Sound

The Use of Sound

Sound can be used to convey information where appropriate. That might mean using sound directly to communicate voice information or using music or sound effects to support and enhance information communicated by other means. We need a more developed language for the discussion of sound issues, however, because, without it, the provision of sound will continue to occupy a much lower status in planning that visual elements in a design.

Sound quality issues deserve as much attention as the quality of the graphics, in view of the fact that user's usually take CD-quality audio for granted in their home listening habits. Often programs are let down by poor sound quality. Of course, the trade-off in sound quality is file size: the higher the playback sound quality, the greater the file size, and the less of it that can be stored on a limited-capacity storage medium such as a CD-ROM. And, of course, choosing mono rather than stereo sound halves file sizes. In general, a CD-ROM can store up to

72mins of CD-DA (audio-CD) quality sound, that is, stereo sound sampled at 44kHz and 16bit (1 min stereo sound = approx. 9MB; 1min mono = 5MB).
Vinyl LP quality sound is about 22kHz at 16bit (1 min mono sound = approx. 2.5MB).
FM radio quality is about 22kHz and 8bit (1 min mono sound = approx. 1.2MB).
11kHz at 8bit is just adequate for speech but can distort sybilants ("S"-sounds), causing the speaker to sound as though they lisp (1 min sound = approx. 0.6MB).

Lower sampling rates are generally not useful for multimedia purposes.

Note that in the production of sound tracks on a computer, it is important not to rely on the computer's in-built speaker for feedback during the editing process. Background noise/hiss may be undetectable through the in-built speaker which otherwise is apparent when the sound is amplified or played on better-quality, dedicated sound equipment. Rather, quality should be monitored using the best-quality sound reproduction facilities available and at least by means of quality headphones or amplified speaker.

Programs can also be let down by voice talent. It is nearly always better to hire professional voice talent for voiceovers than to casually trawl for volunteers. If a particular celebrity is required then they are at liberty to charge what the market can bear, but in general the rates charged by voice-talent actors are not very high. Voices should be selected bearing in mind possible likes and dislikes of the potential user audience and this should be considered in the user-needs analysis in the design stages of the project.

Recorded human voice is generally preferable to computer-synthetic voice, unless the circumstances dictate otherwise. This is bound to remain the case until synthetic voice programs are able to pick up on the rule-bending nuances of speech which are so important to meaning and communication.

If possible, check the effect of the actual ambient conditions on sound from the system. It is important to control sound levels judiciously, allowing playback sound levels to be altered according to ambient conditions. It is not just the user and passer-by who is affected by sound from an interactive display. Repetitive sound from interactive systems in public places can cause annoyance to staff working nearby. This has resulted in systems being sabotaged by staff. Whether volume changes should be left to a supervisor or whether the user can freely adjust the sound should be considered. Whether to allow access to sound controls through screen software or by direct access to a hardware sound control should also be considered.