Mysteries Of Metering

All mechanical meters are VU meters, all bargraph meters read peak levels — and both types will give the same reading if you feed in a test tone. Reasonable enough assumptions, but wrong on all counts, as Paul White explains.

The really wonderful thing about standards is that there are so many of them, and nowhere is this more evident than when you look at metering. This article examines the complicated issue of metering standards, but those unfamiliar with the general terminology of metering (eg. dBu, dbv, and the conventions of 'plus 4' and 'minus 10' operation) are advised to check out my article from SOS February 1994, 'dBs Explained', which should clarify many of the terms used here.

Tape machines have meters, mixers have meters and signal processors have meters, but what do they actually tell you? To take the last point first, most meters are designed to tell you when the piece of equipment to which they belong is being fed the correct signal level. Metering is vitally important, because all electronic devices have a lower signal limit (where the signal is so small it is overpowered by the circuit noise), and an upper limit where the signal reaches the unit's maximum level — whereupon clipping occurs. By using a meter properly, you can choose a signal level which is as high as possible without clipping, which will produce the best possible signal‑to‑noise ratio.

The Needle & The Damage Done

The first type of meter built specifically for audio use was the VU meter, VU standing for Volume Units. The idea was to build a meter that would produce a reading similar to the loudness or volume level perceived by the listener. The way the human ear hears sound is that very short‑duration sounds appear quieter than longer bursts of sound at the same level. Moving coil meters can be built to simulate this characteristic pretty well, because the inertia of the mechanism limits the speed at which the meter can respond to transients. Put a drum beat into a VU meter, and the meter will barely have begun its climb than the beat will have ended, and the meters start back down again.

VU meters measure the RMS (root mean square) value of the input voltage: a sine wave that alternates between plus and minus 1 Volt, peak to peak, will actually produce a reading on a voltmeter of 0.707 volts, which is what you'd get if the voltage in the sine wave were averaged out into a steady DC voltage. Because the dBu scale used for audio is also an RMS‑based scale, steady sine waves or test tones should result in complete agreement between the VU value and the dBu value. For example, a mixer designed to operate at a nominal +4dBu should be outputting exactly +4dBu when the VU meters read 0dB, providing the input is a steady sine wave tone. One exception to this is to be found on some of the newer Mackie mixers, where they have decided to make the VU meter read 0VU for 0dBu. This means that you can use the mixer at either +4 or ‑10, and the meters will always tell you the actual signal level at the output — a practical and sensible option.

VU meters work fine with analogue tape, because analogue tape has quite a lot of headroom above its nominal operating level, during which the level of distortion increases progressively, unlike digital systems which merely clip. For this very reason, when used with digital systems, VU stands for Virtually Useless, because the peak levels produced by something like a drum kit could be driving the digital recorder into clipping while the VU meter is reading around‑10dB or less.

One myth it is important to dispel is that only moving coil meters are VUs: you can also have bargraph VU meters, because the characteristics of a bargraph meter depend entirely on the circuitry driving them. A line of LEDs has no mechanical inertia — so in theory, they can be made to respond as fast or as slow as you like.

PPM

Peak Programme or PPM meters are more in keeping with the digital age, because they are designed to respond fast enough to show any signal peaks that might cause distortion. Some also incorporate a peak hold facility, where the highest peak levels are displayed for several seconds to make sure you don't miss them. However, they still don't read absolute peak values, because clipped peaks shorter than a millisecond or so are generally inaudible. Unlike the VU meter which reads an RMS or average value, the PPM reads the voltage between the negative and positive signal peaks — which explains why you don't see the same reading when a steady sine wave is fed into a VU meter and a PPM meter.

In fact, if you take an EBU standard PPM meter, it will read 8dB higher than a VU meter monitoring the same signal in the same system. This difference equates to the difference between a reading of 0.707 Volts and 2 Volts, the peak‑to‑peak reading you'd get from a +/‑ 1V sine wave. Of course, in real life, you can calibrate a meter to read anything you like, and the BBC have theirs calibrated so that a 0dBu sine wave reads +6dB on their PPMs. Some standards go one further, and rationalise that both types of meter should read the same, so some European and Scandinavian PPMs may read exactly the same as a VU meter for a steady sine wave input.

As if that wasn't confusing enough, you can also have moving‑coil PPM meters. All you have to do is design the drive electronics to hold the peak level until the meter has had time to respond, and you've cracked it. The peaks may register a fraction late, but they'll still register.

Digital Multitrack

A potentially confusing situation arises when using analogue mixers with digital multitrack machines, because the meters on the mixer and multitrack don't match up, not even when you put in a steady state test tone. Even if your mixer has true PPM meters, the chances are that the levels still won't match. Why?

I've spoken to several different people about this, and they all come up with slightly different answers, but as a rule, digital multitracks are calibrated so that a 0dB test tone (measured either VU or PPM) coming out of a mixer will read several dBs below 0dB (clipping) on the digital multitrack. This makes a lot of sense, because mixers are designed to be driven 'into the red', and if you have a model with VU meters, you could be a lot further into the red than you imagine.

Digital machines won't tolerate any overload unless the period of clipping is so brief that you can't hear it, so calibrating the input electronics in this way is one way of helping the user stay out of trouble. It also means that the mixer can be driven a little way into the red, as normal, to make the most of the available headroom. Similarly, when the signal comes back from a digital tape recorder, it's often hotter than you expect, for exactly the same reasons. That means when you're mixing from Alesis ADAT or Tascam DA88, you might find your mixer once again flicking into the red, even though the meter readings on the multitrack are below 0dB.

On my desk, which has moving coil meters and is calibrated to run at +4dBu, a 0VU test tone reads around ‑15dB on my ADAT. Even allowing for the difference of 8dB between PPM and VU metering, this still leaves around 7dB of artificially introduced headroom. To put it another way, your digital machine won't clip until the mixer output exceeds +7dB PPM.

Use Your Head (Room)

When recording to analogue, the meters can only give you a rough guide as to what the right recording level should be, and if they're VU meters, the readings will be different for percussive material and music with more sustained sounds. Add this to the fact that modern tape can often accept a lot more level before saturating than older types, and it soon becomes apparent that the only way you can really define the limits is to make a few test recordings at different levels, to find at what point distortion becomes audible. After a little experience, you get used to what to expect from a VU meter with different types of input material, but as far as I'm concerned, reading a VU meter is still as much an art as it is a science!