Videoconferencing Systems and Applications: An Overview, Thesis of Software Project Management

Download Videoconferencing Systems and Applications: An Overview and more Thesis Software Project Management in PDF only on Docsity! Chapter 19 VIDEOCONFERENCING SYSTEMS AND APPLICATIONS Sandra Brey and Borko Furht Abstract In this chapter we present an overview of videoconferencing technologies, systems, standards, applications, and commercial products with an emphasis on Internet-based video- conferencing. We begin with the history of videoconferencing and then we introduce fundamentals of audio and video technologies for videoconferencing. The videoconferencing standards, such as H320, H.321, H.323, and H.324, are described next. We discuss networks for videoconferencing and evaluate IP-based and ATM-based videoconferencing systems. Various videoconferencing applications are briefly described as well as a variety of commercial videoconferencing products and systems. 1. INTRODUCTION Videoconferencing is the transmission of live video images and audio between two or more disparate participants. Once merely a figment of science fiction writers’ imaginations, it is now used for both business and personal use, on a variety of different types of network media and with varying degrees of quality. Conversations may be one to one (point-to-point) or one-to-many (multipoint), in simplex (one-way only), half-duplex (one way at a time, taking turns) or full-duplex (all parties are seen and heard simultaneously). A range of products is offered, which span a wide spectrum of applications from group (or room) based systems, to desktop videoconferencing systems, to less expensive (and lower quality) personal conferencing/videophone systems. Products are available which can convert a multimedia PC, or even a television set, into a videoconferencing workstation. Networks used include ISDN, IP packet-data LANs, ATM networks, analog phone lines, and even the Internet. Of course, the quality of the video and audio obtainable depends in large part on the characteristics of the network used. Factors such as data throughput rate, delay, and delay variation, differ widely based on the type of network used. Several different standards have been developed which are optimized for use on each of the different network types. Chapter 19452 Taken together, these segments add up to a huge, rapidly growing business. This is partly due to the explosion of videoconferencing via personal computer and the Internet. The videoconferencing market is expected to exceed 3 billion dollars annually by 2001, as illustrated in Figure 1. Videoconferencing Marketplace Sales (Millions) $1,000 $2,000 $3,000 $4,000 19 93 19 95 19 97 19 99 20 01 1 Frost & Sullivan - figures include group and desktop endpoints and network infrastructure Figure 1. Projected growth in videoconferencing sales. 2. HISTORY OF VIDEOCONFERENCING The first demonstrations of videoconferencing in the United States occurred in the 1920s. Other experiments were done in Europe in the 1930s. Put on hold during World War II, research began again in the mid 40s. However, it wasn’t until 1964, at the World’s Fair in New York, that Bell Labs presented the Picturephone, shown in Figure 2, to the world. Figure 2. Advertisement for 1960’s Picturephone. Videoconferencing Systems and Applications 455 speech quality audio which encompasses a much smaller range of frequencies. Telephone quality audio extends from about 300 to 3300 kHz. After filtering, the signal is sampled. The amplitude values of an analog audio signal, representing the loudness of the signal, are continuously varying in time. To encode this signal digitally, the amplitude value of the signal is measured (sampled) at regular intervals. According to the Nyquist Theorem, to have lossless digital representation of the analog signal, the sampling rate must be at least twice that of the highest frequency present in the analog waveform. This is termed the “Nyquist rate”. Figure 3. PCM encoder simplified block diagram. The sampled value is then quantized. This requires the value to be mapped into one of a set of fixed values, which are binary coded for transmission. See Figure 3 for a diagram of the encoding process. The errors, which result from this mapping of analog values to quantized levels, result in “quantization noise”. It follows that the level of quantization noise drops as the quantization levels become closer together. Therefore, more bits of quantizer resolution translate to less quantizer noise, and hence greater dynamic range. Some implementations have a linear quantization. Other approaches may use a logarithmic quantization, which results in a type of audio compression, reducing quantization noise during quiet intervals without wasting bits at higher signal levels. The total bit rate for a monaural PCM signal can be found by multiplying the sample rate by the number of bits/sample. A stereo signal would require twice the bit rate, etc. Because of the fact that telephone quality voice transmission requires about 3 kHz of audio bandwidth and 256 quantum levels, a sample rate of 8 kHz and 8 bits per sample are commonly used, resulting in the 64 kbps channels used for ISDN and other phone applications. Chapter 19456 3.2.2 ADPCM Adaptive Differential Pulse Code Modulation is a compressed version of PCM, which requires a lower bit rate than standard PCM to transmit the same voice information. In DPCM, previous PCM samples are used to predict the value of the current sample. It is possible to do this because of the patterns present in speech samples. This prediction algorithm is performed at both the transmitting and