The telephone was invented in 1876, either by Alexander Graham Bell or Elisha Gray. Bell's attorney registered his telephone patent at the U.S. Patent Office a few hours after Gray's attorney had registered a caveat, an announcement of a telephone invention he intended to file a patent for. The men fought for ownership of the invention in litigation that lasted for years. Finally, Bell was given credit for inventing the telephone. It was later determined that Gray's proposed apparatus would have worked, but Bell's apparatus, as represented, would not.
Whoever invented it, telephone technology evolved tremendously, and progressed from a wired analog system to one that is wired and wireless, and digital. The system's sophistication is impressive; by entering a string of digits, a user's voice is quickly routed over networks to another user and conveyed with good sound quality. The telephone system became the basis for the Internet, a technology that revolutionized the dissemination of information. It would be a tame prediction to say that soon every bit of human knowledge and information will be accessible on the Internet, and that includes music. Whether as MP3 or MPEG-4 files, or proprietary protocols, the Internet can be used to efficiently download recorded music, and to stream live music as it is performed. Whether the connection is between two computers or millions of them, across a room or around the world, via dedicated systems or ordinary telephone lines, digital audio comprises ever more important telecommunications content.
Telephone Services
The digital transmission systems used to convey voice communications synonymously carry high bandwidth digital audio transmission. Both analog and digital signals attenuate over long distances; whereas analog signals are thus subject to increased noise, digital signals can be regenerated with great precision. Telephone companies almost exclusively use electronic switching and fiber-optic cable between telephone exchanges. As far as the telephone company is concerned, any phone call is treated simply as data; a long-distance call has the same sound quality as a local call. However, in most cases, the connection from the consumer's landline phone to the exchange is still analog.
Some information regarding the plain old telephone system (POTS): the telephone lines provided by the phone company are known as subscriber loops or central office lines. They connect users to central offices via analog lines.
Trunk lines connect users to a private branch exchange (PBX) where switching routing takes place. Circuit switching is used, in which a continuous connection is temporarily established between users, exchanging data at the same rate. The audio frequency response for typical subscribers is 300 Hz to 3400 Hz. Speech is coded with 8 bit PCM, at a sampling frequency of 8 kHz to yield a bit rate of 64 kbps; bandwidth is sharply limited to 3.4 kHz.
Amplitude companding and other means are used to improve dynamic range and fidelity.
Before the 1960s, telephone companies used copper wires to convey analog communications between switching offices, one pair of wires per phone conversation. Bundles of wires, each with thousands of pairs, ran through underground conduits. To increase capacity, Bell Laboratories turned to digital communications, and devised the T-carrier network. Illinois Bell installed the first T-carrier digital transmission system in 1962. Today, most long-distance telephone carriers use a variety of T-carrier lines, using copper, optical fiber, microwave radio, and coaxial cables to convey digital audio, video, and data.
T-1 serial digital communication circuits provide a dedicated point-to-point bandwidth of 1.544 Mbps; the signal they convey is called DS-1. Each end of a T-1 line terminates in a customer service unit (CSU), as shown in FIG. 1. The CSU interfaces data from the customer premises equipment (CPE) or a data service unit (DSU) and encodes it for transmission along the T-1; a multiplexer or local area network (LAN) bridge could be used. For example, long-haul T-1 lines can connect a long-distance provider's office in Miami with an office in Seattle; each office is known as a point of presence (POP). Each POP then communicates with local access networks, where the T-1 line terminates with a CSU. The CSU then communicates via a DSX-1 signal that interfaces to data signals such as RS-449 or V.35. A T-1 user sees two conventional copper wire pairs, one for send, and one for receive, and an RJ-48C plug.
A T-1 circuit is capable of carrying Compact Disc data (1.41 Mbps) without data reduction, 24 voice channels, or a single video program with data reduction. A T-1 circuit comprises 24 subchannels (called DSOs or slots) each carrying 64 kbps (8 bits with 8-kHz sampling).
Communications such as data or normal voice traffic can employ these subchannels. Data bytes are applied to frames, with one frame holding 24 DSOs plus one framing bit. Thus a frame holds 193 bits (24 × 8 + 1); the frame rate of 8 kHz yields a 1.544-Mbps overall rate.
Many different applications, such as voice, data, or video, can share one T-1 line, with individual channels assigned to one or multiplexed DSOs, as needed. T-1 is full duplex (bi-directional) and the assignments need not be identical. A line might convey a few high-quality audio channels in one direction, with many voice channels in the other. In some cases a Fractional T-1 line can be used; only a few DSO slots are assigned, as needed. T-1 lines are very reliable, with a bit-error rate (BER) of 10^-9. Individual DSOs can be sent to different destinations with the digital access cross-connect system (DACS).
Figure 1 A T-1 circuit provides long-haul digital communication at a data
rate of 1.544 Mbps; a CSU terminates each end of the line.
