Multimedia conference system based on H.323 protocol - excerpt notes
Ryan Liu
2002/3/14
Multimedia conference system based on group exchange network (PBN) is developed by ITU-T;
These packet networks include LAN / WAN, Internet, intranet, and packet protocols using PPP via PSTN (public switched telephone network) or ISDN dial-up connection or point-to-point connection.
Desktop Video Conference Based on IP Networks, it is a video technology and IP data communication technology to establish visual communication between two or more locations through the IP network (Internet, Intranet, LAN / WAN). A conference form of voice and data exchange.
H.323 Structure of Conference System
Transferred information: audio, video, data, and control information.
All information flows are packaged and transmitted in H.225.0.
H.323 Protocol defined multimedia conferencing system:
It is mainly composed of terminal, gatekeeper, also known as gatekeeping, gateway, multipoint controller (MC), multipoint processor (MP), and multipoint control unit (MCU), and the like.
H.323 terminal
In the use, the audio algorithm used by the codec is negotiated by using H.245 during the capacity exchange. The audio flow should be formatted according to the H.255.0 standard. The H.323 terminal can simultaneously send or receive multiple audio channel information.
Video coding does not perform BCH error correction, and allows for operation with asymmetrical video bit rate, frame rate, image resolution.
H.245 Control the end-to-end control message of the H.323 communication entity operation, including capabilities, logical channels, mode selection requests, traffic control messages, and general commands and instructions.
H.225 Call Control: Use H.225.0 Call Control Signaling to establish two H.323 terminals or terminals between terminals and gatekeepers. The Call Signaling Channel is built first in the H.245 control channel between the H.323 terminal, and therefore, its establishment is not managed by the H.245 control channel.
RAS Control: RAS (Registration, Admission and Status, registration, acceptance, and status protocol) signals use H.225.0 control messages to perform registration, accept, bandwidth changes between terminals and gatekeepers, and processes of the two processes.
The network interface of the H.323 terminal is described in H.225, which specifies the following functionality: 1. Provide a reliable end-to-end service to the H.245 control channel, data channel, call signaling channel (TCP , SPX, etc.), for the audio, video, and RAS channels provide unreliable end-to-end services (UDP, IPX, etc.). These services can be single work, duplex, unicast or multicast.
Neighborhood (net gate)
H.323 Neighborhood must provide 4 basic services in the system:
Address translation
Bandwidth control
License control
District management function.
An optional feature of the gatekeeper:
Bandwidth management, call authentication, call control signaling and call management, etc.
Gateway
The purpose of the typical gateway is to map the characteristics of the packet network terminal to the circuit switched network terminal or opposite. The main application of the gateway is to establish a connection to the remote H.320 compatible terminal through N-ISDN; establish a connection to the remote H.321 compatible terminal through B-ISDN or establish and remote H.324 and remote H.324 and remote H.324 and remote H.324. V.70 compatible terminal connection and the like.
Multipoint Control Unit (MC)
The MC does not directly process the media information flow; MP is mixed, switched, and other processing on audio, video or data information.
Image code
The image coding in the H.323 conference system is mainly: That is, H.261 and H.263. It is H.261 QCIF is a must format, in addition to this, can be used to use other coding forms through capacity negotiation.
H.261 It is recommended to use three technologies in the coding to implement compression: • Predictive coding · Transform domain coding · Retroptrocoding
Like H.261, H.263 uses motion compensation and DCT encoding methods, but it refers to the MPEG standards to introduce three frame modes and interfaces, Intra (inter-frame encoding), intra reference to the MPEG standard. Frame encoding) Two coding modes.
Audio (voice) encoding
The audio (speech) coding in the H.323 conference system has six: G.711, G.722, G.723.1, G.728, G.729 and MPEG Audio. Where G.711 is a must, others are optional. In addition to the above six coding methods, other methods can also be used through capacity to negotiate. At present, the H.323 conference system is mainly used in two types of speech methods.