receiving end. The transmitter compares the actual sample to its predicted value, and computes the error. Because the error signal will have a lower variance than the original speech samples, it can be quantized with fewer bits than the original speech signal. This error signal is then transmitted. Because the prediction algorithm is performed at the receiving end as well, the receiver knows what the predicted value is. It uses the error signal to correct the predicted value and reproduce the original sample. This predict-compare-adjust process is repeated for each input sample, reproducing the original PCM samples at the output. The system is called “adaptive” because the prediction parameters and the quantization levels of the error signal can change dynamically depending on the rate of change of the sample values (i.e. signal level). Many ITU-T videoconferencing recommendations include ADPCM encoding methods. Different flavors of ADPCM encoder/decoders vary in the way the predicted value is calculated and how the predictor or quantizer adapts to signal characteristics. This results in various levels of compression. Standards include G.721, G.722, G.723, G.726, and G.729. Various characteristics of these standards are summarized in Table 1. Higher quality speech (50 Hz -7 kHz, 14 bit resolution) may be encoded by dividing the audio spectrum into two subbands and performing separate ADPCM coding on each. The technique is covered in G.722 and is called “Sub-Band ADPCM”. G.722 specifies three modes of operation: 64, 56 and 48 kbps. 3.2.3 LPC/CELP/ACELP LPC (Linear Predictive Coding) is used to compress audio at 16 Kbps and below. An LPC encoder fits speech signals to a simple analytic model of the vocal tract. The signal is broken into frames, usually tens of milliseconds long, and best fit parameters for each frame are transmitted and used by the decoder to generate synthetic speech that is similar to the original. The result is intelligible but artificial sounding. Plain LPC is not included in videoconferencing standards, but is the basis for CELP (Code Excited Linear Prediction) which is important for obtaining high audio compression rates in videoconferencing. CELP is quite similar to LPC. The CELP encoder does the same frame- based LPC modeling but then computes the errors between the original speech and the synthetic model and transmits both model parameters and the errors. The error signal actually represents indices in a “codebook” of “excitation vectors” shared by the encoders and decoder. Thus the error signal is very much compressed. It follows that the computational complexity and speech quality of the coder depend upon the search sizes of the code books, which can be reduced at the expense of sound quality. CELP makes possible much higher quality speech at very low data rates. ITU-T Recommendation G.728 uses a variation of CELP, LD-CELP (Low Delay CELP). The compressed audio requires a bandwidth of only 16 kbps, but the encoder and decoder are quite computationally complex, requiring special hardware. Videoconferencing Systems and Applications 457 Table 1. Audio Standards G Family ITU Standard Year Approved Algorithm Used Bit Rate Bandwidth (kHz) Typical End- To-End Delay (ms) Application G.711 1977 PCM 48, 56, 64 3 <<1 GSTN telephony, H.323 & H.320 videoconferencing G.723 1995 MPE/ACELP 5.3, 6.3 3 67-97 GSTN videotelephony, H.323 telephony G.728 1992 LD-CELP 16 3 <<2 GSTN, H.320 videoconferencing G.729 1995 ACELP 8 3 25-35 GSTN telephony, modem h.324 GSTN videophone G.722 1988 subband ADPCM 48, 56, 64 7 <2 ISDN videoconferencing 3.3 VIDEO COMPRESSION Since network bandwidth is in limited quantity, and video is inherently bandwidth thirsty, the choice of video compression technique takes on great importance. Video is composed of a sequence of still images, called frames. The sequence of frames are presented at a rate that makes the motion of the depicted video scene appear fluid. The frame rate for television in the United States is 30 frames per second. The frame rate in a business quality videoconferencing session should be at least 15 frames per second. At lower rates the video will appear jerky. Each frame of the video is digitally represented as a two dimensional matrix of pixels. Color images are composed of three image frames, one for each color component. Video compression is typically lossy, meaning some of the information is lost during the compression step. The compression process takes advantage of the functioning of human vision, discarding information that is not perceptible. Further compression can be achieved, further reducing the required bandwidth, but at the sacrifice of quality. The required level of quality will depend on the application. Color space sampling and redundancy reduction is techniques common to most video codecs. Color space sampling is a technique used to reduce the amount of data that needs to be encoded. Because the human eye is less sensitive to chrominance information, an image encoded in YUV space can have the U and V components subsampled. In this way, these components will require one half, or less, of the bits required to encode the more important Y component. Redundancy reduction is also used to decrease the amount of encoded information. Intraframe encoding achieves compression by reducing the spatial redundancy within a picture. This technique takes advantage of the fact that neighboring pixels in an image are usually similar. Further compression is achieved through interframe encoding, which uses the fact that neighboring frames in a sequence of images are usually similar, by reducing the temporal redundancy between frames. 