec having a variable bit rate, A first step of defining jitter, which is a channel monitoring parameter, dividing the range of the parameter into each step, assigning a voice codec mode to each step, and calculating a jitter value in the received packet; A second step of selecting a voice codec mode allocated to a step corresponding to the jitter value among the voice codec modes allocated to each step; And A third step of transmitting a packet to a sender using the selected voice codec mode Voice codec mode allocation method in an internet telephone system to which a voice codec having a variable bit rate including a transmission rate is applied. A voice codec mode allocation method in an internet telephone system using a voice codec having a variable bit rate, By defining jitter and CumLoss, which are channel monitoring parameters, the range of the parameter is divided into stages, the voice codec mode is defined by assigning the voice codec mode to each stage, and the packet transmitted from the transmitter is received. First step; A second step of modifying the voice codec mode table defined in the first step by calculating jitter and CumLoss in the received packet and modifying the range of each step; A third step of selecting a voice codec mode assigned to a step corresponding to the calculated jitter value in a modified voice codec mode table; And A fourth step of transmitting a packet to a sender using the selected voice codec mode Voice codec mode allocation method in an internet telephone system to which a voice codec having a variable bit rate including a transmission rate is applied. A voice codec mode allocation method in an internet telephone system using a voice codec having a variable bit rate, By defining jitter and CumLoss, which are channel monitoring parameters, the range of the parameter is divided into steps, and the voice codec mode table is defined by allocating the voice codec mode for each step. First step; A second step of modifying the voice codec mode table defined in the first step by calculating jitter and CumLoss in the received packet and modifying the range of each step; A third step of selecting a voice codec mode assigned to a step corresponding to the calculated jitter value in a modified voice codec mode table; And A fourth step of transmitting the packet to the sender using the selected voice codec mode, and transmitting information on the voice codec mode selected in the third step together with the packet; Voice codec mode allocation method in an internet telephone system to which a voice codec having a variable bit rate including a transmission rate is applied. The method of claim 3, wherein The voice codec mode allocation method of claim 4, wherein the information on the voice codec mode uses payloadtype parameters included in the RTP header. The method according to any one of claims 1 to 4, The voice codec is a voice codec mode allocation method in an Internet telephone system to which a voice codec having a variable transmission rate is used, characterized by using an AMR voice codec having eight transmission modes.