Other services include the T-0 (or DS-0) service.
Operating at 64 kbps, a T-0 line can deliver one 56-kbps digitized voice channel, sampled at 4 kHz, with 7 bits; that is, one standard voice telephone call. Some video teleconferencing systems use six DS-0 slots (384 kbps).
The T-1C service operates at 3.152 Mbps. The T-2 service operates at 6.312 Mbps; it is not offered commercially. The T-3 (or DS-3) service operates at 44.736 Mbps; it can deliver 28 T-1 channels or 672 voice channels, or one compressed television channel. The T-4 service operating at 274.760 Mbps delivers 168 T-1 channels or 4032 voice channels. The overall data rates have numerical differences from multiples of the basic 64 kbps; this is due to framing information that must be added. All DS levels use alternate mark inversion (AMI ) channel coding; the signal has three levels: positive, negative, and ground reference. The T-1 through T-4 rates in Europe are 2.048, 8.448, 34.368, and 565.000 Mbps, respectively.
ISDN
Integrated Services Digital Network (ISDN) is a dial-up telephone service with full duplex operation. ISDN provides a digital connection between the user and the telephone exchange, and ultimately to long-haul digital transmission systems. An overall rate of about 144 kbps is supported.
Although ISDN uses existing telephone lines, specialized equipment is needed at the send and receive ends. Basic rate ISDN (sometimes called 2B+D) is intended for home use. It uses standard copper pair wire to provide two 64 kbps circuits (B or bearer channels) to send and receive audio or other information, and one 16-kbps (D or data channel) circuit for dialing and other signaling functions.
Audio, video, fax, telephone, or other data can be transmitted. Primary rate ISDN is offered to business customers. It exists as either coaxial cable or fiber cable. It can provide 23 (or more) 64-kbps channels, along with a 64-kbps D channel (23B+D), totaling 1.544 Mbps of bandwidth (equivalent to a T-1 line).
With regular switched telephone lines, communication can be briefly interrupted by concurrent signaling and call directing data (for example, tones in the same bandwidth as the communication data); this is called in-band signaling. With ISDN, the bearer channels are independent of the signaling channel; this is called out-of-band signaling.
The advantage is that the bearer channels are specifically designated for the user and are uninterrupted by any signaling data. Timecode information can be transmitted over ISDN, using, for example, an RS-232 data channel. In addition, numerical data and compressed video signals can be sent over ISDN, with speeds exceeding that of dial up modems that must convert data for analog transmission through the subscriber loop.
Although simple telephones could be used in a basic voice-only ISDN hookup, to exchange audio data the user supplies an A/D and D/A converter, data reduction encoder/decoder (codec), and ISDN hookup as shown in FIG. 2. Data between the user and the local telephone exchange is sent over copper wires; lines are terminated at a dedicated terminal adapter (TA) interface. The copper wire pair connecting to the user is called the user (U) interface and a network termination (NT-1) converts this to a four-wire ST interface. The outgoing data is converted to a telecommunications format where it is directed over a carrier to the receiving party, which must have corresponding ISDN service and equipment. Using an inverse multiplexer (I -MUX), multiple digital lines can be combined to create a higher data-rate service.
Fig. 2 An example of an ISDN connection between two users. Each user
needs appropriate interfacing equipment. The long-distance connection can be
via fiber optic cable, microwave, or other means.
For example, a 384-kbps signal could be transmitted over six 64-kbps channels. The I-MUX resynchronizes the different lines (they might go through different long-distance routing) to produce a coherent stream at the receiver. I-MUX devices are placed at the terminal adapter or the codec. Depending on the number of channels and fidelity required, numerous data reduction methods can be used.
The coding delay in an ISDN system must be accounted for. For example, when recording an overdub from a remote location, the user might send the mix over ISDN and listen to a delayed mix over the monitors. The received (delayed) overdub part is then remixed and recorded.
Later, the overdubbed part can be slipped back into synchronization with the mix.
Similar to ISDN, the Switched-56 dial-up service provides a bi-directional 56-kbps data channel. The 56k Digital Data Service (DDS) is a dedicated 56-kbps service, with long-distance calls billed by the minute, at prices similar to those of regular long distance calls.
Channel service units (CSU) are used to interface the user's application to a Switched-56 line; they can be purchased or leased. Units are required at both the send and receive ends of the line.
Various commercial systems provide multiple channels of full duplex digital audio connection over reserved T-1 telephone lines. For example, a connection might use Dolby AC-2 data reduction at 256 kbps for dual monaural, with timecode occupying a third channel. Alternatively, a monaural or stereo voice-over service can be provided over Switched-56 and ISDN lines using MPEG data reduction. In addition, video services are available.
Timecode can be conveyed through the system, permitting synchronized overdubs.