3.3.1 Discrete Cosine Transform Discrete Cosine Transform is a video compression technique that forms the basis for the two important video compression standards, H.261 and H.263. This compression algorithm is Chapter 19460 Table 2. Video Picture Formats. Picture Format Number of Luminance Lines Number of Luminance Pixels Number of Chrominance Lines Number of Chrominance Pixels Supported in H.261 Supported in H.263 sub-QCIF 96 128 48 64 not supported optional QCIF 144 176 72 88 optional mandatory CIF 288 352 144 176 mandatory mandatory 4CIF 576 704 288 352 not supported optional 16CIF 1152 1408 576 704 not supported optional Where the H.261 standard was limited to full pixel precision for motion compensation, H.263 has required support of half-pixel precision. The half-pixel refinement greatly improves the picture quality, particularly in low-resolution video. New to the H.263 standard are negotiable coding options offering improved performance. One of these options is the support of P-B frames used in interframe encoding. This is a technique that is also used in MPEG video. Although it is computationally more expensive, it allows for much higher compression, and therefore a potentially higher frame rate. 4. COMPONENTS AND FUNCTIONS OF A VIDEOCONFERENCING SYSTEM A typical videoconferencing system and its components are shown in Figure 5. Videoconferencing stations are equipped with video and audio capture and compress subsystems, and decompress and display subsystems. The communication media can be POTS (Plain Old Telephone System), LANs, or WAN. Figure 5. Components of a videoconferencing system. The common software architecture of a videoconferencing system, shown in Figure 6, consists of videoconferencing application, middleware, video and audio codec and data, stream multiplexer and demultiplexer, and line drivers. Video capture and compress Video capture and compress Audio codec Audio codec Display and decompress Display and decompress Comm. Comm. POTS, LAN, WAN Camera Camera Speaker Microphone Speaker Microphone Videoconferencing Systems and Applications 461 Figure 6. Common software architecture of a videoconferencing system. Videoconferencing systems support a number of functions including: Multipoint connection set up. The system should be able to negotiate for network resources and end-to-end conference capabilities. Dynamic session control. The system should have an ability to add and delete participant to/from an existing videoconference. Such a change may require modification of the underlying network configuration. Conference directory. The system should provide a conference directory service that will support conference registration, announcement, and query. The directory should contain various information such as: the title and the brief description of the conference, a list of participants, start and end time for the conference, audio and video coding schemes, their protocols and QOS requirements, and shared working space. Automatic conference scheduling and recording . The conference scheduling function combined with resource reservation mechanisms will allow planning of network resources. Automatic conference recording is a useful function that does recording of conference sessions in a form of multimedia documents. Conference termination. The system should be capable to release all reserved resources when the conference is complete. 5. VIDEOCONFERENCING STANDARDS AND NETWORKS 5.1 VIDEOCONFERENCING STANDARDS The original, widely accepted videoconferencing standard was H.320, which defines a methodology for transporting videoconferencing traffic over ISDN. However, in 1996, additional standards for videoconferencing emerged - H.323, H.321, and H.324. These standards define methodologies for videoconferencing over other various networks such as POTS, IP networks such as LANs and the Internet, and ATM networks. Each of these standards brings with it certain capabilities and quality levels, illustrated in Figure 7. Each Application Videoconferencing middleware Data Audio codec Video codec Stream mux/demux Line drivers (Async, LAN) Shared apps still images (JPEG) fax, file xfers Audio in Video in (digital) Chapter 19462 has advantages and disadvantages in videoconferencing transmission. The various characteristics of these standards are summarized is Table 3, and discussed in the following sections. Figure 7. Quality of videoconferencing standards. Videoconferencing involves the transmission of video, audio, and data. Data can be in the form of whiteboard data or shared application data, used in a collaborative conference. These different types of information have different reliability and delay variation requirements of the networks over which they are being transmitted. Table 3. Characteristics of Various Videoconferencing Standards H.320 H.321 H.323 H.324 Approval Date 1990 1995 1996/1998 1996 Network Narrowband Switched digital ISDN Broadband ISDN ATM LAN Non-guaranteed bandwidth packet switched networks POTS, the analog phone system Video H.261 H.263 H.261 H.263 H.261 H.263 H.261 H.263 Audio G.711 G.722 G.728 G.711 G.722 G.728 G.711 G.722 G.728 G.723 G.729 G.723 Multiplexing H.221 H.221 H.225.0 H.223 Control H.230 H.242 H.242 H.230 H.245 Multipoint H.231 H.243 H.231 H.243 H.323 Data T.120 T.120 T.120 T.120 Communication Interface I.400 AAL I.363 AJM I.361 PHY I.400 TCP/IP V.34 Modem Videoconferencing Systems and Applications 465 The delay between audio and video is very critical and limited to 20 to 40 milliseconds in order to provide good lip synchronization. 