United States Department of Defense Design Criteria Standard: Human Engineering, better known as
MIL-STD 1472 has a fairly complete section (5.3) on audio signals. A quick reading gives the following standards that seem to apply directly to your question (edited here):
220.127.116.11 Meaning of the signal. Each audio signal must have only one meaning.
18.104.22.168 Apparent urgency. The attention gains characteristics of the signals in a set (for example, the speed of the pulse pattern, the frequency,
intensity) must correspond to the relative priority of the signal.
22.214.171.124.1 Range. The frequency range must be between 200 and 5000 Hz and, if possible, between 500 and 3000 Hz.
126.96.36.199.1 Attention and prevention of the surprise reaction. To minimize surprise reactions, increase the sound level for 0.5 seconds
period should not be greater than 30 dB. In addition, the first 0.2
dry of a signal should not be presented at the maximum intensity, use
square top waveforms, or abruptly rising rising waveforms
188.8.131.52.1 Use of different characteristics. When several different audio signals must be used to warn an operator of different types
of conditions, discriminable differences of intensity, height, beats
and harmonic, or temporal patterns must be provided. So absolute
discrimination is required, the number of signals to be identified
must not exceed four. Signal strength should not be used alone
means for discriminating the signals.
It describes in detail the characteristics of warning signals, but remember that the military defines "warning" as "you must react". now where you will die Everything will not apply to you, I do not think so.
It also tells you if you need to have audio signals and what commands to provide to users.
The latest Windows UX interaction guidelines also provide instructions on the use of sound (p686 in the Windows 7 PDF version). They point out that all sounds provide only additional information, namely that users should never need to rely solely on sound for their comments. Thus, all sounds, including those of errors and warnings, must be brief, subtle and calm, if present.
Use sound only when there is a distinct advantage. If in doubt, do not
use the sound … The ideal sound effect is the one that users hardly notice,
but they would miss it if he was absent.
As for the quality of each sound, it's more of art than science. Windows guidelines recommend:
Clear and pure sounds, as well as glassy and airy sounds, with a progressive fade and
fade-out ("soft" edges) to avoid sudden, jerky, percussion
effects. They are designed to feel subtle, sweet and consonant.
Windows sounds use echo, reverb, and EQ to get
The single beeps and buzzes of the computers of yesterday are well out. You must exploit all the advanced sound reproduction features of modern systems. This means that you have to create your sounds as you create a piece of music with one or two notes. If so, you may need to hire an experienced musician in digital synthesizers, describe the information and effects you want to communicate, and allow him to create candidate sounds.