If you're serious about recording and mixing you need to set a consistent reference level to which you can always return. Our seven‑step guide explains how to do it.
Setting up and maintaining optimal signal levels through an audio chain, from the microphone preamp right through to the DAW, brings many important practical and sonic benefits, as Matt Houghton explored in his article about gain staging in your DAW (SOS September 2013, http://sosm.ag/sep13-gainstagingDAW). There's also a growing movement to set levels according to the material's overall loudness, rather than the signal peaks, as I described in my feature on the BS.1770 loudness standards recently (SOS February 2014, http://sosm.ag/feb14-endofloudness). In both cases, it makes a lot of sense to extend these gain staging and loudness concepts all the way to the monitors, the listening environment, and ultimately our own ears.
In essence, what I'm talking about is establishing a reference listening level in the control room, to use as a reliable base from which to make aural decisions. The good news is that this is very simple to achieve, and doesn't require any major investment in new equipment, although it's certainly easier and more intuitive if you have a well‑designed monitor controller and a simple SPL meter.
Gain staging is all about setting a reference signal level — the operating level — through both hardware and software signal paths, which provides a sensible headroom margin above while simultaneously ensuring signals remain well clear of the system noise floor below.
This headroom margin (typically 20dB in professional systems) helps to avoid accidental clipping and unplanned distortions — and it is frequently overlooked! In fact, one of the most common complaints about digital working, the 'harsh sound of digital', is often caused by working with a negligible headroom margin in the DAW. What that means is that the analogue front end and the analogue monitoring chain electronics are all operating with much higher average signal levels than their designers intended, and are often on the verge of clipping. So it's no wonder they can sound a bit strained!
From a practical recording technique perspective, working with sensible headroom margins makes tracking far less fraught (because the risk of clipping is removed), so you can concentrate on performance rather than level watching. It also makes mixing much easier for similar reasons, but also because many plug‑ins, especially vintage emulations, simply sound better when used with typical analogue levels and headroom margins.
The nominal operating level we establish with our headroom margin can usefully be translated through to a reference listening level from the control room's speakers too. Our ears quickly become used to a 'standard' volume, and we can then judge levels, loudness and dynamics by ear — and surprisingly accurately too. In that way, mixes end up being more consistent, and working to the new loudness standards becomes easier because we know intuitively when something is 'too loud' or 'too quiet'.
This concept was embraced in the cinema industry back in the late 1970s, with a rigorously enforced standard reference acoustic level in all dubbing theatres, and is mirrored in the presentation cinemas. In the broadcast world, TV sound balancers now working in BS.1770‑compliant facilities are also finding that by adopting a sensible reference monitoring level at the start of the day, their mixes fall in line with the R‑128 requirements almost automatically, again because their ears naturally tell them when things are too loud or too quiet.
In short, establishing a stable, reference monitoring level — one which can be returned to instantly and reliably, and which conforms with the reference monitoring level of other facilities — is a great aid in maintaining consistency between different mixes, or when a mixing project evolves over many days.
Another aspect of setting a reference speaker level is to work in an optimal part of the Fletcher‑Munson equal‑loudness curves of the human ear. The idea is to provide a consistently uniform frequency response and minimise problems with too much or too little perceived bass. However, while there is some validity to this, we must also be careful because of the psychoacoustic effects of different room sizes, of which more later!
If you're not used to it, working with calibrated monitoring levels will involve a few small changes in working practices. The most obvious change is to always starting a mix session with the monitoring at a 'calibrated' reference point. For some it may also involve finding and using more appropriate meters, or using existing meters in a slightly different way, too.
So, down to the nitty‑gritty! How do you establish a reference monitoring level in your control room? The process is fairly straightforward and involves only seven steps...
Step 1: The first stage in establishing a calibrated monitoring level is to decide on a preferred digital 'operating level'. This can be anything you like, but I prefer to stick with the professional convention that calls for a 20dB headroom margin, so the digital operating level becomes ‑20dBFS. To make the chosen operating level obvious we need to find a suitably calibrated meter scale or some other way of making the reference point. Analogue meters had suitable reference points built in — like the 0VU mark — but standard digital sample meters in DAWs usually have no such provision.
However, some systems allow the meter to be customised with different coloured regions and, for example, in some of my DAWs I set the metering up with green below ‑20dBFS, yellow from ‑20 to ‑10 dBFS, and red up from there up to 0dBFS. If you are using an analogue‑style meter plug‑in you will simply need to set the reference level (0VU, PPM4, etc.) in the configuration panel to the preferred digital reference level (‑20dBFS or whatever). Or you could embrace Bob Katz's K‑20 meter — plenty of manufacturers offer K‑System meter plug‑ins.
