Levels for Recording Acoustic Guitar

darknight

New member
How quiet is too quiet when recording through an interface into a DAW?

I'm recording an acoustic guitar and when I playback I see it's around -30 on the meter (using Reaper). If I crank it up it still sounds fine to my ears as there isn't too much background noise.

However, convention seems to be to record as close to clipping-point on the interface as possible without actually causing clipping. I'm way off from that point with my recording but I'm wondering will this cause any problems down the road as I start mixing/would this be an issue for someone who was mastering?

I'm happy with my takes but would consider doing them again with more input gain to avoid problems later on.

Thanks :)
 
How quiet is too quiet when recording through an interface into a DAW?

I'm recording an acoustic guitar and when I playback I see it's around -30 on the meter (using Reaper). If I crank it up it still sounds fine to my ears as there isn't too much background noise.

However, convention seems to be to record as close to clipping-point on the interface as possible without actually causing clipping. I'm way off from that point with my recording but I'm wondering will this cause any problems down the road as I start mixing/would this be an issue for someone who was mastering?

I'm happy with my takes but would consider doing them again with more input gain to avoid problems later on.

Thanks :)

Don't know WHERE you found "convention is as hot as possible" from! It is surely well known now that you should aim for an average level of -18dBFS and peaks around -10? A check in "Dummies" shows Jeff Strong is even more cautious with a -12dBFS max level for a 24bit* recording system.

Neg 30 will be perfectly acceptable. Yes, you could up the gain a bit but guitars are spikey and you do not want to clip. Digitally boosting the signal post tracking will only bring up analogue noise that was there anyway (assuming a 1/2 decent AI!) which would happen anyway if you boosted the gain.

*I do not agree with his condemnation, out of hand, of 16 bit systems. Yes, some cheap systems are shit and are 16 bits but not all 16 bit systems are shit! If you are recording a live event, especially say a pub band. The PUB sets the noise floor and 16 bits is all you need!

Post a clip chap (top Q MP3)

Dave.
 
View attachment Birnam.mp3

Here's a snippet of one of the tracks (it's doubled in my DAW with another recording). I left about half a second in at the beginning of near-silence so you can hear my laptop whirring but that's the loudest it gets in terms of background noise. There's no FX on the track either, just the pure take.

And when I say 'convention' I mean a lot of what i've read online, although some of it may have been referring to tape recording in fairness.

Cheers.
 
-30 is getting a bit low but whoever told you the convention is "as hot as possible" has it wrong.

Perhaps a bit of explanation:

Although people talk about -30 or +8 or whatever but those terms are actually pretty meaningless without some initials after the number. For computer recording, people assume we're talking about "dB(FS)". Back in the analogue days, people were usually talking about dB(VU) or dBu. What's the difference?

Well, in the analogue world, 0dBu was based on an totally arbitrary voltage (0.7746 Volts if you care but it's not really important). What IS important is that analogue gear was built so it could handle much higher voltages before clipping distortion--+18dBu was typical. This meant that, when recording, it was normal to have your VU meter bouncing around zero for the average and maybe peaks at about +8, leaving lots of head room before an odd loud transient could cause distortion.

Digital recording, on the other hand, uses dB Full Scale (dB(FS)). What the Full Scale indicates is that it's the absolute maximum the signal can be and even a fraction more gets into digital clipping. Quite seriously, if you could analyse the digital signal, something at 0dB(FS) would be 111111111111111111111111 in 24 bit. You have no room to go any louder.

Anyhow, since 0dB(FS) is the loudest you can possibly have, it's normal to say 0dBu=-18dB(FS). This means that a recording averaging -18dB(FS) with peaks around -10 is roughly equivalent to a typical analogue recording.

Is it a problem to record louder than that? Not usually. However, you're risking getting closer to clipping (which is technically known as a "bad thing") and also you can sometimes find some outboard hardware or software imitating analogue that doesn't play nicely with higher levels.

So...while -30 is a bit low, as hot as you can go is even worse.
 
Not entirely arbitrary Bobbs! As I am sure you know, the 775mV comes from one milliwatt into 600 Ohms and THAT is lost in the mysteries of ancient telephone technology!

One of the things that might confuse the newb (and infuriates me!) is the fact that no "domestic" AI maker I know of gives input sensitivities referenced to 0dBFS. Also, most AIs are simply not capable of +18dBu output.

Just pulled a few speccs out of my .pdf box...

Roland Opto capture: max out + 16dBu

Steinberg UR22 (bus powered) max out +10dBu
" UR44 (Rat powered) max out +16dBu
But! for budget interfaces such as these (and others to be fair) you would not want to get within 3dB of that max out specc.

One stranger in the camp? Behringer X222 USB mixer, max out +22dBu!

The "hit the tape hard" school is really only quite a recent one and really only applies to the "rock" scene?
Those recording classical and jazz were constantly frustrated by the noise at one end and the distortion at the other! Many "classical" adopters of Dolby A forsook some 6dB of noise reduction to get lower distortion.

Dave.
 
How quiet is too quiet when recording through an interface into a DAW?

The answer is really relative to noise.
If there's no noise then there's no such thing as too quiet.
There is alway such a thing as too loud though. There is an upper limit at which your input signal will be irrepairably clipped.


In the old days you tracked hot and with compression and all sorts of tricks to keep above the noise floor.
These days that's so much less of an issue.

Maybe some people want to track hot to push their tubes or transformers or whatever but for basic healthy recording, go with the above.
 
Some interfaces tell you to track as hot as possible without clipping. They do. Read the manuals. Is it silly? Sure, but that's probably where the confusion comes from. You CAN actually record right up to clipping and it's perfectly fine. It'll make mixing suck ass, but for one track, you can do it if you want to. Don't though. -30 is pretty fucking low, but whatever.
 
But will there be any detrimental effects to raising the level of the track after I record?

Like I said, I'm fine with the level of background noise relative to the guitar but i'm just wondering if it will introduce any other issues.
 
But will there be any detrimental effects to raising the level of the track after I record?

Like I said, I'm fine with the level of background noise relative to the guitar but i'm just wondering if it will introduce any other issues.

All that background noise will increase with the whole track. That's the detrimental effect. Will it be a problem? I don't know, I haven't heard your track. Ideally you'd have set your pre gain to a usable level but with room to spare in a quiet environment. You live you learn.
 
But will there be any detrimental effects to raising the level of the track after I record?

Like I said, I'm fine with the level of background noise relative to the guitar but i'm just wondering if it will introduce any other issues.

No. The ratio will remain the same and there are no other issues I can think of.

Just to be clear, I didn't mean background noise in my previous post.

If you want to do something about background noise you need to make your instrument/source louder, or place the mic closer, or......anything that physically alters the ratio in the room.
 
But will there be any detrimental effects to raising the level of the track after I record?

Like I said, I'm fine with the level of background noise relative to the guitar but i'm just wondering if it will introduce any other issues.

If the noise floor is okay after you boost it, then no more technical problems. However, if you're mixing multiple tracks together and they're all near 0dB(FS), i.e. just below clipping, they you'll have to lower them all anyway because levels in a mix are cumulative. Normally, another reason to keep levels will below maximum is simply for convenience in the mix.
 
Back
Top