Asymmetric Digital Subscriber Line (ADSL)
Companies invest billions of dollars in fiber optics, switchers, and routers to develop high-speed Internet access. But the most stubborn problem is "the last mile"- the actual hookup to the consumer's home. Several methods are used to cost-effectively establish that final link.
Existing telephone modems are too slow. Attention has turned to cable modems connected to CATV (cable TV) lines to take advantage of cable's high bandwidth. Other groups are looking at electric power service lines to move data to and from homes. In addition, small dish satellites, as well as terrestrial wireless delivery, are employed.
Asymmetric Digital Subscriber Line (ADSL) provides high-capacity data channels, along with regular telephone service, on a single pair of copper wires. Like ISDN, ADSL needs two modems (one in the home, and the other at the telephone switching office) but unlike ISDN, ADSL uses existing wires and existing phone numbers. However, ADSL requires the installation of DSLAMs (Digital Subscriber Line Access Multipliers) in telephone switching offices. ADSL was originally developed by telephone companies in 1989 to provide interactive television, and thus bandwidth is its forte. Moreover, unlike dial-up modems that require dialing and connecting waits, ADSL provides a continuous always-on connection. However, as the name "asymmetric" implies, the data rate is not evenly balanced. The rate out of the home is much lower than the incoming rate. Since most applications require much higher input than output, asymmetry is acceptable for most users.
ADSL uses DMT (discrete multi-tone) technology as the coding method. DMT divides the telephone bandwidth into discrete frequency subchannels and simultaneously transmits data over each band. DMT also monitors and analyzes each subchannel, continuously and variably distributing data to each subchannel for optimal data rate.
DMT is part of the ADSL standard, formally known in North America as T1.413. Many systems require a signal splitter, a separate line to the computer, as well as a modem and software. Nonstandard proprietary equipment may be used.
The Universal ADSL Working Group (UAWG) was established to simplify ADSL and lower costs. Its goal was to replace competing incompatible ADSL standards with a simplified consumer ADSL, known as Universal ADSL.
Sometimes called ADSL "lite," it is compatible with standard ADSL but eliminates a splitter and separate wires. The modem is supplied by the telephone company or purchased in retail stores and installed by users as a plug-and-play feature. PCs and other devices may have ADSL lite modems built-in. The lite version may offer relatively low data rates, making high-quality live video problematic, but it is acceptable for other applications.
Whichever method is employed, ADSL competes with cable modems. Both have pros and cons. Both systems are inherently asymmetric. CATV was originally designed to move data in one direction-into homes. Cable modems also suffer from the party-line syndrome in which many homes share a common line back to the head-end. The more users who log on, the slower the traffic. ADSL's growth has been slowed by competing implementations, but cable modems have had a unified standard, called Multimedia Cable Network System (MCNS), for years.
However, although the standard is in place, there are still practical incompatibility issues that remain unsettled. An initiative known as OpenCable may resolve this. Although cable passes by 90% of all U.S. homes, it is only connected to about 60% of them, and not all of these homes have access to cable-modem quality cable.
ADSL is generally not as fast as cable modem hookups.
But ADSL users enjoy a private line to the central switching computer. However, ADSL's speed depends on the distance away from the switching computer; top speeds are possible only for relatively short connecting distances.
Rates are slower for longer distances. Beyond that, service may not be possible. Perhaps 30% of all U.S. homes are either too far from a switching office, or use lines that cannot accommodate ADSL. For all ADSL users, the return path is always relatively slow, and that limits applications. For example, a high-quality video-conference between PC users would require high bandwidth in both directions. Cable companies, telephone companies, and satellite companies (and others) will continue to compete as data access providers to American homes.
Cellular Telecommunications
A number of technologies have been developed for cellular telecommunications. Three representative examples are summarized here. The Group Special Mobile (GSM) system is a Time Division Multiple Access (TDMA) technology that was first developed in Europe and has been adopted worldwide. Frequency Division Duplexing (FDD) is used for both the uplink and downlink. The aggregate data rate is 270.8 kbps and the per-user data rate is 33.85 kbps with eight users per channel. The supported user data rates for full-rate mode are 12.2 kbps and 13 kbps. Gaussian Minimum Shift Keying (GMSK) modulation is used.
The Edge system is an improvement on GSM technology. GSM was primarily designed for CS (circuit switched) communication such as voice and fax. Edge is designed for greater efficiency with PS (packet-switched) communication such as multimedia and other Internet traffic. Edge provides higher capacity and greater data throughput data rate. PS data flow is bursty and application of dynamic resource allocation to user needs is key to good performance; bandwidth on demand is required.
Although Edge retains the channel baud rate and time-slot structure of GSM, it uses 8-PSK modulation and other means to accommodate varying throughout demands. Nine coding schemes are available. The maximum data rate using 8-PSK modulation is 473.6 kbps (59.2 kbps × 8).