If you always work with low dynamic‑range mixes (ie. the 'bang it against the end‑stops to make it louder than everyone else's' school of mixing!) then you may wish to employ a smaller headroom margin with a correspondingly raised operating level. However, I certainly wouldn't advise setting an operating level higher than ‑12dBFS in any circumstances, and a digital operating level of ‑14dBFS might be more practical. (K‑System meters support both of these options, too.)
Step 2: The standard test signal for acoustic alignment is pink noise, which sounds like a waterfall, with a smooth top‑end. Ideally, it should be a band‑limited pink noise signal in mono, and at an RMS level that precisely matches the chosen digital operating level. (RMS is short for 'root mean square', which is a kind of averaging process.) The band‑limiting means that the frequency response has been curtailed deliberately, and in this case it is usually restricted to 500Hz to 2kHz. The idea is to minimise the likelihood of any low-frequency standing waves or strong local mid and HF reflections within the listening room causing significant measurement errors. (You can use full-range 20Hz to 20kHz pink noise if you like, but the measured acoustic level is likely to be less accurate, unless you're in a well‑treated room.)
Mono, band‑limited, pink noise files with a guaranteed RMS level of ‑20dBFS are freely available on the Internet or can be generated in some DAWs. Bear in mind, though, that very few meters indicate true RMS levels, so don't be surprised if your guaranteed ‑20dBFS RMS pink noise file appears to replay at a slightly different level! (For this reason, it is often worth first checking the electrical system alignment with a ‑20dBFS sine wave test signal, which will register accurately on all meters). If you want to work at a different operating level you'll need to adjust the pink noise signal accordingly.
Step 3: All Monitoring systems have volume controls, of course, and although that might appear to make a mockery of the concept of a calibrated monitor system, we really do need to be able to adjust the volume sometimes! Often we need to turn the volume up to identify unwanted low‑level noises, or to rock our clients. And sometimes we need to turn it down because we want to check how the mix works at low levels, or just to audition commercial reference tracks without going completely deaf.
So, what we need is a volume control that can be set to a precise calibrated position when we want to adopt our reference level. Some monitor controllers have a dedicated preset level setting for this purpose, most have a specific '0' mark on the volume control scale, and at worst you can add a chinagraph pencil mark to the control panel yourself! In a perfect world, the monitor volume control will also be calibrated in decibels so that you can adjust the replay volume in precise increments away from the reference position. This is a useful option as it allows you to assess the relative 'loudness' of auditioned material in a qualitative way. So you can assess how a compressor or limiter has made the track louder by noting how far you have to adjust the volume control to match perceived loudness with your reference level.
Most computer audio interfaces have some form of monitor volume control built in, and they can be used directly if required, but I recommend using an external analogue monitor controller whenever possible — even if it's just a simple passive controller like the SM Pro Nanopatch. There are several reasons for this, but the most important is that if the computer goes mad, or you accidentally create a howlround, you will generate a peak‑level signal. When working with an operating level of ‑20dBFS, that means the volume will be 20dB louder than usual and so you'll really want to be able to turn things down in a hurry! That's a lot easier to achieve if the monitor controller is (a) right in front of you, and (b) not reliant on the computer still recognising control inputs! A separate monitor control also makes it easier to optimise the gain structure through the monitoring chain between converter output and active speaker (or power‑amp) inputs.
Step 4: An SPL meter will be required to measure the acoustic sound-pressure level produced by the monitors. The critical requirements are that it has a C‑weighting filter option (a flatter response than the A‑weighting curve usually employed for general measurements), and a slow averaging mode. I use a cheap analogue SPL meter for this purpose — the kind of thing you'll find on Amazon for about £15$30. There are many variations on the theme, but I recommend the analogue VU‑meter versions, as the average level of a noise signal is much easier to determine visually than digital displays, which may be constantly cycling around the digits! Despite the low cost, I've found this kind of SPL meter perfectly adequate for simple monitor alignment duties like this, although if you can check it's accuracy against a properly calibrated SPL meter, so much the better.
An SPL meter app on a smartphone can also be used, of course, but do be aware that their accuracy can vary wildly depending on the quality of their design and whether they can be configured for the specific phone's mic sensitivity and frequency response. Again, if possible, compare the phone SPL meter reading with a properly calibrated SPL meter to check its alignment, and adjust the app calibration parameters as necessary.
Having said all that, the absolute precision of the SPL reading is not very critical for setting up the monitoring in an independent project or home studio. Matching levels consistently between different speakers, and for the system as a whole at different times (such as when the monitors are changed) is what matters for most people. The only time absolute SPL accuracy is critical is if trying to conform with a specific SPL standard used in other studios with which you want to share work.
Step 5: The next step is to decide what reference SPL level to use, and this, I am afraid, is where it gets a little vague! There are basically two options: one is to adopt an industry‑recommended reference SPL (as detailed below), and the other is to arrive at your own preferred reference SPL empirically (see side box). Both approaches are equally valid in a private project studio, but the former is desirable in commercial setups where compliance with other studios is important.
