It was originally developed in 1938 by the BBC as a superior alternative to earlier types of meter, which were not much use for monitoring peak audio levels. Despite the name though, the PPM is actually a quasi-peak meter with quite a slow integration time of around 10 ms. This was deliberate, the intention being that while the PPM could be used to avoid overload caused by peaks, it would ignore fast peaks because the ear is tolerant of distortion lasting a few milliseconds.
There are nowadays many types of peak and quasi-peak indicating audio level meter, but this article confines itself to the BBC designs and their derivatives. Differences from UK practices are listed under National variants below.
The BBC used a number of methods of measuring programme volume in its early years, including the 'volume indicator' and 'slide-back voltmeter'.
By 1932, when the BBC moved to purpose-built facilities in Broadcasting House, the first audio meter to be known as a 'programme meter' was introduced. It was developed by Charles Holt-Smith of the Research Department and became known as the 'Smith meter'. This was the first meter with white markings on a black background. It was driven by a circuit that gave a roughly logarithmic transfer characteristic, so it could be calibrated in decibels. The overall characteristics were the product of the driver circuit and the movement's ballistics.
The first of the PPMs was designed by C. G. Mayo, also of the BBC's Research Department. It came into service in 1938. It kept the Smith meter's logarithmic, white-on-black display, and included all the key design features that are still used to this day with only slight modification: full-wave rectification, fast attack and slow decay time constants, and a simple scale calibrated from 1 to 7.
The time constants were determined after a series of experiments. At first it was intended to create a true peak meter to prevent transmitters from exceeding 100% modulation. A prototype meter was created with an integration time of about 1 ms. It was found that the ear is tolerant of distortion lasting a few ms, and that a 'registration time' of 4 ms would suffice. The return time had to be a compromise between a rapid return, which was tiring to the eye, and a slow return, which made control difficult. It was decided that the meter should take between 2s and 3s to drop back 26 dB.
|Specification||a.k.a.||Scale marks||Reference level||Attack time||Decay time|
|IEC 60268-10 / I||Nordic||−36, −30, −24, −18, −12, −6, TEST, +6, +9||TEST = 0 dBu||80% in 5 ms||20 dB in 1.5 s|
|IEC 60268-10 / IIa||BBC||1, 2, 3, 4, 5, 6, 7||4 = 0 dBu||80% in 10 ms||24 dB in 2.8 s|
|IEC 60268-10 / IIb||EBU||−12, −8, −4, TEST, +4, +8, +12||TEST = 0 dBu||80% in 10 ms||24 dB in 2.8 s|
|ANSI C 16.5|
|VU meter||−20 to +3 dB||0VU = +4 dBu||99% in 300 ms||99% in 300 ms|
BBC installations used the M3 standard until 1999 in which the M and S needles indicated 3 dB higher. This was chosen as a compromise to ensure mono compatibility when stereo programmes were listened to on mono receivers, which pick up only the M signal. The BBC has now begun to use the M6 Standard, but this has not been rolled out universally - many parts of the corporation still use 'traditional' M3.
With the M6 standard, a widely-panned sound that peaked to its maximum 6 on, say, the A (left) channel would peak only 4.5 on M. The same sound in the centre of the stereo image would peak 6 on M. This is a 6 dB variation for the mono listener.
With M0, i.e. simple M = A + B, a widely-panned sound peaking 6 would peak to 6 on M, but when panned centre would need to peak just 4.5 on each channel to keep M at 6. This is a 6 dB variation for the stereo listener.
With M3, any variations as sounds are panned are kept to 3 dB. Moreover, for most non-phase-coherent stereo sounds, the sum of the two channel voltages averages 3 dB (the full 6 dB sum is only achieved by exactly in-phase signals, i.e. a mono signal panned centre), so with M = A + B – 3 dB, most stereo sounds are a 'good fit' to the maximum permissible signal (PPM 6) on M, A and B.
The sensitivity of the S indication can be increased on some meter installations by 20 dB; this is to aid line-up procedures, e.g. of stereo mic pairs, or the azimuth of analogue tape machine heads, which rely on cancellation of the S signal.
Although the PPM is preferred by many engineers to the much slower VU meter used in the USA, it does require some interpretation in use. Although it gives a precise meaningful indication of overload it cannot be assumed to represent subjective loudness. The BBC have tables showing recommended settings for different types of programme, such as speech, classical music etc, which attempt to take account of this fact.
Regardless of the kind of programme, 6 on the meter is nearly always considered the nominal maximum level, with anything more considered an overload.