Data rate can change by changing the modulation method and the number of parity bits used. A technique called incremental redundancy is used to maximize the data rate.
To determine link quality, a packet with minimal overhead is sent and this high data rate is used if the communication is successful. If not, coding overhead is increased until communication is successful.
Wideband Code Division Multiple Access (CDMA), also known as WCDMA, is a 3G technology. It uses FDD for both uplink and downlink communication. The 3G UTRA and GSM standards are specified by the Third-Generation Partnership Project (3GPP).
Networks and File Transfers
Arguably, the first digital network was installed by Samuel Morse, linking Washington, DC to Baltimore with 35 miles of copper cable. His first transmitted telegraphic message on May 24, 1844, was: "What hath God wrought?" Today, thousands of networks interconnect millions of communications devices worldwide.
Whereas most dedicated audio interfaces operate point-to-point, with continuous data flow, and in real time, computer networks are typically asynchronous, transmit data in discrete packets, and can interconnect many disparate devices. Successful communication across a network calls for arbitration to avoid usage conflicts.
Although a dedicated audio interface uses a dedicated audio format, a network is not concerned with the type of data being transmitted, and uses a common file structure for all data types such as e-mail, graphics, audio, and video. In addition, unlike dedicated audio interfaces, data delivery over a network is usually not continuous, and often not in real time. For example, delivery depends on network data rate as well as on current network traffic; transfer can occur at speeds well below, or well above real time. In some applications, a portion of the network bandwidth can be reserved to allow continuous real-time multimedia exchanges, such as video conferencing. Networked data transfer is increasingly supplanting the traditional practice of hand-carrying physical media, using the "sneaker net." Whereas a dedicated digital audio link provides bandwidth that is sufficient to transmit a program in real time, a network is designed to interconnect multiple devices as networked nodes. For example, personal computers equipped with network interface cards, each with a unique address, can form nodes on a common network. Data is sent when a path is available, at the speed determined by the network interface. File exchange as well as random-access functions among computers are both permitted. The task of conveying high-bandwidth data (one channel of 44.1 kHz, 16-bit samples requires 705.6 kbps) in real time (and over an extended time) without interruption requires special consideration in the network design. In addition, one node may transmit multiple audio channels or a video channel; bandwidth of 30 Mbps to 40 Mbps might be needed. It might be necessary to lock out lower priority transfers, while a high priority time-sensitive transfer is in progress. A typical office LAN will not suffice. On the other hand, a multimedia network allows one major concession- data reduction. Low bit-rate coding of audio and video data can effectively multiply the bandwidth of the network.
Networks can be configured in a variety of ways. Most commonly, a bus configuration places nodes along a serial bus, a ring configuration places nodes along a closed circuit, and a star configuration gives each node direct access to the central controller. A star configuration is preferred because a disabled node does not affect other node performance. A central concentrator unit is needed to monitor and direct bus traffic in a star configuration. When the maximum number of nodes are placed on a network, or the system's longest length is reached, it can be extended with a repeater, a device that receives signals, then resynchronizes and retransmits them. A bridge isolates network segments so that data is only transmitted across the bridge when its destination is another segment. A router is a computer with two network interfaces. It can pass data between different types of networks; moreover, it can optimize the routing for faster communication. A hybrid configuration is sometimes recommended. For example, in addition to a ring joining workstation nodes and a central storage area, nodes might be star-configured to central storage using SCSI . An example of local networks bridged to other local networks, also showing different network configurations, is shown in FIG. 3.
Generally, it is the job of a network to break a message into data packets of uniform size and code them with a destination address and a header that describes where each packet fits within the message. The packets are transmitted, and received where they are assembled into the message. A packet that is corrupted with errors can be quickly retransmitted. Packet-switching is extremely efficient at conveying bursty communication through a distributed computer environment. However, because the arrival speed of packets cannot be guaranteed, real-time transfer is not always possible. All networks must define rules for access to physical storage and terminal devices; this limits bandwidth of the network. Two common control methods are token passing and collision detection. With token passing, when a node finishes a transmission, it sends a token (a bit pattern) to the next node; a node can transmit only when it has the token. Each token ring node has both an input and output, so that connections are passed from one node to another, forming a ring. The ARCnet, operating at 2.5 Mbps uses token passing. A Fiber Distributed Data Interface (FDDI ) network configured in a ring and using token ring can achieve 100 Mbps; FDDI is defined in the ISO 9314 standard. With Carrier Sense Multiple Access with Collision Detection (CSMA/CD), transmission occurs on command; if a collision occurs, priority is assigned, and data is retransmitted. Utilization (measuring successful transmission) of CSMA/CD networks decreases considerably with heavy use, as more collisions occur. A CSMA/CD protocol can achieve an overall efficiency of approximately 40%.