Historically, Dolby‑approved professional cinema mix rooms are aligned so that ‑20dBFS (RMS) pink‑noise produces 85dB SPL from each individual speaker. This level is right in the middle of the flat‑ish portion of the Fletcher‑Munson curves, and it produces an acceptable listening level in a large cinema‑like space. It's worth appreciating that this reference level requires speakers capable of delivering 105dB SPL individually to the listening position for a peak‑level digital signal — and few project‑studio monitors can achieve that.
This 85dB SPL Dolby standard has been widely adopted in other professional audio circles too, and there are references to it from numerous audio standards bodies (for example, the SMPTE's RP200 and the ATSC's A/85). Confusingly, though, it was subsequently discovered that the way in which Dolby measured the pink noise was slightly inaccurate, and so the reference level has been tweaked slightly, and the revised standard is now 83dB SPL (as measured on a full‑bandwidth SPL meter with C‑weighting and slow averaging from a ‑20dBFS RMS pink noise source).
This 83dB SPL reference level (with 103dB peaks) is perfectly acceptable if you're listening in a big space, like a cinema or a film dubbing theatre, or even a very large and well‑treated commercial studio control room. Unfortunately, it will be completely overwhelming in a smaller space, because the listener is inevitably sitting much closer to both the speakers and the room boundaries. The very different nature of early reflections makes the level seem, psychoacoustically, much higher than it would be in a larger room.
Consequently, the optimum reference level for smaller rooms needs to be lower, on a scale which is dependent on the enclosed volume of the room in question. You can work out the volume of a room simply by multiplying together its length, width and height, of course, so a home studio which is four metres wide, six metres long and 2.5 metres high has a volume of 60m³. If you prefer to work in imperial measurements, the example above would be roughly 13 x 19.5 x 8 feet, and 2028 cubic feet in volume.
The recommended reference monitor SPLs for different room sizes are shown in the Room Size vs Reference Level Table.
For the example room cited above, this table suggests the ideal monitor calibration level is 76dB SPL. It should be noted, however, that these figures are recommendations rather than absolutes, so consider them as starting points and if you find that you prefer working with a slightly higher or lower reference level, that's fine. Just keep a note of the level you've calibrated your system to work at so that you can match the level if you need to recalibrate for new monitors at some future point.
Step 6: All that's needed now is physically to set the monitor chain's gain structure to set the correct reference level. Start with the individual monitor or power‑amplifier volume controls at their minimum position, and if you have a variable output on the computer interface or D‑A, set that to maximum. The SPL meter should be placed in the listening position, and most SPL meters are intended to be used with the microphone pointing at the ceiling, not at the speaker! There shouldn't be any obstructions or reflective objects around the meter or between the meter and speaker (including you!) and make sure the meter is set to C‑weighting and slow/average response. Start with the volume setting on the monitor controller at minimum until we've established correct gain-staging in the DAW.
We now need to replay the appropriate pink noise from the DAW to just one of the speakers. However, there is a potential trap here, since most pan controls alter the level slightly at the extremes relative to the centre. This is where that ‑20dBFS sine‑wave tone signal becomes useful because it's the level at the output meter that matters. So play the tone, and offset the channel fader slightly to compensate if necessary for any panning loss or gain. You need to get the output meter for the selected channel to read precisely ‑20dBFS.
With the DAW gain staging correct, substitute the ‑20dBFS RMS pink noise signal for the tone and carefully increase the monitor controller volume control to its reference position. With the active monitor or power-amp controls at minimum the acoustic level will probably be too low. You can now start adjusting the input sensitivity of the active monitor or power amp gradually to raise the acoustic level up to the chosen reference SPL level indicated on the SPL meter. It is important to change the gain slowly to allow the meter time to react (because it is set up for slow averaging).
If necessary, it's permissible to juggle the interface/D‑A output level and the monitor/amplifier input sensitivity against one another to achieve the correct SPL with the monitor controller volume control at its reference position.
Once you're happy with the calibration of this first speaker, route the pink noise signal to the other speaker and repeat the process. Both monitors are now calibrated to the chosen reference SPL, and if the SPL meter and your alignments have been accurate, everything should be spot on. In practice, though, small discrepancies may exist and as little as a 0.5dB error is enough to offset the phantom image centre position slightly.
To correct for this, route the pink noise signal equally to both speakers (ie. pan centre) and listen carefully. The noise should come from a point midway between the two speakers. If it is offset slightly to the right, decrease slightly the input sensitivity control on the right monitor (or amplifier) and check again (and vice versa if the offset is to the left). Be aware that the amount of adjustment required will be very small!
Having followed steps one to six, you can now get used to working with your newly calibrated monitors.
