Welcome to Day 14 of 31 Days to Better Recordings.

A lot of people ask me about setting levels for recording. It seems simple enough, but people tend to be a little nervous about it.

What if I record it too loud? What if it clips?

What if I record it too quietly, and it’s never loud enough?

These are legitimate concerns, but I would say that it doesn’t matter nearly as much as you think it does. Yes, proper gain staging is important, especially when using outboard gear, but when setting levels coming into your recording platform, it’s not as tricky as it may seem.

24 Bits to the Rescue

I wrote an article entitled 24-bit vs 16-bit that might be helpful for you. In it, I explain the differences between 16-bit and 24-bit recording, and why you should always record at 24-bit.

In summary, 24-bit recording allows for a huge amount of dynamic range.

In the past (back when we were recording to analog tape), things were a bit tricky. If you recorded the signal too quietly, it would be covered up by tape noise (tape hiss). If you recorded it too hot, there would be a lot of tape saturation (which could be either a good or a bad thing, depending on your goals for that track).

In both digital and analog systems, you certainly want to prevent clipping, but the way they each handle recording levels is important to know.

With analog tape, the actual sound of the signal would change the harder you hit the tape (i.e. the louder the signal was). There is a fine line between getting that nice tape saturation and getting distortion and clipping.

With a digital system, however, the signal stays exactly the same. It doesn’t get “warmer” if you record it at a louder level. And the noise floor of a digital system (the inherent noise of the system itself) is much lower than the hiss of analog tape. Even if you record at a low volume, there’s not much chance of the signal being very close to the noise floor.

But Joe, what does it mean?

At one point, early in my recording career, I would try to peg the meters every time I recorded. I would try to get that little dancing light to hit as close as possible to clipping without going over. I thought I was maximizing the sound. I wasn’t.

Since the signal is going to sound exactly the same at -10 dB as it will at -2 dB, don’t bother trying to get really hot recordings. Let the lights dance between 1/2 and 3/4 of the way up the meter. That’ll leave plenty of room for louder sections (like a particularly loud snare hit), and you’re not in danger of losing out to noise.

All I’m saying is don’t add more stress to a recording session by trying to get really, really high levels. It’ll sound just as good (and you’ll be much more comfortable) if you just turn everything down a little bit.

Recommended reading: Setting Levels for Recording

Day 14 Challenge

Your challenge today is to try easing up on the gain next time you have a recording session. Report back here and leave a comment. Does it sound just as good? Were there a lot less clipped signals and wasted takes?

  • Pingback: Ask Joe #29 – The Resurrection of the HSC Podcast! | Home Studio Corner()

  • angel

    hello,

    ok to all above but…

    if you record in 24b using just a half of the dinamic resolution, don´t you end up with a lower quality track?

    • Nope. The sound quality is exactly the same whether your mix peaks at -1 dBFS or -12 dBFS. Good question, though.

    • angel

      Thinking it twice, there´s no reason to have lower quality, unless you record so low that you have to deal with noise floor.

      My mistake it´s about thinking that the dinamic at higher inputs could be higher, but it doesn´t; you just raise the signal level but its dinamic range it´s the same (but situated higher), so it´s coded in the same x bits of depth.

      …I think… 😉

      Anyway, I´m glad having stopped thinkig about it…

  • Arjun Ramesh

    I recently recorded my own vocals for a song and I tried to keep the meter hitting right around where the fader was at nominal level, which is still enough to get great signal above the noise floor and a little breathing room for accidental louder parts. It turned out great. I used to try to peg the meter, too and I knew that when I would hit the louder parts, I would either have to back away from the mic and belt it out (which also changes the sound compared to the other parts), or overdub it by bringing the gain down (which is just more work), or compressing it to tape (which also changes the sound). These methods were just ridiculous for me and mistakes I used to make when I first started out in digital recording.

  • Phil Harmon

    Just to be specific, we are talking about input gain levels on one’s audio interface. Just turning down the faders on one’s DAW so that the input looks like 1/2 to 3/4 of peak is not the same.

    • YES! Great point, Phil. We’re talking about the level of the actual recording itself.

