How should you set your instrument levels prior to recording?

Question

That title isn’t very descriptive, but what I’m asking is kind of hard to summarize. I’ve been digging into some basic tutorials and introductions for recording, but so far I haven’t come across what I need.

I’m looking to get some recordings into my DAW software. I’m working with a simple little USB interface (M-Audio Fast Track: http://www.amazon.com/M-Audio-US44010-Interface-GT-Player-Software/dp/B00061ZM2Y)

It’s a pretty humble setup, but I’d like to make the most of it.

I intend to record:

• An acoustic guitar with an onboard preamp (four band EQ)
• An acoustic bass with an onboard preamp (three band EQ, with, uh, a button marked “PHASE” and a slider marked “SHAPE.” What are those for?)
• An electric guitar (LP style; two humbuckers each with tone and volume)
• A basic dynamic mic for vocals

I’ve done a bit of recording with all of these in the past, and every time I’ve felt like I’ve been chasing myself in circles trying to get volume and tone dialed in right. The USB interface has an input level control, then the instruments have their own controls, then there are countless controls in the software. Change one, then change the other, then the other, then they’re all at weird extremes and nothing sounds right. I need to get this stuff straight.

So, my questions all sort of feed down to, what are the rule-of-thumb basic settings I should be using for all of this stuff as a baseline, and then where should I be making my tweaks? To get more specific:

• How high should the level control on the USB interface be while recording? As high as possible without clipping? Is there something I should be watching for in the DAW’s metering?
• What should I be doing with the onboard EQ for the acoustic instruments? Should I worry about dialing in the tone at all there, or should I just flatten the EQ in the middle setting and worry about that in the software? And what about the volume control? Is there a specific spot I should be setting that? And what the heck are those shape and phase controls on the bass, anyway? :(
• How about the electric guitar? Should I be thinking ahead to what I want my tone to be like and select which pickup / tone settings I want to record with based on that? Or does it make more sense to record with both pickups selected / tone knobs in the middle / volume knobs maxed out, and then carve out the tone with equalization and amp simulation controls in the software?
• The microphone is the most straightforward of the bunch, I suppose, but is there anything special I should be looking out for there?

Maybe that’s a lot to ask in a single question. If someone could even point me toward some “recording basics 101” type site that outlines the stuff I asked, I would be incredibly grateful!

Answer 1

First, you should consider forgetting about those on-board pickup/preamp systems. Acoustic instruments sound best when recorded with an external condenser microphone, full stop. If the pickup signal is no good, there's not much you can do in mix. If the acoustic guitar has decent quality, then with a standard small-diaphragm condenser mic and no processing whatsoever you should get a nice sound, better than even the most delicately fine-tuned post-processing of a piëzo-pickup signal.

While the acoustic bass is also... well, acoustic, that might not quite apply to it: if it's just another of those slightly-oversized western guitar things, then the body probably doesn't offer enough low-frequency resonance to give a convincing microphone sound. So you should probably use the pickup; I'd use linear EQ and level all up.

But levels and EQing are less critical, because their influence is largely reversible: if you've boosted 200 Hz on the preamp and later realise it was too much, attenuating the same band in your DAW gets you more or less back to the original sound. There will remain some phase unevenness, but that's not too much of a problem. Similarly, if you've had the gain a bit too low, normalising up the recorded audio gives you a fine full-levelled result. Indeed there will be some more ADC / quantising noise, but with a modern 24-bit interface it's not really notable (your guitar's preampt has likely a much higher noise level; that and any interference from the unbalanced instrument connection has the same level no matter if you normalise the gain in hardware or software. The only thing you should assert is that there is no clipping: that would be very audible, and much more problematic to repair. So, I'd recommend setting your levels a bit lower than "just not clipping", about -6 dB rather than 0 dB¹: then even an unexpectedly loud accent will still be recorded cleanly.

Electric guitar is quite another issue. While a completely clean "DI" sound (i.e. recorded straight from the pickup into the mix with only some compressor on it) can sometimes come out pretty nice, it is by no means standard, and will rather resemble a straightly played steel guitar than what you'd think of as an "electric guitar sound". You've probably experienced this yourself already: the amp makes up a really large portion of the final sound. That means amongst other things, it's hard to judge from just an unprocessed monitor sound how various pickup settings will come out through the amp. Only if you've done this a lot and know your guitar well you'll just have a reliable feeling for this.

There are two possible conclusions:

• Leave you guitar's controls all the way up. Works surely for volume and also tone, those can be emulated later on (though in a much less obvious way as for your acoustic preamps' graphic EQs. Basically, you need to consider the system pickup+controls+cable as a state variable filter, tone&volume modify not just gain and "treble content" but subtly infuence the frequency an q-value of that filter), but the pickup selector is more critical: pickup position acts as a comb filter, and for each string differently, so there's no real way to change it in post-processing.
• Use a proper amp sound as monitor during recording, so you can set up the guitar as it's ultimately supposed to sound. The classical way is to actually have an analogue amp and cabinet, with a microphone. You can also still record the clean guitar signal, with a DI box, so the final sound can still be done more flexible with a plugin. It is also possible to use that simulation itself for monitor, but it requires your interface's drivers to be set to low latency.