Step 7: You should find that material that hovers around or just above the digital operating level on the DAW's meters is comfortable to listen to, without being too loud or too quiet. Moreover, material that replays high up on the meters should sound too loud, and material that is well below the operating level is correspondingly quiet. As you get used to this arrangement you should find that you are able to build mixes entirely by ear without looking at the meters at all — and find that they actually sit well in relation to the defined digital operating level.
However, I must issue one very important warning at this point. When using commercial CDs as reference material, or when mastering your own material within the peak‑normalisation paradigm, the normal working headroom margin is absent because it has been deliberately stripped off. The signal will be banging against the digital end‑stops and the average level will therefore be some 12 to 15 dB higher than expected in a system aligned for a digital operating level of ‑20dBFS. It's therefore essential that you turn down the monitor level in advance, or attenuate the source material within the DAW to make it compliant with your standard operating level.
This is another situation where a calibrated monitor volume scaled in decibels is quite useful, as the amount you need to turn the level down to restore the expected loudness indicates just how much louder the music is compared with the system reference. In this way you can compare the loudness of mixes in progress with reference tracks in a meaningful way. Bob Katz has an 'Honor Roll' of well‑mastered pop CDs on his web site (www.digido.com/media/honor-roll.html), listed with the appropriate monitoring level settings for equal perceived loudness.
Of course, many home studio enthusiasts feel that they need to use headphones more than most to take the room out of the equation. So what is a sensible level when working in this way?
It's actually very difficult to accurately measure loudness on headphones at home. Therefore, the best approach is simply to set sensible speaker levels, as discussed in the main article, and then adjust your headphones to give a similar subjective level — and, if needs be, make a mark at the reference level on your headphone amp's level control.
If you'd rather define your own reference SPL, instead of accepting the pre‑defined value for a given room volume, this alternative approach can be employed. It is a perfectly valid option for individual project studios, where absolute conformity with other studios is not important. The only difference in approach is that you determine the preferred reference SPL by ear, but every other aspect of the alignment process is identical to that described above.
So, in place of Step five, replay some well‑mixed material from the DAW with the replay level adjusted such that the average level hovers around or slightly above the digital operating level chosen previously. In other words, adjust the track level in the DAW so that it looks right relative to the chosen digital operating level on the meters.
Next, adjust the monitor volume controls as necessary so that the track sounds appropriately loud — the level at which you would typically mix. Bear in mind that you will be working at this reference level for long periods, so don't get carried away; a little quieter is far preferable (and safer) to a little louder.
Now, without touching any of the monitor controls, replay the calibrated pink noise file appropriate to the chosen digital operating level over just one loudspeaker, and use the SPL meter to measure the actual acoustic level received at the listening position. This is the value that you will use as your preferred monitor SPL calibration level. You can now follow Step six to optimise the gain staging through the monitor chain with the operational volume control set at its reference position.
An acoustic reference monitoring level has to be related to something tangible in terms of source signal levels, and that basically means a reference operating level on the system meters. Some engineers like to use DAWs in conjunction with analogue consoles, partly because the familiar and practical analogue‑desk metering encourages sensible gain staging, and analogue meters all have a readily apparent reference level (0VU, PPM4, or whatever). In the digital world many manufacturers offer alternative meter scales or colour coding of standard digital sample meters, while the BS.1770 loudness standards throw away conventional sample meters completely, replacing them with a 'target loudness' level on an integrated loudness (LUFS) scale.
I'd recommend that everyone acquire and start using a BS.1770‑style loudness meter because this is likely to become the standard in the future. A colour‑coded digital sample meter can also be useful, though, and I've described in the main article a typical arrangement with green increments up to ‑20dBFS, then yellow up to ‑10dBFS, and red above that. This kind of configuration naturally encourages the user to set levels to loiter mostly around the boundary between the green and yellow regions — the digital operating level — with transients and intentionally louder sections kicking towards the top of the yellow area. Incursions into the red region inherently look wrong and so the optimal gain structure and headroom margin is encouraged semi‑subconsciously.
Bob Katz, the American mastering engineer, took this approach a few steps further in his K‑System metering. First, he relocated the '0' mark to provide a very definite 'nominal operating level' indication with a defined headroom margin above. This '0' reference level is always aligned with an acoustic monitoring level (notionally 83dB SPL in large control rooms), providing a fixed correlation between what the meter shows and what the ear hears. The K‑System is quite elegant and popular, although in many applications the BS.1770 loudness standard makes it as obsolete as every other historical form of metering. Nevertheless, any metering scheme that defines a sensible operating level and headroom margin is acceptable for a calibrated monitoring system — we just need to tie the meter reading and its nominal operating level to a reference acoustic level reaching the ear.
Cubic Metres (m³)
Cubic Feet (ft3)
dB SPL (C)
284 - 566
10,000 - 20,000
143 - 283
5,000 - 10,000
42 - 142
1,500 - 5,000