Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
A.2.4 Multimedia Communications Networks Multimedia networks refer to networks which integrate voice and/or video web data on a common network infrastructure. This integration requires special consideration to account for different characteristics of traditional data and voice~v~deo networks. Basically, data networks are most efficient when implement as packet networks (e.g., LANS) while voice and video are most efficiently implemented using TI, SONET (i.e., circuit-switched, synchronous technologies) networks. These requirements have been discussed in Section A.~.5 and, in particular, in Table I.5.13-~. Thus, ITS multimedia communication network design must address several issues. Multimedia networks require the conversion of voice and video to digital information so that it can be combined win other data through multiplexing techniques for transfer between points within He network. Voice and modon video have timing constraints to achieve real-time fidelity criteria for listeners/viewers. Networks which are synchronous, such as TI and SONET, guarantee timing and are capable of supporting multimedia Asynchronous network technologies (e.g. LANs, etc.) must have a means of guaranteeing timing for voice and video signals. One method is to give He video and voice data a higher network transfer priority Han over data. Examples include the emerging Asynchronous Transfer Mode (ATE standard and FDDI-D draft standard. Another method is to provide separate channels for isochronous versus asynchronous data as is provided in the recently adopted ~ ~ HE 802.9 isochronous ETHERNET standard. Traditional LAN standards (e.g., 802.3) do not support these methods; however, a clear mend is to provide these capabilities in emerging standards. When a network is not inherently synchronous the term "isochronous" ("iso for equal" and "chronous for delay) is used to indicate Hat He network design will accommodate voice and video. The major advantages of a multimedia network are presented in Table A.2.41. L::WCHRP`Phasc2rp ~NCHRP 3-51 · Phase 2 animal Report A2-14
l Table A.2.4~1 Advantage of Multimedia Networks _ . Feature Benefit . Single communications technology rather Lower life cycle cost than mixed technologies All communications needs accomplished Optimum use of medium's bandwidth and over a single medium for segments on the thus lower installation cost; Single fiber pair total network does alla Networ < distribution addressing added to Facilitates interoperability with distributed voice and video users and thus, lower information distribution cost . _ Voice, data, and video in digital form Computer fnendly~ format simplifies interfaces with information processing, storage, and retrieval environment Network management technology supports Built-in test and status reporting supports voice and video, as well as data use of open standard protocols; minimizes maintenance cost _ Fault tolerant features of networks also High availability for all information, applied to video and voice enhancing safely and reducing maintenance cost Unlike point-to-po~nt or analog overlay link video distribution, multimedia networks allow video and voice to be "addressed" to users. Thus, single users, work groups, or broadcast distribution techniques may be used. Voice is digitized supporting "voice maid' features of systems. Information is accessible from any compatible terrn~nal or the network assuming need and authority to use We ~nfonnabon. Access control is accomplished by a variety of techniques from addressing gateway locks (channel access "code key" locks and information encryption/descnption). There is a significant distinction between single frame video and Fill motion video. S~ngle frame video has no frame-to-frame synchronizad0n requirement. It is essentially a "stand alone" picture which is digitized. Compression aIgonthms that handle frame video (such as JPEGS) do not accommodate full-motion video. Full-motion video requires an algorithm that accommodates changes from frame-to-frame. MPEGS ~ is the modern algorithm handling fuB- motion video. 1 ., L;wcHlwha~n NCHRP3-51 · IPnase2F'nalReport A2-15
hnage quality in teens of horizontal and vertical resolution maintained through Me compression/decompression process and equivalent frames per second, defines needed data rate for compressed video. Generally as data rate is decreased, image quality is decreased. In fun motion video, not only is resolution decreased but a point is reached where equivalent frame rate causes image "jumping" and a rapidly changing image, especially in terms of percent of pixels consumed by the image from flame to frame (such as wig a rapidly approaching truck) will result in '~blocking." Blocking is when He aIgori~m cannot convert the compressed image back to pixels and Bus small colored blocks (perhaps "c x c") are displayed. These blocks are in fact the cells used to process pixels in Be compression algorithm. Thus, in Me design of a muldme~a network, one must define Be acceptable image quality. For full-motion, color video, two TIs (3.08 Mbps) is Be limit of acceptable quality condor surveillance video MPEG ~ provides. As data rates for MPEG ~ approach four, TIs (6.6 Mbps), picture quality approaches Cat of a nonnal analog television broadcast image of a high motion scene (such as a ball game or action movie). Older algorithms, such as Be 11 U (CC11-1) 261 standard, require higher data rates for equivalent image quality. ~stoncaDy, 45 Mbps have been used for video CODEC. Currently, 6 to ~ video channels can be supported at T3 data rates using MPEGS B. There are a number of voice CODECS which reduce the histor~caBy used DS-0 (64 kbps) data rate to as low as 4.5 kbps. Digital cellular telephones are deploying 7.5 kbps voice CODECS which provide high quality voice, where pitch is maintained, Be speaker's voice is identifiable and words are easily recognizable. Wig selection of video and voice CODECS, data loading on Be multimedia network can be defined. Digital video and voice data are added to Be computer-generatec] data to provide total multimedia Ott load for network segments. One important consideration: If video from surveillance CCTV cameras is required continually by someone on Be network, nothing is gained by using ATM over SONET. It is more cost effective to use Be synchronous channels of SONET for video. Continuous video placed over ATM, will continuously have highest priorly and continuously consume Me bandwidth, Bus LARCH - P~2'pt\ NCHRP3-51 · Phase2FmalReport A2-16