5.3 CIRCUIT SWITCHING vs. PACKET SWITCHING NETWORKS Circuit-switched communication is a manner of data transmission where the communication path is established and reserved for the duration of the session. The bandwidth allocated for a session is used exclusively by it alone. The resources used by the session are freed, and available for other calls, at the end of the session. The dedicated bandwidth is an advantage for videoconferencing, providing predictable delays. However, circuit-switching transmission underutilizes network resources, as the dedicated bandwidth can go unused at times of limited activity. Unlike circuit-switched transmission, a packet-switching environment has no dedicated bandwidth circuit set up; the bandwidth is shared with other network users. The information is divided into packets, and each packet is routed individually through the network. Since the packets may take different routes, they may arrive at their destination at different times and out-of-order. This variable delay in the delivery of information can cause problems in the quality of videoconferencing applications. Video packets received out-of-order may have to be discarded. Various protocols have been developed to try to overcome these inherent problems associated with packet switching such as RSVP, and RTP. These attempt to provide some quality of service over this type of transmission, and are discussed in later sections. 5.4 ISDN VIDEOCONFERENCING Integrated Services Digital Network, ISDN, is a circuit switched end-to-end digital service. ISDN was designed to support the transmission of a wide variety of services, including voice, data, and video at high speeds over the public switched telephone network (PSTN). ISDN relies on 64 kbps channels, originally chosen to support digitized voice traffic. Two access rates are defined for ISDN: Basic Rate Interface (BRI) and Primary Rate Interface (PRI). The user information is carried over bearer channels, known as B channels, each having a capacity of 64 kbps. The Basic Rate Interface provides 2 B-channels, while the Primary Rate Interface provides 23 B channels, in North America and Japan, and 30 B channels in Europe. A separate channel, the D-channel, is used for signaling. The D-channel has a data rate of 16 kbps in the BRI and 64 kbps in the PRI. While a single BRI service is not sufficient to support a business quality videoconference with a frame rate of 15 frames per second, this service does satisfy requirements for many videoconferencing applications, such as desktop videoconferencing. A business quality videoconference will require at least 6 B channels with a data rate of 384 kbps (6 x 64). This setup is normally only seen in group based systems. In the past, ISDN videoconferences were just point-to-point connections. However, today it used for multipoint as well, by using a Multipoint Control Unit (MCU). Figure 10 illustrates a typical ISDN videoconferencing session. Despite its reputation for difficult installation, lack of availability, and expense, many videoconferencing products on the market utilize ISDN. One of the main reasons for ISDN videoconferencing's broad acceptance is the timely emergence of ITU standards supporting it in 1990. The standards for other communication networks were not seen until 1996. However, with so many changes taking place, and the ratification of further standards for videoconferencing over such diverse networks as POTS, and IP based networks such as LANs and the Internet, ISDN's days as the dominant videoconferencing communications medium Chapter 19466 are over. While ISDN will continue to provide the quality required for business applications, POTS and Internet products have become more widespread among home and less serious users. ISDN run to Each Location IMUXISDN Network 384 B R II PRI Gateway PRI Figure 10. Example of an ISDN videoconference connection. ITU-T Standard H.320 ITU-T’s ratification of standard H.320, titled "Narrow-Band Visual Telephone Systems and Terminal Equipment", in December 1990, gave the videoconferencing industry a much needed jump-start. This standard was optimized for popular circuit-switched media such as ISDN. For the first time, a standard existed that made it possible to link equipment from different vendors in the same conference. H.320 is an umbrella standard. It contains other standards concerning audio and video encoding and compressing, specifications for how calls are negotiated, how the data is multiplexed and framed, multipoint conferencing, data transmission, and communication interfaces. 5.5 PLAIN OLD TELEPHONE SERVICE (POTS) VIDEOCONFERENCING The telephone system gradually began converting its internal connections, once purely an analog system, to a packet-based, digital switching system in the 1960s. Today, nearly all voice switching in the US is digital within the telephone network. However, the final link from the local central office to the customer site remains primarily an analog line, although it too is being replaced. This connection is often called Plain-Old Telephone Service (POTS). The standard telephone system is the most widely pervasive network in existence. The availability of this existing infrastructure, combined with the latest CPU's, compression techniques, and advanced modem technologies, has brought videoconferencing to the home consumer. In May of 1996 the ITU-T ratified the H.324 standard for videoconferencing, which defines a methodology for transporting videoconferencing across POTS. Videoconferencing Systems and Applications 467 POTS videoconferencing has been technically difficult because of the low bandwidth of audio phone lines. However, with improved coding methods, there is sufficient bandwidth available to support audio, video and data sharing with this media. And