Finally, on the microphone: also for vocals, condenser (large-diaphragm in this case) is pretty definitely superiour to dynamic. If you can afford it, you should get one of those.

---

¹Mind that is is specific to digital recording. On tape, it's perfectly reasonably to level instruments to more than 0 dB to get natural saturation/compression, and indeed that effect is one you might want to put on your tracks as a digital plugin.

Answer 2

Whenever you have a digital device be sure that the levels won't reach 0 dBFs. On analogue devices you can go beyond 0 dB(u or v in most cases) but you will get some overdrive. This harmonic distortion can be a desired effect.

So when you record something be sure that the analogue digital converter -> for example your USB interface isn't reaching 0dB. But you want to set the levels as high as possible, else you get an unnecessary noise floor. A good level is, if you set the really high peaks of a signal at around -9dB (yellow area on a lot of level meters)

On the Preamps -> be sure to play with the eq and really go for what sounds best. Thinking on how you want to use the instrument before recording is very important.

I don't know shape, but phase just inverts the waveforms polarity. So if you record a di signal and the amped bass -> press it to see if the di signal is cancelling something of the amped signal.

Always think ahead on how a guitar should sound. I would never record the DI of a guitar if i am not planing to use guitar rig. I feel more comfortable to commit to sounds. And so as a sound engineer i always push a band/producer to commit to the guitar sound at the recording day. This helps to get things done -> else you need to decide about everything at mixdown.

Good luck recording!

Answer 3

You want to avoid having anything too low. If there is a unity mark, then as long as impedance match, you should be able to use that. If not, adjust things so that you can have moderate settings that are not too low on anything. The lower the volume level gets, the harder it is for the signal to be differentiated from the noise and the lower the quality gets. Similarly, depending on the quality of the gear, if you crank something all the way up, there is a good chance that the electronics won't handle extremes as well as they would if things were a little lower.

You'll need to experiment a bit to figure out where the best balancing point is to avoid extremes though.

Answer 4

You should set your preamp as high possible without clipping to get the best signal/noise ratio. You can than adjust the monitoring level in a comfortable way. About the onboard EQ, I suggest you try to get as close to the sound you want to have at the end of the mix, but some people try to have the most neutral sound during recording (but it doesnt make sense when you want a heavy distorted guitar, you do this when recording not during the mix). I think that the closer you get to the sound you want, the easier the mix will be. But you won't be able to go backwards, you must make choices. In your configuration, if you change your mind it wont be difficult to do another recording :)

Good Luck

Answer 5

This general process is called "gain balancing", and it can be time-consuming to get just right, especially if you have long signal chains or you change your system's configuration often (both are often seen in home studio systems). Overall, the ideal gain structure is the one that maximizes signal-to-noise ratio without inducing distortion.

If, at any time, you have an analog audio component producing a very high signal level, that you then attenuate to very low levels before reamplifying to higher levels, you will increase the presence of noise in your signal. The attenuation will reduce the level of the desired signal (and any "upstream" noise) being input into circuitry "downstream", which will then introduce a constant amount of noise to the signal. Reamplification of this signal will increase both the desired signal and the introduced noise, reducing the signal to noise ratio and increasing the "noise floor". If, instead, the input signal were not attenuated (at least not more than necessary), then the node further downstream would not have to amplify the signal as much or at all, minimizing the increase in the noise floor and maximizing SNR.

So, the best gain structure is the one in which the gain at each gain stage is as high as possible without clipping the waveform, which occurs whenever a gain stage is required to produce more voltage than it is given by its power source in order to accurately reproduce the waveform. Clipping is usually bad as it distorts the waveform from the ideal, and is to be avoided, but you want the highest signal voltages you can get below that threshold, at every step along the way, in order to minimize the relative effect of any noise added to the signal.

Usually, the way you accomplish this is by paying attention to signal level indicators, like clip lights. However, not all audio equipment has a clip indicator. To adjust these pieces of equipment properly, you have to use your ears; adjust all other gain stages, in order from upstream (closest to the audio source) to downstream (closest to the recorder or power amplifier), as high as possible without any clip indicators lighting, then listen to the resulting waveform; if you hear distortion, turn down the furthest upstream source that doesn't have a clip indicator until you don't hear distortion, then turn up the next source downstream from that to compensate. If you immediately hear distortion, the problem is further downstream; set the two upstream sources back to their original levels, then repeat with the next node further downstream that doesn't have a clip indicator. Obviously, the fewer gain stages in your signal chain, especially the fewer that don't have signal level indicators, the better.

Once you have your gain structure balanced, the recording you produce by converting this analog signal to a series of digital samples can be further manipulated without as much fear of introducing noise. You can still clip the sample (in digital audio, clipping occurs when a sample requires a digital numerical value higher than the highest value available for that sample size), but once the signal is digital, it's a series of numbers represented as digital bits, and those bits are unaffected by any noise introduced into the electrical circuits along which they're carried, until the electrical noise becomes so present as to prevent the computer being able to distinguish between a "one" and a "zero" (by this time an analog signal carried along the same circuit would be unusable as well).