G.723.1 is a two-rate speech encoder with two coding rates of 6.3k and 5.3k, respectively. High speed (6.3k) uses multi-pulse excitation maximum likelihood quantization (MP_MLQ) algorithm, low rate (5.3k) using an algebra, a linear prediction (ACELP) algorithm. These two algorithms have the same theoretical basis, which are based on linear prediction (LPC), which uses a non-periodic component excitation source. The difference is that MP_MLQ uses multi-pulse maximum likelihood quantitative incentives, and the ACELP is an algebra.
G.723.1 The voice quality is better.
5.3K rate coding, the voice quality is superior to VCELP (8kb / s);
Its 6.3k rate coding, the speech quality is equivalent to G.726 of 32kb / s suggests the corresponding indicator.
Both are basically able to meet the requirements of long-distance telephone quality.
The disadvantage of G.723.1 is that there is a large delay.
G.729 is a 8kb / s speech coding standard, which uses the algorithm that the conjugate structure is digitally excited, linear predictive coding (CS_ACELP), can reach 32kb / sadpcm voice quality.
The CS_ACELP algorithm is very characteristic:
After nothing big, it can reach 6.4kb / s down, up to 13KB / s up to 13KB / s, and can transmit better voice quality.
data communication
Data communication in the conference system is a communication between multiple participants, so it is a communication architecture for establishing a multipoint communication service (MCS). Data communication is high for error control, and it is not sensitive to time delay.
Data communication uses T.120 series recommendations, divided into four levels to complete data communication in the conference system:
The first level is the next four-layer communication protocol stack, which is specified in T.123. The second level is a multi-point communication service (MCS), which is specified in T.122 / T.125. The third level is a general conference control (GCC), which is specified in T.124. The fourth level is the application level, which is specified in T.121, T.126, T.127 and T.128. The binary files specified in the electronic whiteboard and T.127 specified in T.126 have been applied.
Code stream
The H.323 conference system is based on packet exchange, so the code stream in the conference system must be packaged before transmission, and statistical multiplexing according to the tag on the packet.
Audio and video code flow are high at real-time requirements, so as to avoid delayed problems. However, it is less sensitive to small amounts of packet loss. Therefore, for the audio and video stream, real-time transmission protocol RTP is used to packed them and then use the UDP protocol for unconnected UDP protocols. The RAS signal is also transmitted using the UDP protocol. Data and control signals are high for service quality, and small amounts of packages or errors are unbearable. Therefore, for the data and control code stream, the TCP protocol for the transport layer is used to complete the TCP protocol, thereby completing reliable transmission of them. H.323 system in system code stream represented as shown below:
QoS (service quality) guarantee
Multicast technology can effectively solve the network bandwidth problem of multi-point data communication. During transmission, the tree path of data transfer is determined and optimized and optimized according to the network topology distribution of the transmitting and receiving the parties, and the data stream of the same content is only transferred once. The multicast address can use a universal multicast address, but this is not a valid communication method, preferably, can dynamically allocate a set of addresses.
Resource reservation refers to bandwidth resource allocation based on the QoS requirements of business data, and provides a complete path on the IP online. IETF RFC2205 Resource Reserved Agreement (RSVP) is an agreement to provide this management mechanism.
On the unconnected network, it increases the connection-oriented characteristics; it uses both a variety of business bearing capabilities facing the network, and provides a quality assurance that approaches the connection network.
RTP provides end-to-end transfer services for data with real-time characteristics of interactive audio, video, and more. If the underlying network supports multicast, RTP can also send data to multiple destination endpoints.
RTCP is the control protocol of RTP, which is periodically communicated with participants of all sessions and transmits the control packets with the same mechanism as transmitted data packets.
supplement:
Video transmission of TV quality is approximately 33 frames / s; ISDN-based point-to-point video transmission can theoretically reach 30 frames / s. The actual transmission speed of the desktop video conference is generally 10 frames / s, and if the Internet is transmitted during the Internet or the large file is transmitted on the local area, the speed will drop to 5 frames / s.
With the development of standards such as H.323V2 and V.90, video conferences based on telephone network (using H.324 standard) and IP network (ie, based on local area network, using H.323 standard) will become the future development of this field. Mainstream.