the equipment is inexpensive and easily installed, relative to other existing methods. Although most Pentium PC’s can digitize local video at 15 fps and higher with little problem, quality suffers because the POTS modems cannot transmit the data fast enough to maintain the same frame rate. This typically restricts POTS-based videoconferencing to just a few frames per second. Thus, POTS videoconferencing does not approach the levels of quality required for even the most casual of business needs, and is better suited for home and recreational uses. A POTS connection can be established either modem-to-modem, or over the Internet. An Internet POTS connection requires the user to first establish a dial-up connection to the Internet. A call is then made to someone else on the Internet, who is also probably using a modem. The quality of the connection varies, as Internet traffic can severely compromise the quality of the videoconference. A typical frame rate is 2 to 5 frames per second, and basically looks like a series of stills. The advantage to this method is that a call can be made to anyone in the world for the price of accessing the Internet. A modem-to-modem connection can achieve frame rates of 5 to 10 frames per second, because the Internet traffic is avoided. But, long distance phone charges will apply if calling out of area. POTS videoconferencing is finding acceptance in the consumer marketplace, due to the ubiquitous nature of the phone system. Also, recreational videoconferences do not require the same quality level of a live business meeting or a training class. Furthermore, today’s PCs come multimedia equipped, with modems, are inexpensive, and readily available. ITU-T Standard H.324 The H.324 ITU standard, titled "Multimedia Terminal for Low Bitrate Visual Telephone services over the GSTN", was the first standard to support point-to-point video and audio compression over analog POTS lines. Specifically, H.324 is designed to optimize videoconferencing quality over the low-speed links associated with the POTS system, typically operating at the speeds of modems - 28.8 kbps - 56 kbps. This standard allows users to interoperate across diverse endpoint such as ISDN, ATM, POTS, or mobile devices, and makes it possible to hold modem-based, POTS video calls that connect equipment made by different vendors. H.324 terminals may carry real-time voice, data, and video, or any combination, including videotelephony. H.324 is an umbrella recommendation with a similar structure to H.320, its ISDN counterpart. See Section 4.7, Summary of Standards, for specifics on the various standards that H.324 supports. 5.6 IP-BASED VIDEOCONFERENCING Networks are a fundamental part of today's information systems. They form the backbone for information sharing in enterprises, governmental and scientific groups. This shared information comes in a variety of forms such as e-mail and documents, files sent to colleagues, and real-time applications such as videoconferencing. Local Area Networks (LANs) are commonly used on campuses and in companies to connect desktop computers together. At the physical layer, LAN’s are frame-based and usually consist of 10 Mbps Ethernet, 16 Mbps Token Ring segments, or ever 100 Mbps Fast or Switched Ethernet. Chapter 19470 probable that small 'latency sensitive' frames carrying videoconferencing could easily be stuck in a buffer behind much larger 'data' frames. This is the case for 10Base-T, 100Base-X and Gigabit Ethernet. RSVP is a simple signaling system, where every hop-by-hop link must be negotiated separately. There is still no end-to-end guarantee of a minimum service level. ATM, on the other hand, sets up an end-to-end connection with a specific QoS class that commences at call set-up and ends at call teardown. 5.6.4 RTP In conferences with multiple audio and video streams, unreliable transport via UDP uses IP Multicast and the Real-Time Protocol (RTP) developed by the Internet Engineering Task Force (IETF) to handle streaming audio and video. IP Multicast is a protocol for unreliable multicast transmission in UDP. RTP works on top of IP Multicast, and was designed to handle the requirements of streaming audio and video over the Internet. A header containing a time-stamp and a sequence number is added to each UDP packet. With appropriate buffering at the receiving station, timing and sequence information allows the application to eliminate duplicate packets; reorder out-of-sequence packets; synchronize sound, video and data and achieve continuous playback in spite of varying latencies. RTP needs to be supported by Terminals, Gateways, and MCUs with Multipoint Processors. 5.6.5 RTCP The Real-Time Control Protocol (RTCP) is used for the control of RTP. RTCP monitors the quality of service, conveys information about the session participants, and periodically distributes control packets containing quality information to all session participants through the same distribution mechanisms as the data packets. 5.6.6 Multicast In some videoconferencing applications it is necessary to send the same real-time video and/or audio streams to multiple destinations throughout the global Internet. Typically this would be accomplished by sending multiple streams of redundant packets, one for each destination. This can be very inefficient and slow. RTP-based applications can use “IP multicast” capabilities in conjunction with the MBONE (Multicast BackBONE), a virtual network designed to facilitate the efficient transmission of video and audio signals simultaneously over the Internet. The network is composed of “islands”, Internet sites supporting multicast and linked by virtual point-to-point links called "tunnels". The IP multicast packets are encapsulated for transmission through the tunnels, so that they look like normal unicast datagrams to intervening routers and subnets. Once they reach the destination island, they are copied and forwarded to destinations as required. 