Fig. 3 An example of star, ring, and bus network configurations, connected
through long-distance bridging circuits.
In client-server networks, most applications, data, and the network operating system, are centrally placed on a common file server. In peer-to-peer networks, data is kept locally under the control of individual computers. For example, a user could access files, ROM drives, and a printer-all in different physical locations; any computer can act as the server. Because of the size of audio and multimedia files, emphasis on distributed storage is often most efficient, in which nodes act as peers, each serving as both client and server, each with local storage. A distributed network with modest bandwidth of 10 Mbps to 15 Mbps might be adequate for many applications. For example, this would allow transfer of audio data at a rate several times faster than real time. However, this cannot always be achieved when multiple users are on the network, particularly when relatively slow SCSI hard-drive transfer rates limit data transfer speed. In many applications, data reduction algorithms are used to reduce the size of files, speeding up communications.
A wide area network (WAN) connects multiple stations beyond a single building, reaching up to global distances.
For example, a WAN network can use an ISDN interconnection. A local area network (LAN) distributes data through an office, building, or campus. A LAN can use FDDI , CDDI , or ATM protocols.
Ethernet
Ethernet is a kind of computer network used to connect personal computers, printers, disk drives, and other equipment over coaxial cable, twisted pair, and optical fiber. Ethernet uses asynchronous transmission and collision detection. The following bit rates are often used:
Ethernet at 10 Mbps, Fast Ethernet at 100 Mbps, Gigabit Ethernet ("GigE") at 1000 Mbps and 10 Gigabit Ethernet (10GE) at 10,000 Mbps. Gigabit Ethernet is based on IEEE 802.3 Ethernet and ANSI X3T11 Fibre-Channel.
Generally, because of technical overhead, Ethernet is not used as a facility backbone to carry audio/video files. In many cases, a Storage Area Network (SAN) using Host Bus Adapter (HBA) cards and 1000BaseT or Fibre Channel fiber-optic cable, is a better choice for moving large, multiple files among many storage devices.
An FDDI installation uses duplex (two-conductor) fiber optic cables to permit 2-kilometer runs between nodes, and up to 200 kilometers on a single ring. It uses a token ring arbitration method. The FDDI standard specifies bandwidth of 100 Mbps. FDDI uses a Token Ring protocol. Copper Distributed Data Interface (CDDI) is a copper implementation of FDDI , offering 100 Mbps performance on Class 5 twisted-pair cable. It is more economical than FDDI, but does not provide the immunity to radio frequency (RF) and magnetic interference afforded by optical fiber, and it is limited to 50- to 100-meter runs between nodes.
CDDI and FDDI can be freely mixed using converters.
Fiber-optic technology is discussed in Section. 13.
Generally, unshielded, twisted-pair cable (similar to that used in telephony) is classified in five levels: Level 1 (voice), Level 2 (RS-232 low-speed data), Level 3 (10 Mbps Ethernet), Level 4 (16-Mbps Token Ring), and Level 5 (100-Mbps high speed). For example, a 10Base-T network implementation requires Level 3 unshielded twisted-pair cable. It is used in a star configuration with a multi-port hub; it can be terminated with RJ-45 modular plugs. Daisy-chained Ethernet cabling is specified as 10Base2 (thin coaxial), or RG58 coaxial cable; it is limited to 185-meter segments. It is fitted with 50O? BNC plugs and attached to each node with a BNC T adapter.
10Base5 (thicker coaxial), also known as RG8 or RG11 cable, is used in Ethernet backbones; it is limited to 500 meter segments. Category 5 (Cat5e) cable is satisfactory for short runs of Gigabit Ethernet; Category 6 cable is needed for more sophisticated installations.
Asynchronous Transfer Mode (ATM)
Asynchronous Transfer Mode (ATM) is a high-bandwidth network standard using low-delay switching, variable bandwidth, and multiplexing to provide flexible and efficient communications. Audio, video, and other data can be simultaneously delivered at a rate of 1 Gbps or more. ATM is used primarily in wide area networking and local area network backbones with many users. Traditional networks pool many channels of information, with each receiver picking its information from the stream; considerable routing data is needed, and this increases overhead. The ATM architecture uses switching and multiplexing to form a temporary dedicated, virtual channel within the transmission path bandwidth. The virtual channel is allocated a sufficient data rate, providing bandwidth on demand. Data packets, called cells, carry the name of the channel with much less routing data required. In addition, many ATM users can share a channel through multiplexing, yet maintain a fixed-time relationship between data cells. In this way, slower moving data can be efficiently combined with fast data, with every time slot of the channel packed with data. In particular, this helps ensure continuous data flow, an important criterion for real-time audio and video transfers. ATM is sometimes referred to as Broadband ISDN.
