1 .3 S ETTING LEVELS You don’t want just any signal reaching your recorder track, though—you want something that actually sounds good! The fundamental prerequisite for achieving this is that you set appropriate signal levels throughout your recording chain. For one thing, low-level noise will inevitably be added to your recording by any recording equipment you use, so you want to keep your audio signal level higher than this “noise fl oor”—the higher the better, in fact, in order to maximize the recording’s signal-to-noise ratio. On the other hand, overloading or “clipping” your recording gear by feeding it levels that are too high will produce unwanted distortion, and the further you push the level beyond any unit’s capabilities, the more audible this distortion will become. So in the normal run of things your aim is to keep signal levels high enough to minimize noise, but not so high that you trigger undesirable clipping distortion.
SAMPLE RATES AND BIT DEPTHS The limit of CD-quality sound is largely determined by its standardized 44.1 kHz sample rate (which restricts the upper frequency-response limit to around 20 kHz) and 16-bit sampling resolution (which results in a noise fl oor at roughly −96 dBFS). Given that most commercial music is distributed in this form (or in a data-compressed fi le format directly derived from it), you should at the very least record at 44.1 kHz/16-bit if you’re planning on releasing anything to the general public. However, to make best use of the CD noise fl oor, it actually makes sense to record at a higher bit-depth, so that the digital noise fl oor of your recording won’t rise above that of a CD even if you increase the levels of your recordings during mixing and mastering. For this reason I suggest working at 24-bit resolution instead, which drops the digital noise fl oor well below the noise fl oor of any other equipment in a typical small studio—at which point you can stop worrying about it! The downside of 24-bit audio is that it takes 50% more storage space than 16-bit, but nowadays this really isn’t a big deal given the ridiculously low cost of digital storage. Some DAW platforms give you the option of recording at “32-bit fl oating point” as well, but I wouldn’t waste further disk space on that—frankly, it’s overkill for practical recording purposes, unless you’re the kind of person who stores their CDs in the fridge to keep them fresh… The choice of sample rate for recording is a more contentious issue. In addition to the CD-quality rate of 44.1 kHz, a 48 kHz rate has long been standard in the broadcast and fi lm industries on account of its ease of synchronization with video equipment. To be honest, it matters very little which you choose for music work, although I marginally favor 44.1 kHz so that I don’t have to convert the sample rate for CD mastering—a process that can have audible side-effects. However, in recent years manufacturers have begun offering higher rates as well, based on doubling and redoubling the 44.1 kHz and 48 kHz standards to 88.2 kHz/96 kHz, 176.4 kHz/192 kHz, and beyond, extending the upper limit of the captured frequency range well beyond the 20 Hz–20 kHz zone commonly regarded as the range of human hearing. The extent to which frequencies above 20 kHz do actually infl uence our listening experience is very much a moot point, but there are also wellunderstood technical reasons why elevated sample rates can actually sound better even below 20 kHz, and many professionals have already voted with their ears and wallets by moving to 88.2 kHz/96 kHz in particular. On the face of it, this should be a strong incentive for ambitious small-studio owners to follow suit, but there are two big downsides to factor in. The fi rst is that working at a doubled sample rate not only doubles the storage space you need, but it also doubles the strain on every data buss and digital signal processor in your entire studio, which frequently translates into fewer simultaneous record/playback tracks, fewer effects plug-ins at mixdown, and more complicated digital cabling. The second thing is that the difference in resolution brought about by the increase in sample rate demands equal resolution of your recording and monitoring hardware, and many lower-cost devices simply haven’t been designed with frequencies above 20 kHz in mind. As a result, although it’s diffi cult to dispute that elevated sample rates make an audible difference to the audio quality, I wouldn’t recommend for anyone on a budget to bother with them. In my opinion, the 44.1 kHz/48 kHz rates are more than a match for practically every small studio I’ve ever been into, so upgrading your entire studio rig to handle elevated sample rates will rarely ever be an effi cient use of resources—especially when seen within the wider context of a music market increasingly reliant on MP3/AAC fi les and media-streaming technology, which fall well short of even CD fi delity.