5.6.7 ITU-T Standard H.323 In 1996, the ITU ratified the H.323 standard for videoconferencing over packet-switched networks, such as Ethernet and Token-Ring, and ultimately the Internet. The standard is platform independent and runs on top of common network architectures. To ensure that critical network traffic will not be disrupted by videoconferencing traffic, the standard includes network traffic management capabilities and supports multicast transport in multipoint conferences. H.323 defines four major components for a network-based communications system: Terminals, Gateways, Gatekeepers, and Multipoint Control Units, shown in Figure 11. Videoconferencing Systems and Applications 471 Terminals are the client endpoints on the LAN that provide the user interface. All terminals must support voice communications; video and data are optional. Also required of H.323 terminals is support of H.245, for negotiation of channel usage and capabilities, Q.931 for call signaling and call setup, Registration/Admission/Status (RAS), a protocol used to communicate with a Gatekeeper; and support for RTP/RTCP. Gateways provide translation between H.323 conferencing endpoints and other terminal types, such as H.320 and H.324. Gateways are not required if connections to other networks are not needed, since endpoints may directly communicate with other endpoints on the same LAN. Terminals communicate with Gateways using the H.245 and Q.931 protocols. Figure 11. H.323 network and interfaces. The Gatekeeper acts as the central point for all calls within its zone and provides call control services to registered endpoints. It translates addresses from LAN aliases for terminals and gateways to IP or IPX addresses, as defined in the RAS specification, and manages bandwidth. While a gatekeeper is an optional part of an H.323 system, all compliant terminals must support its use. The Multipoint Control Unit (MCU) supports conferences between three or more endpoints, handling negotiations between all terminals to determine common capabilities for audio and video processing, and controlling multicast. 5.7 ATM-BASED VIDEOCONFERENCING Unlike many specialized networks used in the past, Asynchronous Transfer Mode, ATM, is a networking technology that was designed at the outset to be service independent. The flexibility of ATM allows it to support a variety of services with different bit rates, burstiness, and acceptable delay characteristics. The bit rate can be constant for the whole transmission, or it can be variable over time. ATM satisfies all the various types of data, each with unique service requirements, found in videoconferencing such as video, audio and data. Chapter 19472 ATM was designed to be efficient in the use of its available resources. All available resources are shared between all services, such that the optimal statistical sharing of the resources is achieved. 5.7.1 Basic Principles of ATM ATM is based on the concept of asynchronous data transfer. It combines the best qualities of circuit-switched and packet-switched communication. ATM does not use fixed bandwidths as in circuit switching techniques such as ISDN. It is implemented by using virtual circuit packet switching technology with fixed sized packets, called cells. In the virtual circuit technique all the cells in a call follow the same route, and arrive in the correct sequence. Each cell is 53 bytes in length, 48 bytes for the information field and 5 bytes for the header. To guarantee a fast processing in the network, the ATM header has very limited function. Its main function is the identification of the virtual connection by an identifier, which is selected at call set up and guarantees a proper routing of each packet. The information field length is relatively small, reducing the network queuing delays. This in turn leads to small delays and a small delay jitter as required by real-time applications such as videoconferencing. No processing is performed on information field of the cells inside the network, further reducing delays. ATM is connection oriented. Before information is transferred from the terminal to the network, a logical/virtual connection is set. When a user requests a connection, it provides information to the network about its requirements such as peak cell rate, acceptable cell delay variation, sustainable cell rate, and its highest expected cell burst. If the network can support this connection without jeopardizing the promised quality of service for the existing connections, the connection is accepted. On admission into the ATM network, a virtual channel is set up between the end users, through the network. ATM provides a Quality of Service (QoS) system that is vastly more sophisticated than that found in IP networks. While IP networks attempt to provide quality with protocols such as RSVP and RTP, it is not guaranteed. A constant good quality videoconferencing application with a fixed frame rate, no audio gaps, and lip synchrony, can only be guaranteed on a network with a guaranteed QoS. Although possible on dedicated LANs or unoccupied high- speed links, in general it requires ATM or ISDN connections, with guaranteed QoS. 5.7.2 ITU-T Standard H.321 H.321 is the ITU-T’s standard defining the implementation of videoconferencing over ATM. It is an enhancement of the ISDN standard H.320, and is fully compatible with existing H.320 systems. Like