Audio networking using commodity hardware
This is an idea proposed in 2004 as a Cambridge Computer Science Part II project, and has been completed by
Many everyday electronic devices such as laptops, PDAs and telephones are able to play and record audio. Invariably this ability is used to interface with people, playing or recording music, speech or sound effects. However, it is just as possible for the devices to communicate with each other, exchanging data encoded into sounds. This project aims to allow computers, separated by short distances within a room to communicate by transmitting sounds between them.
In certain ways, audio based data communication already occurs. The prime example, the modem, encodes digital data into sounds suitable for transmission over a telephone line. Similarly, telephones communicate with exchanges using tones to encode telephone numbers. Historically, old computers would sometimes store data on standard audio cassettes.
The techniques used in modems will likely be a useful point of reference; however the challenges of sending audio reliably across a room are a little different. Telephone lines have fairly well known bandwidths and noise characteristics, whereas rooms vary greatly in their acoustics and hence the distortions they apply to sound. Additionally, the variety of noise that must be dealt with will be larger, since people and machinery may also be present. Therefore different encoding schemes may be needed.
Audio networking has some important differences when compared with other wireless networking technologies. Radio based systems such as 802.11 and to some extent Bluetooth will radiate throughout a building, permeating walls and floors. Because buildings are usually architected with soundproofing in mind, audio signals are much better confined to a particular room. Ignoring glass windows, such confinement is also applicable to infrared signals. However, line of sight is normally needed for infrared, whilst sound can be more omnidirectional.
The most obvious difference to a user of audio networking will be that they are able to perceive it themselves - assuming that the audible frequency range and not ultrasound is used. Since noise is objectionable to humans, causing headaches, irritation and lowering productivity, it would be imprudent to use audio for continuous data transmission. Short bursts of communication however should not be too distracting, particularly if initiated at the user's request.
Since audio networking will have a much lower capacity than existing systems, the most effective use will exploit its architecturally limited range and human awareness. Such a use could be a system for portable computers to discover their location within a building. If there were a fixed computer in a room which knew its location, then portable systems could configure themselves by sending an audible signal and receiving a reply from the fixed station. Such a system could for example be used to obtain information on the nearest printer or the configuration for the room's 802.11 wireless network.
See also