This seems quite simple on the face of it, but there are a couple of complications in practice: Firstly, different stages in your recording chain may require different optimum signal levels; and, secondly, there are often so many meters and gain controls on hand that it’s easy to get confused about which ones to use. Neither does it make things any easier that there’s no “standard” recording rig these days, and that many small-studio users are frequently working with shared or borrowed gear that’s unfamiliar. So in response to all this I’d like to explain a step-by-step procedure that I’ve found to be pretty foolproof for setting decent levels, no matter what studio setup you happen to be faced with.
1.3.1 Find the First Gain Stage Firstly, try to locate the very fi rst gain stage in your signal chain. To give a simple example, imagine that I’m using an unbalanced splitter cable to record the TRS minijack headphone output of a portable MP3 player into the TS jack inputs of one of those little all-in-one sampling workstations (perhaps one of Akai’s iconic MPC range) as illustrated in F igure 1.9. My fi rst gain stage there will probably be the MP3 player’s headphone level control. A more complicated setup is shown in F igure 1.10: a vinyl turntable being recorded via a DJ mixer and a studio mixing console to a stand-alone stereo analog-to-digital converter, which in turn feeds the digital inputs of a software DAW system. In this case the fi rst gain stage will likely be the Input Gain knobs on the DJ mixer’s turntable channels.
1.3.2 Identify the Important Meters Next, narrow down which meters you need to concentrate on while setting your levels. The way to do this is to work your way through the signal chain, all the way from the fi rst gain stage to the destination recorder track, looking for “checkpoints” where the signal: ■ passes through an analog cable (I call this a “cable” checkpoint); ■ is converted between the analog and digital domains (a “conversion” checkpoint).
Optimizing the signal level at a given checkpoint should pretty much guarantee that you’ll steer clear of noise and distortion problems between that checkpoint and the previous one—provided that you were conscientious in setting unity gain throughout the signal path back in Section 1.2.1. The most suitable meter for assessing the level in each case will be: ■ the closest meter before a cable checkpoint; ■ the digital meter closest to a conversion checkpoint. If you can’t fi nd a meter for a given checkpoint that satisfi es these conditions before reaching another checkpoint (or the fi rst gain stage), then it means that it can’t reliably be measured, so strike it from your list and continue working through the remaining checkpoints as if it didn’t exist. (Theoretically speaking, disregarding the signal level at any checkpoint raises the possibility of noise/distortion concerns, but in practice problems very rarely arise, because manufacturers of studio equipment tend to provide metering where it’s required.) Bear in mind that some meters may be able to measure signals in a variety of signal-chain positions—for example, the main meters of a small analog mixing console will often display the control-room monitoring signal, allowing you to measure the level of different channels, groups, and returns just by hitting their Solo buttons. (Solo buttons may sometimes be labeled PFL (Pre-Fader Listen) or AFL (After-Fader Listen) to indicate where in the channel path the signal is being measured.) Let’s look at how these principles apply to our two setup examples. In F igure 1.9 there’d be two level checkpoints: the headphone splitter cable and the sampler’s analog-to-digital converters. However, assuming there was no metering within the MP3 player (which is quite common), then there’d be no way of measuring the level prior to the fi rst checkpoint, so I’d disregard that one and focus all my attention on the second checkpoint instead, relying on the sampling workstation’s digital input meters. (In theory, by ignoring the fi rst checkpoint I’d risk overloading the MP3 player’s analog output circuitry if I turned the headphone level up too high. In practice, though, this would be very unlikely given that most MP3 players are designed to play back even the loudest digital fi les at maximum headphone volume without signifi cant distortion— which is why they seldom require output metering!) Turning to the more complicated setup in F igure 1.10, the checkpoints would be: the cables from the DJ mixer’s main outputs, measured from the DJ mixer’s
master output meters; the cables from the studio mixer’s group outputs, measured from its group output meters; and the analog-to-digital conversion stage, measured from the digital meters most closely following it, i.e., those within the standalone converter unit itself. (Note that the cable between the standalone converter unit and the audio interface wouldn’t qualify as a level checkpoint because it’s not an analog connection.) I’d have to keep my eyes on the meters for all three checkpoints when setting levels, to be sure of capturing the cleanest signal.