ISDN, H.321 is implemented at transmission rates, which are multiples of 128 kbps (128 Kbps, 384 Kbps, 768 Kbps, etc.). The standard was designed to take advantage of the inherent QoS capabilities of ATM, delivering the highest quality videoconferencing. Implementing an H.321 ATM-based videoconferencing is less costly and less complex than ISDN, for several reasons. ATM switches are significantly cheaper than ISDN switches. Also, in an ATM implementation, single gateways give centralized access to other networks, where an ISDN implementation requires separate IMUXs for each end-point ISDN user. Not only does this extra cabling required in ISDN implementations increase costs, it also makes the installation more complex. H.321 is claimed to deliver high-quality videoconferencing at significantly lower costs and with greater flexibility than with ISDN based standards, yet to provide significantly higher quality levels than IP-based implementations. ATM's unique suitability to the transport of Videoconferencing Systems and Applications 475 Figure 13. A virtual field trip using videoconferencing. 6.4 DISTANCE TRAINING Most large corporations have determined that it is in their critical business interest for their workers to have access to the educational opportunities that will help them increase their knowledge level and skill sets. They know this is necessary to compete, and accordingly are investing large sums of money in workforce training and education. Today over $50 billion is spent annually on corporate training, and over $6 billion is spent annually on the purchase and maintenance of facilities and equipment for this purpose. Videoconferencing is quickly becoming part of many companies' overall employee training programs, because it provides a way to offer high-quality training while cutting course delivery costs. Because employees do not have to leave their offices for extended periods of time, disruptions are minimized and short-term productivity is not sacrificed. 6.5 PERSONNEL RECRUITMENT Videoconferencing can be used as a screening tool for corporate personnel departments. It allows face-to-face interviews to occur without the expense of flying each potential candidate to the job location. 6.6 HEALTH CARE Videoconferencing is used in numerous clinical applications. The most common application is in radiology, but it is also used for cardiology, dermatology, psychiatry, emergency medicine, home health care, pathology, and oncology. The healthcare industry is a rapidly growing and evolving market. As healthcare corporations merge and form huge networks, successful organizations will be those that can leverage the medical expertise and resources of their member institutions across multiple sites. The use of videoconferencing increases operational efficiency through better use of staff resources, provides rapid information transfer, achieves better physician collaboration, and facilitates medical education, as illustrated in Figure 14. Chapter 19476 Figure 14. A remote consultation. Figure 15. MedLink bedside terminal. Uses include patient interviews, remote medical examinations using remote sensors, medical education via “interactive rounds”, remote consultation with a specialist, and even such mundane uses as enabling hospital administration to negotiate medical supply contracts. Videoconferencing is used in some hospitals to facilitate the care of newborns in the neonatal intensive care unit (NICU). These babies may be hospitalized for the first several months of their lives, see Figure 16. Videoconferencing systems are installed in the NICU, as well as the homes of families with newborns in the unit, enabling family members to see and talk to their babies and review their progress with doctors, all from their homes. Once the baby is sent home with his or her nervous parents, the hospital staff use the system to monitor the progress of the baby and provide comfort and support to the parents. By using videoconferencing, the hospital can smooth the transitions of NICU babies from the hospital to their homes, while reducing overall costs. Figure 16. Baby in NICU. Psychiatrists can us videoconferencing technology to monitor their patients such as those with severe mental illnesses, personality disorders, and adjustment problems. For example, a psychiatrist employed at a state hospital may have patients who live over a wide area and who are unable to travel. Even hearing- or speech-impaired patients may be reached in this way, provided that the doctor knows sign language. One example of videoconferencing used by the medical industry is the MedLink Mobile Videoconferencing Unit (by PictureTel), designed to operate at the patient’s bedside and Videoconferencing Systems and Applications 477 allow him to speak with a remote specialist face-to-face. This system is shown in Figure 15. The MedLink delivers separate or simultaneous medical data along with videoconferencing to any of PictureTel's other products. It is based on the Venue 2000 Model 50 platform and is network independent. The network link can range from simple ISDN through T-1/E-1 or the unit can be attached to the local area network to take advantage of existing topologies. 6.7 LAW ENFORCEMENT AND SURVEILLANCE Videoconferencing technology is used for remote video monitoring of secured locations, as well as for replacing human operators at monitoring locations, such as at drawbridges, which must be raised and lowered based on boat and car traffic. Figure 17 illustrates TeleEye, a remote videoconferencing surveillance system. Figure 17. TeleEye videoconferencing surveillance system. 