Some implemented ATM channels provide transmission rates of 155 Mbps, 620 Mbps, or 2.4 Gbps along copper or fiber cables. However, data is transmitted at variable data rates, using only the rate needed at that moment. Billing is based, for example, on the number of cells transmitted.
Moreover the data flow can be asymmetric; for example, a remote user can query a database with little data, but receive considerable data in return.
Specifically, ATM transmits data in fixed-length data fields. It uses 53-byte (8-bit) cells each composed of 48 bytes of user (payload) data, a 5-byte routing and control header, and an 8-bit error correction header. ATM switchers can thus route cells efficiently to proper destinations where they are assembled into useful information. In addition, packets are sequentially delivered along the virtual channel. Corrupted data cells are detected by the receiver, which requests a retransmission of cells.
The ATM data protocol is based on the Open Systems Interconnection (OSI ) model that specifies the physical medium and links.
Each ATM header is composed of six data fields. The Generic Flow Control (GFC) controls usage between several terminals with the same access connection. The Virtual Path Identifier (VPI ) designates the path, or bundle of individual channels connecting two points in an ATM network. The virtual channel identifier (VCI ) is a number allocated to every connection on the ATM during the call.
The Payload Type (PT) defines maintenance cells, congestion conditions, and other types of cells. The Cell Loss Priority (CLP) assigns a priority rating to a cell, designating which cells can be discarded if necessary. The Header Error Check (HEC) is a checksum used to detect errors in the 5-byte header.
The AES47 standard describes the transmission of audio over ATM channels. PCM samples can be inserted into ATM cells in a number of ways including time order, channel order, or in multichannel groups. Data in the AES3 format can be conveyed, and there is a provision to carry up to 60 audio channels.
Bluetooth
The Bluetooth specification (BTSIG99) is an open standard describing a protocol used for wireless short-range audio and data communication. The short-range proximity network uses a peer-to-peer method; two devices can establish a link when they come within range of each other.
This forms a wireless personal area network (WPAN).
Radio transmission is used to transmit and receive signals.
Class 1 devices have a maximum permitted power of 100 mW and a range of 100 meters; Class 2 devices have a maximum power of 2.5 mW and a range of 10 meters;
Class 3 devices have a maximum power of 1 mW and a range of 1 meter. Because of relatively low power consumption, Bluetooth is often used in small, battery powered devices. Bluetooth is a worldwide standard and is not constrained by otherwise incompatible telecommunications standards. Moreover, Bluetooth operates in a spectrum that can be used without license worldwide, with certain restrictions. Bluetooth can replace wired connections and is used in mobile phones, MP3 players, car audio receivers, hands-free headsets, printers, keyboards, video game controllers, and other mobile and fixed devices.
When a Bluetooth connection is established, one device is a master and the other is a slave. The device initiating the communication is usually the master, but either device may assume either status. The master device sets up the frequency hopping spread spectrum (FHSS) synchronization and pattern, and controls which devices can transmit. When conveying data, a slave can only transmit in response to a master device. When conveying voice, a slave periodically transmits in its designated time slot. Each device has a 48-bit address. A device can be a master for one link and a slave for another. A master may directly communicate with up to 7 active slaves and up to 255 parked slaves which the master can bring into activity.
All slaves linked to a single master form a piconet; all devices are synchronized. A device may be linked to more than one piconet, establishing synchronization with more than one master; overlapping piconets are called scatternets. In a scatternet, each piconet retains one master and has its own synchronization pattern. A device can act as a bridge, operating as a master in one piconet and a slave in another.
When in active mode, a slave device can continuously receive transmissions from the master device and remain synchronized; response time is fast but power consumption is relatively high. In sniff mode, a slave device is considered active but is only active periodically and the master only transmits packets at certain times. Response time is somewhat slower, but power consumption is also somewhat lower. In hold mode, a slave device is considered active but is inactive for a certain time interval.
Response time is slow, but power consumption is low.
When in parked mode, a slave device is considered inactive on the piconet, but it is synchronized with a master using a beaconing method. Response time is slow, but power consumption can be very low. By placing devices in and out of parked mode, a master can communicate with many devices, while observing the limit of seven active devices. An adaptive transmission power feature lets each slave use a received signal strength indictor to tell a master that its power is not appropriate and that it must cut or boost its transmission power to that slave.
Bluetooth uses frequency division spread spectrum communications, also known as frequency hopping spread spectrum (FHSS). Hopping occurs at a nominal rate of 1600 times per second. Bluetooth operates in the ISM (Industrial, Scientific, Medical) frequency band at 2.4 GHz to 2.4835 GHz. The spectrum is divided into 79 channels;
bandwidth is limited to 1 MHz per channel. Each channel is divided in time (TDMA) into time slots; each slot corresponds to an RF hop frequency. In basic mode, Gaussian frequency shift keying modulation is used.