1.3.3 Adjust Gain Through the System Now make sure your monitors are muted and restart your playback device— it’ll save time if you can shuttle to the loudest section for level-setting purposes. Adjust your fi rst gain stage while watching the meter for the following level checkpoint. The reading you’re aiming for will vary depending on the type of meter you’re using, so here are some guidelines: ■ Digital Peak Meter. This shows the instantaneous level of a digital signal, sample by sample. At the top of a digital meter is 0 dBFS, which is the highest level the system can capture before overloading. The basic level-setting tactic here is to make sure the signal is as hot as possible, but without ever hitting the top of the scale. (Digital distortion is one of the nastier-sounding varieties.) There’s no need to be too fi nicky about things, though—if the loudest signal from the playback device registers within the top 6 dB of the meter’s scale you’re fi ne. ■ Volume Unit (VU) Meter. This is normally used for analog signals, and doesn’t respond nearly as quickly to fast-moving waveforms as a digital peak meter does. As such, it tends to favor lower frequencies in general, as well as seriously under-reading short-term level spikes (often called transients) that are responsible for both the percussive attack of drums and the note-onset defi nition of many other instruments such as acoustic guitar, piano, and tuned percussion. For this reason, the 0 dBVU “reference level” marking on such meters is usually designed to correspond to an electrical.
level about 20 dB below the actual overload point of the surrounding analog circuitry. Therefore you can set your gain control to give a reading at the meter’s reference level and you’ll still have around 20 dB of “headroom” to accommodate the unmetered transient peaks without distortion. (If your VU meter eschews the dBVU scale in favor of dBu markings, then you can usually treat the +4 dBu level as 0 dBVU.) ■ Peak Program Meters (PPMs). These meters are pretty uncommon in small studios, so you’re only likely to come across them if you’ve scavenged some broadcast gear for your recording rig. Like VU meters, they’re primarily intended for use with analog equipment, but they respond much better to transients. For this reason, you can set a recording level only about 10 dB below the analog overload point on a PPM without any real danger of distorting unmetered signal peaks. Irritatingly, though, there are several different labeling standards commonly used for PPMs, some of which use decibel markings, while others use an arbitrary numeric scale, so you’ll need to work out roughly where the “10 dB below overload” point appears in your specifi c instance if you’re going to use PPMs to judge your gain settings. ■ Uncalibrated Meters. Metering is one area where manufacturers of budget equipment tend to cut corners, most commonly by replacing fully-featured metering with a single LED. The simplest design just lights up when it senses any signal signifi cantly stronger than the unit’s own noise fl oor, and is therefore mostly just a line-checking tool. More useful for level-setting is the overload/clip LED, which warns you of impending distortion on signal peaks, so your primary concern while setting recording levels is to turn up the gain as far as you can without triggering that. These two single-LED meters complement each other quite well, so they’re often twinned in practice, or else have their functions combined via a single variable-color LED.
When you’ve achieved the meter reading you’re after for your fi rst checkpoint, grab whichever gain control is immediately after the fi rst checkpoint and adjust it (if necessary) while looking at the next checkpoint’s meter. Continue in a similar manner until you’ve got appropriate meter readings for all the checkpoints.
Returning to our examples, in F igure 1.9 there’s only one measurable checkpoint, so I’d just turn up the headphone volume control until the loudest playback signal registered just under 0 dBFS on the sampling workstation’s digital input meters. In Figure 1.10 I’d fi rst grab the Input Trim knob on the DJ mixer’s turntable channel, and set that for a 0 dBVU reading on the unit’s master output VU meters. Then I’d shift my focus to the studio mixer’s group-output VU meters and check whether those were reading around 0 dBVU—if not, I’d adjust the Input Gain knob on the mixer channel(s) receiving the DJ mixer’s stereo output. Finally, I’d check the digital meter on the standalone converter unit and if necessary adjust the input sensitivity controls on the converter to achieve the most sensible reading.
One potential fl y in the ointment, though, is if a specifi c gain control doesn’t have enough juice to achieve your target meter reading. Here’s what to do in that scenario: ■ If you need more boost, use the next gain control in line—as long it still precedes the metering point. ■ If you need more cut, use the previous gain control in line.
So if, for example, the DJ mixer’s Input Trim in F igure 1.10 didn’t give me an adequate level on its master output meters, then I might apply further gain by pushing the turntable channel’s fader above its 0 dB mark, and even the DJ mixer’s master fader too. Alternatively, if setting the optimum level on the studio mixer’s group output meters resulted in overloading the standalone converter unit’s digital meters (even at the lowest input-sensitivity setting), I might turn down the studio mixer’s group output fader to reduce the level.
Niciun comentariu:
Trimiteți un comentariu