6.8 TELECOMMUTING Some companies use videoconferencing to allow employees to work out of their homes. This is part of a trend termed telecommuting. For example, IBM has put 95% of their US marketing and services personnel into telecommuting. They have been able to close or reducing the size of the field sales offices accordingly, and claim a 15% gain in productivity and a 40-60% savings in real estate per location. One benefit of the system is that it allows cheap customer follow-up “video visits”, which be cost-prohibitive if carried out in person. 6.9 LAW AND CRIMINAL JUSTICE Many state and federal courts now use videoconferencing to arraign criminals, for appeals and parole hearings, and to provide telemedicine to inmates. Lawyers use desktop system to hold meetings with their clients and to gather data from expert witnesses without incurring large expenses. Chapter 19480 Figure 20. Desktop Figure 21. A desktop videoconferencing session. videoconferencing equipment. Table 5. Desktop Videoconferencing Products. Manufacturer Product Product Description Camera Mic H.320 H.323 T.120 Intel ProShare Video System 500 Computer Add-in Card Π Π Π Π Π PictureTel Live200 Computer Add-in Card Π Π Π Π PictureTel LiveLAN Computer Add-in Card Π Π Π Π VTEL SmartStation Desktop 384 2 Computer Add-in Cards Π Π Π Π Π Zydacron OnWAN350 Computer Add-in Card Π Π Π Π 7.3 PERSONAL VIDEOCONFERENCING WITH VIDEOPHONES Videophones are also considered a personal conferencing tool, but are generally of lower quality than the desktop-to-desktop variety, and appeal to a different market segment. Videophones may be special telephones that include a small video screen, set-top boxes operating with the user’s TV, but are most often implemented as software running on the PC. They communicate via POTS analog phone lines, and conform to the ITU-T’s H.324 standards. The dedicated telephone with integrated video screen, known as a desktop videophone, shown in Figure 22, is the most expensive option in the videophone arena, however installation and operation is simple. These devices simply plug into any analog telephone outlet and are ready to make a video call. Depending on the vendor, features such as adjustable picture quality, size and frame speed, electronic pan, tilt, zoom, snapshot, caller ID, and auto answer can be found. There is no software to install, or special wiring involved. These videophones are designed to work with other H.324-compatible videophones including computer-based videophones. Videoconferencing Systems and Applications 481 8X8 VC150 MM220 Videophone Figure 22. Desktop videophones. Figure 23. Comtrad C-Phone system. Figure 24. Set-top videophone system. The next, slightly less expensive, option is the set-top box device, shown in Figures 23 and 24. This device is about the size and shape of a cable TV converter box, and hooks up to the TV just like a VCR. It may be operated via remote control. The product includes a camera, a modem, and an H.324 industry standard videoconferencing codec. The system uses a television set to present the audio and video of the person being called, thereby making this solution more economical. Figure 25. Videophone kit. Figure 26. Intel videophone. Chapter 19482 A third option available for videophone conferencing is available in the form of software that is installed on the user’s PC. These products are similar to the desktop videoconferencing systems, except that these systems are operating over POTS. Table 6 shows a list of videophone manufacturers and products. This solution takes advantage of the existing capabilities of today’s multimedia PC, making this it by far the most economical solution available. In fact, new PCs, with an Intel® Pentium® III processors, come preinstalled with the latest Intel® Video Phone software, shown in Figure 26. All that is required is a camera. Most videoconferencing vendors provide the products as a kit, which includes the software, video capture board (optional), and a camera, shown in Figure 25. The user can choose to make video phone calls over regular telephone lines or through the Internet. When used over regular telephone lines, the audio and video quality is much better than that of video phone calls made through the Internet. The quality of an Internet videophone call will vary depending on Internet traffic at the time of the call. Many of these products support both H.323 and H.324 standards, and support a variety of broadband Internet connections, including cable modems, DSL, ADSL and LAN. Table 6. H.324 Videophone Products. Manufacturer Product Number Product Description 8x8 VC150 ViaTV Desktop Videophone 8x8 VC105 ViaTV Set-Top Videophone W/Camera Winnov Videum Conference Pro Videoconferencing Kit Winnov VideumCam Videoconferencing Kit Panasonic Eggcam Desktop Video Camera w/Cu-Seeme 3Com Bigpicture VideoPhone PC Videoconferencing Kit, includes video capture card Intel Create and Share PC Videoconferencing Kit References 1. E. Brown, “Videoconferencing Systems,” New Media, December 1998. 2. K. Nisenson, “Tune in to IP Videoconferencing,” Network World, September 1998. 3. W. Wong, “Video Conferencing for the Enterprise,” Network Magazine, April 1998. 4. T. Trowt-Bayard and Jim R. Wilcox, “Video Conferencing and Interactive Multimedia: The Whole Picture”, Flatiron Publishing, March 1997. 5. M. Herman, “The Fundamentals of H.324 Desktop Videoconferencing,” Electronic Design, October 1996. 6. C. Tristram, “Video Conferencing: A Work in Progress,” Network Magazine, April 1998. 7. S. J. Bigelow, “The Many Faces of Videoconferencing,” Bay Area Computer Currents, October 1998. 8. K. Cholewka, “IP VideoConferencing: Beat the Clock,” July 1997, http://www.data.com/tutorials/video.html.