  • Canadian Guy

    After doing some reading on the science of audio and experimenting with various recording levels, I ended up following two rules: record everything at a max of -10dB and make sure the final mix for bouncing is topping out at -3dB. This achieves a couple of things simultaneously: it forces me to really concentrate on the mixing levels and bring everything down because I’ve set a “clipping limit” lower than 0dB; and it gives the mastering engineer plenty of headroom to play with. When I finally get my hands on my mastered CD and play it through my stereo system, it’s awesome experience to hear how it’s been beefed up during the mastering!!

  • Preshan

    I’ve learnt this recently too. I record about 1/2 to 3/4 on average, but a little lower than this for drums because drummers need lots of headroom…

  • Matt

    I tend to set the levels around 3/4 of the way up the meter. I’d rather record a lower signal than ruin a great take by clipping.

  • I have a question long needing some sort of answer. I use ,[ PC DRUMMER ] AND SET UP THE PARTS, [ DRUM PARTS OF THE SONG IN ORDER ] AND THEN EXPORT THEM INTO pro tools. However there is a volume percentage slider, from 0 to 100% and wonder if there is any norm for setting this here as would be with alot of other dig. programs being exported.?? Appreciate any help.
    Take Care,
    TOMMY…

    • Not sure on that one, Tommy. It could be anything. If it’s quantization, then I wouldn’t put it at 100%, as that would make the part sound very robotic.

      Your best bet would be to try it and see which one you like better.

  • Frank Adrian

    This is something that I learned recently. I used to try to get the meters right up to the point where they were almost (but not quite) clipping. This often led to re-takes. I’ve since dialed down my audio interface inputs with very little change in final result.

    Older books say that you should go in with the hottest signal possible, but they were written during the day when analog equipment was noisy and, if you didn’t stage gains correctly, you had a noisy mess on your hands. This myth is kept alive by people who trained in that era and who dream of the days of going back to an all-analog chain. In addition, the old analog chains were more forgiving of overload, and most components saturated in a musical manner rather than giving hard clipping distortion. Digital chains are much less forgiving of overload and changed everything.

    These days, with an all digital chain, the S/N issue is much less critical – once your A/D converter gets your signal into the digital realm, there’s not going to be a lot of additional noise introduced in processing. If you get up to about 3/4 on the meter, you have enough signal that you’ll do fine. As I said, digital changed everything.

    The new rules are:

    (1) Each transition between the digital and analog domain (and vice versa) will add noise – minimize conversions between analog and digital domains.
    (2) Analog stages are noisier than digital stages. Get into the digital domain as quickly as possible with a non-clipping signal having a reasonable level.
    (3) If you want to add “analog warmth”, the best place to do it is in a preamp placed directly before your audio interface.
    (4) A well-written plug-in will generally not add noise. Certain plug-ins like compressors, EQs, etc., can make existing noise more obvious, but on their own, assuming that the arithmetic routines in them are done well, they will not add noise (and, no, if you record in more than about 22 bits of resolution, you will not hear quantization noise).
    (5) If you’re going to use outboard gear, it had better sound beaucoup better because the conversions going from the DAW to the analog gear and back through A/D, as well as the gear itself is going to penalize you in noise level. This goes for other analog processes like re-amping, as well. Proper gain staging from the D/A converter into the analog effect is still critical here.
    (6) If you want to run multiple outboard effects, chain them directly without going back and forth into and out of your DAW. Proper gain staging through each of these analog stages will be critical.
    (7) In most cases, properly staging the gain in outboard analog equipment will be such a pain, you’ll either (a) have a permanent external chain set up so you don’t have to screw about with re-staging the gains or (b) find yourself using plug-ins, because most outboard equipment that you’ll be able to afford doesn’t sound that much better.

    • Matt

      Frank, thank you so much for the information you posted!!

    • Well said, Frank. Thanks!

  • Everett Meloy

    I do record between ½ to ¾ level because of the unexpected loud surge. Sometimes expected or not, these “or not’s” usually are directly proportional to the “did that clip?” shutter. Then the dreaded words, we have to do that again, is such a buzz kill that sometimes the performance doesn’t recover to how good it was going before the clipping. That loss is worse than not recording something great because you were recording already recording it.

  • The analog components in your gear are still designed to work optimally at line level which ends up being around -18dB digital.

    • Wish somebody had told me that 10 years ago… 🙂

      • Rob

        Glad somebody told me today 🙂

  • Christopher w

    I always leave room for a louder part. I tend to set the levels just under 3/4 of the meter which gives the signal enough to play about in.