Version 1.2 enables a gross data rate of 1 Mbps (maximum unidirectional asynchronous bit rate is 723.2 kbps), and Version 2.0+EDR (Enhanced Data Rate) enables 3 Mbps (maximum unidirectional asynchronous bit rate is 2.1 Mbps). Bluetooth data can be conveyed in asynchronous connectionless (ACL) mode and synchronous connection-oriented (SCO) mode. The ACL mode uses packet-switching; headers contain destination addresses. Up to three SCO links are allowed in a Bluetooth channel, each link operating at up to 64 kbps. An SCO link between two devices may be used to convey real time audio. Both synchronous and asynchronous audio channels may operate simultaneously.
Audio in conventional (narrow-band) Bluetooth is conveyed with a sampling frequency of 8 kHz yielding 64 kbps with companding for voice communication. The Advanced Audio Distribution Profile (A2DP) can also be used to convey high-quality audio via Bluetooth ACL mode in either monaural or stereo. For example, music can be streamed from a phone to a car audio receiver. The maximum bit rate is 320 kbps for monaural and 512 kbps for stereo. Use of a SBC subband codec is mandatory.
SBC was specifically designed for use in Bluetooth and features four or eight subbands, adaptive bit allocation, simple adaptive block PCM quantizers, and low complexity.
Block diagrams of an SBC encoder and decoder are shown in FIG. 4. In the SBC encoder, the input PCM signal is split by a polyphase analysis filter bank into subband signals. A scale factor is calculated for each subband, and scale factors are used to calculate bit allocation and levels for each subband. Subband samples are scaled and quantized and a bitstream is output. In the SBC decoder, scale factors are used to calculate bit allocation. The number of quantization levels is derived for each subband, subband samples are calculated, and a polyphase synthesis filter bank is used to output the PCM signal.
Fig.
4 An SBC subband codec is used in the Bluetooth A2DP specification. It can
be used to convey high-quality audio via Bluetooth. A. SBC subband encoder.
B. SBC subband decoder.
Other codecs including MPEG-1/2 Audio, MPEG-2, 4 AAC, and ATRAC can be used optionally when both the source and receiver support the optional codec. When an optional codec is not fully supported, audio data is transcoded into SBC.
The decoder in an SBC receiver must mandatorily support sampling frequencies of 44.1 kHz and 48 kHz. The encoder in the SBC source must mandatorily support at least one of the sampling frequencies of 44.1 kHz or 48 kHz. Lower sampling frequencies are optional. The decoder in the receiver must support monaural, dual channel, stereo, and joint stereo modes. The encoder in the source must support monaural and at least one of the other modes. Both the encoder and decoder must support block lengths of 4, 8, 12, and 16. The decoder must support four and eight subbands, and the encoder must support at least eight subbands. Similar requirements are imposed for other codecs. For example, both the encoder and decoder must support at least one of the MPEG-1/2 Layer I , II , or III codecs. For MPEG-2, both the encoder and decoder must support AAC LC; other object types including those of MPEG-4 are optional. The Audio/Video Control Transport Protocol (AVCTP) can be used to control devices, and the AVCTP and Audio/Video Distribution Transport Protocol (AVDTP) can disseminate audio.
Bluetooth was conceived by Ericsson in 1994, and Version 1.0 was published in July 1999. The specification is publicly available and is royalty-free. The Bluetooth trademark is owned by Ericsson and licensed to adopters of the Bluetooth Special Interest Group; the founding companies of the SIG are Ericsson, Intel, IBM, Nokia, and Toshiba. The Bluetooth specification is described in the IEEE 802.15.1 standard. The specifications for a number of networking technologies are shown in Table. 1.
As noted, Bluetooth uses frequency hopping spread spectrum (FHSS) transmission. This helps overcome interference from other broadcast channels as well as microwave ovens and other interfering devices in the unlicensed 2.4-GHz to 5-GHz band. The spectrum is divided into different channels, each at a different frequency. A message packet is transmitted on a channel, then a new channel is selected for the next packet, and the process repeats; the message is thus spread across the entire spectrum. The receiver uses the same channel pattern to receive each packet. FHSS reduces interference from another device operating at one channel frequency, and collisions rarely occur with other FHSS devices in the same band. A single packet loss can be retransmitted at a new frequency. Adaptive hopping can be used to avoid an interfering frequency, and more power can be delivered to any particular broadcast frequency. The broadcast signal appears as noise-like impulsive interference to a narrow band receiver.
Table 1 Specifications for several networking technologies.
The invention of frequency hopping is often attribute actress Hedy Lamarr and musician George Antheil. T 1942 patent (2,292,387) proposed a "secret communication system" to guide torpedoes by radio.
Lamarr proposed that the signal be switched from one frequency to another, and Antheil used his expertise w player pianos to devise a synchronization method. FH also provides some security against eavesdropping; o the receiver knows the hopping pattern.
