Analogue audio signal levels have historically been categorised for professional usage in two groups – microphone level (low) and line level (high). Domestic equipment generally falls into a third category, just below professional line level. Signal levels are measured using units based on the Decibel. There are many different types of Level Meter used for signal measurement and monitoring.
Decibels specify ratios, not absolute levels. To specify an absolute level, we state the ratio of the quantity to be specified referred to a standard reference value. For example, signal power in electronic circuits is usually referred to one milliwatt. The level would then be given as so many decibels re: 1 mW.
In all formal writing, and wherever there could possibly be any doubt, it is essential to give the reference quantity. This may be done by putting the reference quantity in brackets. For example, where the reference is 1 mW, the level would be given as so many dB(mW).
The abbreviation dB is commonly modified by adding an extra letter to indicate the reference value: for example dBm indicates that the reference level is 1 mW. Such abbreviations are not recommended by the AES and IEC but are likely to remain in use for some time.
The abbreviations dBA and dB(A) do not indicate a reference value but a frequency weighting: see under Sound Levels.
In the early days of broadcast audio, level standards were derived from the telephone industry. At that time, the standard signal level was referred to as 0 dBm, and this was the amount of signal required to dissipate 1 milliwatt of power in a 600 ohm termination (600 ohms send and receive impedances in equipment was also a standard of the time).
The ‘m’ in 0 dBm indicates that the measurement was made to this standard (ie. 600 ohm terminations in place). If you do the sums, you will find that the voltage needed across a 600 ohm load to produce the 1 mW of power required is 0.775 of a volt.
Today, we generally don’t use 600 ohm impedances any more in professional audio since we are not interested in the efficient transmission of power. Instead, we use low output impedances and high input impedances (typically 100 ohms or less, and 10,000 ohms or more, respectively). However we still use the same signal voltage as the original dBm reference level – ie. 0.775 V.
To differentiate between this new standard and the old (matched 600 ohm impedance) one, we now use the suffix ‘u’ – hence the dBu. In essence, then, the dBu means that our reference signal is 0.775 V irrespective of impedance.
The standard Alignment Level employed by UK broadcasters is 0 dBu, shown on peak programme meters as PPM4, and on standard VUs as 0VU (or -4VU in some establishments). However +4 dBu is a common operating level in many music studios and professional equipment is often aligned to this standard.
Domestic/consumer equipment is usually built with a much lower reference signal level which is normally defined as -10 dBV. In this case the V indicates that the reference signal is 1 Volt. Again, if you do the sums you discover that -10 dBV implies a signal of 0.316 Volts. If you convert this into (professional) dBu terms, it equates to -7.78 dBu.
In other words, the output of a consumer device working to the -10 dBV standard is going to be roughly 8 dB or 12 dB below the nominal level expected of a professional systems.
In digital equipment, the only defining point is the level at which you run out of quantising levels – the digital overload point. All digital equipment defines this as 0 dBFS, where FS stands for Full Scale.
For practical usage, we need to define a working headroom to allow for normal and unexpected peak levels (bearing in mind that PPMs deliberately don’t show fast transients), and so various standards bodies have specified alternative Alignment Levels between -20 and -12 dBFS.
In Europe, the EBU recommend that -18 dBFS equates to the Alignment Level, and for UK broadcasters, Alignment Level is taken as 0 dBu (PPM4 or -4VU). This works well with original recordings and allows a reasonable safety margin against overloads. In a post-production environment, where levels are likely to be much better controlled (through manual or automatic means), many organisations use a higher reference level. One common alternative is to use -12 dBFS which takes advantage of the fact that less ‘contingency’ headroom is required, and trades that for an improvement in the noise performance of the medium.
Another approach, often taken with systems offering 24 bit wordlengths, is to provide an increased headroom margin while maintaining the same (or better) noise performance of a 16 bit system. Typically, people are using -20 or -22 dBFS in this situation. The American SMPTE standard defines -20 dBFS as the Alignment Level for all systems anyway, regardless of digital wordlength.
In August 2010 the EBU made Recommendation 128 in respect of “Loudness normalisation and permitted maximum level of audio signals”, introducing the term LUFS (Loudness Unit, referenced to Full Scale).
This states that “In addition to the average loudness of a programme (‘Programme Loudness’) the EBU recommends that the descriptors ‘Loudness Range’ and ‘Maximum True Peak Level’ be used for the normalisation of audio signals, and to comply with the technical limits of the complete signal chain as well as the aesthetic needs of each programme/station depending on the genre(s) and the target audience.”