The glossary pages provide definitions for over 2680 PA-related terms. If you can't find the term you are looking for, or would like any of the existing definitions to be expanded, please email me − likewise of course if you find any errors in the links etc. Use of this information is conditional upon acceptance of the Disclaimer on the PAforMusic home page.
In the list below, the most commonly looked-up terms are in bold, lighting-specific terms are in pink, and video-specific terms are in orange.
Q * QTBMC * Quad cable * Quality factor * Quantisation * Quantisation distortion * Quantisation noise * Quarter track * Quasi-balanced * Quasi-floating * Quasi-parametric equaliser * Quick-connect terminal * Quiescent current * Quiet
The definitions for these terms are given on the assumption of their use in the context of PA systems; many of the terms have more general meanings when used in a wider context. Where more than one definition is given for a term, the definitions are numbered (1), (2) etc.
Some of the definitions themselves use terms (such as "signal") in a specific way − most of these are links (just the first time they are used, in each definition), so just click on them to see the meanings that are intended.
The term 'Q' is an abbreviation of 'quality factor', and derives from the fact that the bandwidth of early equalisers, which used a passive design, was reduced by the use of better quality inductors, which have lower losses (see Q (2)).
Numerically, the value of Q is given by the centre frequency divided by the bandwidth. If we consider a graphic equaliser, we can assume that the bandwidth of each band runs from a point logarithmically midway between its own centre frequency and the centre frequency of the band below, to a point logarithmically midway between its own centre frequency and the centre frequency of the band above. On this assumption we can calculate that the approximate Q value for the bands of an octave equaliser is 1.4, the Q value for a 2/3 octave equaliser is 2.1, and the Q value for a 1/3 octave equaliser is 4.3. (The values for 2-octave and half-octave types would be 0.7 and 2.9 respectively.) So, the range of Q values that can be set by the Q control of a parametric equaliser would typically be 0.5 to 8. See also Proportional Q and Constant Q.
Note that the effective series resistance is not just the resistance of the conductor that forms the inductor, but also includes magnetic losses (which in the case of cored inductors can be considerable at high frequencies). Numerically, the value of Q is given by the reactance divided by the effective series resistance.
Q values of this kind that are specified for a speaker or driver are more properly designated the Qes (electrical Q), Qms (mechanical Q) and Qts (total Q). Note that these values of Q are completely unrelated to the directivity Q (see Q (4)) value for the speaker or driver. See also Resonant frequency.
Assuming linear quantisation, the level of quantisation noise voltage is halved for every additional bit employed for each sample. This means that for 16-bit sampling its level is 96 dB below full scale level (0 dB FS). So, 16-bit sampling is sometimes quoted as providing a 96 dB signal-to-noise ratio (SNR), but this is somewhat misleading on two counts. Firstly, it is only a theoretical optimal figure as it takes account only of quantisation noise, whereas in practice there will also be other sources of noise acting to reduce the dynamic range. This ambiguity is corrected by use of the term 'signal to quantisation noise ratio' (SQNR). Secondly, both these terms are comparing average signal level with noise level, and since the average signal level will be set somewhat below the 0 dB FS level in order to provide headroom, the actual SQNR will be less than 96 dB by that amount.
What about '96 dB dynamic range' − is that a better description of a 16-bit system? Well firstly, like '96 dB SNR' it takes account only of quantisation noise, whereas in practice there will also be other sources of noise acting to reduce the dynamic range. Secondly, it implies that a signal at −96 dB FS can be accommodated by the system, albeit at the same level as the noise. Whilst an analogue signal of that level at the input of the analogue to digital converter would indeed be at a level equivalent to the level of quantisation noise about to be introduced, a signal of such a level cannot reasonably be considered to exist in meaningful form in the digital domain, as it would have only one least-significant bit's worth of digital amplitude to represent it and therefore would not be recognisable as the original signal when eventually converted back to analogue. So it is very dubious indeed to imply that the system can handle signals of any level between −96 dB FS and 0 dB FS, in the same way that a purely analogue system with a 96 dB dynamic range can.
Quantisation noise may sometimes be referred to as 'quantisation distortion', but is not a form of distortion in the true sense. Note that the American spelling is 'quantization noise'.
This behaviour has two benefits:
However, some types of quasi-floating output can become unstable or provide an excessively noisy signal if no connection is made to either its 'hot' or 'cold' legs, and the level appearing between the other leg and earth may not be as expected under these conditions. If connecting this type of output to an unbalanced input, it is therefore generally advisable for the unused leg (normally the 'cold' one) to be intentionally shorted to earth, as described above. (Though note that such intentional shorting is generally not advisable with other types of output.)
In the event of a disconnection fault occuring on one of the legs of an interconnection, the level provided on the other leg by this type of output should (with a well-designed output circuit) be unaffected, and so the level received by a balanced input of the destination equipment should fall by 6 dB (just as in the case of a conventional fully balanced output). However, as mentioned above, other unexpected results may be obtained under these conditions, notably a substantially increased noise level. See also One-legged.
A table comparing the most common types of balanced interconnections is provided under the 'Balanced' entry. Diagrams illustrating various different types of signal interconnections are available here (opens in a new window). Compare Balanced, Semi-balanced, Ground-compensated and Pseudo-balanced.
There are no more definitions on this page. (The space below is to facilitate linking to the last few terms above.)
This page last updated 05-Sep-2017.