Direct sequence spread spectrum (DSSS) techniques are used to code data into patterns of bits, and portions of redundant data are delivered to different carrier frequencies in a wideband. A spreading ratio as defined by the level of redundancy is a factor that determines the system's robustness.
IEEE 802.11 Wireless LAN (Wi-Fi)
Increasingly, data that was once confined to copper wires or optical fiber is being conveyed wirelessly via radio. Wi Fi is used for wireless LAN (WLAN) applications. Wi-Fi technology is described in the IEEE 802.11 standard. Four standards are widely used: 802.11a, b, g, and n. A Wi-Fi system establishes a local area in which devices such as cell phones, computers, and printers can wirelessly communicate with each other. Wi-Fi is used to convey audio data between computers and peripherals, as well as audio playback devices. Wi-Fi is sometimes known as wireless Ethernet because it shares some parts of its specification with Ethernet (IEEE 802.3). For example, Wi Fi uses carrier detection and collision avoidance, as well as authentication and encryption.
Some terminology: a WPAN is a wireless personal area network, usually operating over personal distances. A WLAN is a wireless local area network operating over longer distances. A hot spot, for example, can encompass a home, office, or school. A WMAN is a wireless metropolitan area network operating over a large area. A WWAN is a wireless wide area network, potentially providing a global reach, for example, a cellular telephone system such as the General Packet Radio Service (GPRS).
Table 2 --- Wi-Fi standards and specifications.
Specifications for the four Wi-Fi standards, IEEE 802.11a, b, g, and n, are summarized in Table. 15.2. They may differ in terms of operating frequency band and data transfer rate. In addition, they provide different levels of performance and various features and security enhancements. For example, an 802.11b system might provide a theoretical maximum bit rate of 11 Mbps over a 300-meter range, while an 802.11a system might provide a bit rate of 54 Mbps over 50 meters. Actual user data throughput is much less. Lower bit rate 802.11a systems use FHSS and DSSS techniques, and higher bit rate 802.11b systems use only DSSS. Higher transmission frequencies allow higher bit rates, but that generally dictates a shorter operating range.
The 802.11a standard operates in the 5.15-5.35-GHz and 5.725-5.825-GHz bands. It uses orthogonal frequency division multiplexing (OFDM) with 52 subcarriers (48 main plus 4 pilot) using Binary Phase Shift Keying/Quadrature Phase Shift Keying (BPSK/QPSK), and 16-Quadrature Amplitude Modulation (16-QAM) for higher bit rates, and Forward Error Correction (FEC) convolutional coding. A variety of data rates are supported: 6, 9, 12, 18, 24, 36, 48, and 54 Mbps; rates of 6, 12, and 24 Mbps are mandatory.
Different data rates are supported in the 802.11b standard. The basic rates of 1 Mbps and 2 Mbps use a DSSS system; extensions allow rates of 5.5 Mbps and 11 Mbps in which complementary code keying (CCK) modulation is used. Some worldwide frequency bands are 2.4-2.4835 GHz ( United States, Canada, Europe), 2.471- 2.497 GHz ( Japan), 2.4465-2.4835 GHz (France), and 2.445-2.475 GHz ( Spain).
The 802.11g standard provides data rate extensions in the 2.4-GHz frequency band. Supported data rates are 1, 2, 5.5, 6, 9, 11, 12, 18, 24, 36, 48, and 54 Mbps; rates of 1, 2, 5.5, 6, 9, 11, 12, and 24 Mbps are mandatory. DSSS/CCK and OFDM are supported. 802.11g hardware is backward compatible with 802.11b hardware.
The 802.11n standard uses four spatial streams with a channel width of 40 MHz to provide a significant increase in overall data rate. A multiple-input multiple-output feature employs multiple antennas and Spatial Division Multiplexing (SDM). Wi-Fi standards are developed, and Wi-Fi products are certified, by the Wi-Fi Alliance trade group.
MediaNet
The MediaNet is an example of a dedicated high-speed multimedia LAN network. It lets multiple users simultaneously access materials on central disk drives.
MediaNet is implemented on CDDI and FDDI protocols, and supports the Apple Filing Protocol, Networked File System, and ATM. Using CDDI , the network can simultaneously handle multiple channels of compressed audio and video with throughputs of 24 Mbps from node to node. With ATM, transfer rates of 120 Mbps can be achieved. Computers can be interfaced either as servers or clients; both have SCSI controllers and ports for connection to hard-disk drives. Disk drives are local in node Apple computers; however, any node can access data on any other node hard drive in the same way that one would access a printer or other network device. Token ring is used for data traffic control. Using such networks, multimedia data can be directly manipulated remotely without copying files from place to place. MediaNet was developed by Sonic Solutions.