What are your average recording levels...

[field scarecrow]

... for instrument tracks, and the whole mix, prior to mastering?

Do you record each drumkit part as loud as possible then mix the volumes down accordingly?

My average overall mix level is about 12 - 15 db quieter than mastered cd's, so when I multiband-compress and maximize, i'm bringing the mix up by that much, does that seem right? are my levels going in too low?

Any replies will be appreciated!

 

[Impy]

It doesn't matter all that much since you're not on analogue tape (i assume). But i always try to record at as high volume as possible, that way you know you won't get screwed (unless you clip it).

 

[Andy Sneap]

I'll always get a good level into pro tools, just make sure it's not too low and theres no clipping.

 

[Kenny Lee]

It doesn't matter all that much since you're not on analogue tape (i assume). But i always try to record at as high volume as possible, that way you know you won't get screwed (unless you clip it).

WHOA!! Are you tracking to tape?

 

[MR NINE]

For instrument tracks and everything else get the highest signal as possible without cliping or distorting (sometimes just because it isn't clipping doesn't mean that you are not getting a distorted signal!)
As for mixing I usually mix it with about -3/-4 dbs so that after you have enough headroom for mastering. By the way what are you using for recording? DAW or Analogue?

 

[Muttley]

I've been trying to record everything as loud as possible (without clipping). Seems to work, but IANAE.

Do you really need to mix at -3/-4 dbs to leave room for mastering?

Muttley

 

[Plec]

Digital levels isn't that important as you might think. Clipping is of course no good, but there's no problem recording levels really low and still end up with the same result.

You have a total of 24 bits to work with (144db of dynamic range), so the only time you really need to be a little bit more picky about your levels is when you record nothing more than a stereo pair of microphones, and that's it. Those tracks will be represented by all your available bits. What happens when you start adding more tracks is that they all will share the maximum 24 bits at the mix bus, so every track in the mix will be represented by fewer bits when played together at one point in time.

That's all some fun theory that you can explore, but what it really comes down to in the end is that you're probably going to CD with this stuff. CD is of course a 16bit medium (98db of dynamic range). If you're recording at 24 bits, you could record every track peaking at -48dbFS and you would still end up with the same resolution when going to CD in the end.

Another aspect is that there are very few converters that actually give you more than a 20bit resolution in practice (120db) even though they are 24 bit converters. So If you REALLY want to be on the safe side for just about every medium, you don't need to record any levels above -24dbFS.

 

[Impy]

WHOA!! Are you tracking to tape?

No, i just said that the only time that you need to be really observant of the levels is when recording to analogue tape.

That's what i meant anyways...

 

[field scarecrow]

For instrument tracks and everything else get the highest signal as possible without cliping or distorting (sometimes just because it isn't clipping doesn't mean that you are not getting a distorted signal!)
As for mixing I usually mix it with about -3/-4 dbs so that after you have enough headroom for mastering. By the way what are you using for recording? DAW or Analogue?

I'm using digital, a PC with Cubase. Thanks so much for the replies guys, I appreciate it! =)

 

[MR NINE]

I've been trying to record everything as loud as possible (without clipping). Seems to work, but IANAE.

Do you really need to mix at -3/-4 dbs to leave room for mastering?

Muttley

I mean that when you record your mix to a Dat or Cd or a stereo track in your DAW for mastering purposes.
Not when you are actually mixing the instruments, vox, etc.

Imagine that your Mix is a glass of water!?!
Now if you fill it to the top and then if you want to put something more in it, the water comes out! Do you see what I mean?!
0 Db it's the full scale and the maximum volume that you get (or want!) starts in the way that you capture the sound/instrument

 

[Kenny Lee]

Ahhh, I see Impy Wasn't trying to be a A-hole

 

[SPLASTiK]

What happens when you start adding more tracks is that they all will share the maximum 24 bits at the mix bus, so every track in the mix will be represented by fewer bits when played together at one point in time.

Most Mix Buses are a higher resolution that the individual tracks. This was a problem in early digital days when they didn't know a lot about digital audio theory so the early 16bit digital mixers with 16bit mix buses sounded like shit, but most mix buses nowadays are of a higher resolution.

Plus on top of that, you aren't really using all 24 bits unless you're faders are maxed to the top volume which is a reason to set your record levels higher so you get the maximum resolution possible at lower volumes.

I try to peak the tracks around -4 to -6dbfs since for every 6db loss in signal you lose 1 bit of resolution.

 

[Plec]

What do you mean higher resolution than the individual tracks
You record at 24 bits usually, but your digital mixbus is usually no less than 32bit float, this is good 'cause of the immense math that's going on, but 24 bits is the absolute max that you're ever going to hear. Mainly because the ear can't handle any more and secondly because in practice all you're going to get is 20-21bits due to the limits of todays electronics.

Plus on top of that, you aren't really using all 24 bits unless you're faders are maxed to the top volume which is a reason to set your record levels higher so you get the maximum resolution possible at lower volumes.

Actually... you're always using the 24 bits you recorded... if you max your faders you're not gaining resolution, you just gain gain Everything you do in the DAW takes DSP... volume changes, panning etc.. This is normally done at 32 bit float or higher so the system can handle the size of the math in order to produce a clean and believeable result. If you move the fader in any way you're recalculating the audio which makes it go 32 bit float instead of your original 24bit fixed.

The resolution of the mixbus us all you have. For example...

Let's say you have a 24bit sine wave on a single channel at -0dbFS (no gain change or other DSP processes) going into the mixbus. This sinewave will then be represented by all the available 24 bits. If you have 8 channels of -0dBFS peaking same sine wave, going into the mixbus it's not going to fit in there since 8 channels of full 24bit material playing at the same time will not fit into the 24bits of the mixbus. Fact is that you will only be using 3 bits per channel (that is 16db of dynamic range per channel effectively) to fit it into the mixbus without clipping.

This is an extreme example, but the theory is correct and applies to all.

The point is... don't worry about your recording levels being to low. Most people record their tracks peaking at -8 to -4dBFS or something, but you don't need to ever think about it if you don't go beneath -20dBFS really.

 

[SPLASTiK]

Actually... you're always using the 24 bits you recorded... if you max your faders you're not gaining resolution, you just gain gain Everything you do in the DAW takes DSP... volume changes, panning etc.. This is normally done at 32 bit float or higher so the system can handle the size of the math in order to produce a clean and believeable result. If you move the fader in any way you're recalculating the audio which makes it go 32 bit float instead of your original 24bit fixed.

That's DAW though, not if you're a shitty digital board where the mix bus is the same bit depth as the tracks you recorded I know many who still use them.... Every 6 decibel loss of gain is a 1 bit gone.
And that's why levels should be a concern, why not get the highest level of resolution before it's processed?
Maybe not to the point you're nit picky about everything, but you should be at least hitting Unity, usually -12dbfs or -18 depending on what you're using.

 

[Soundscaper]

Plec IS SO RIGHT!!!!

Forget as this bitfucking, you'll end in no headroomland soon enough, and your gtrs will starting to sound as if there's some extra (UUUUGLYYYY) distortion on top of them.

I learned it the hard way too... We're talking digital, there's NO reason to record loud in rockmusic! Crank your amps, not your AD-DA nor mixbus. 1 bit is not 6dB btw

Soundscaper.

PS stay away from floating point if you're so scared to loose information in your tracks. Floating point makes you loose EVERYTIME when you go over or under the line

 

[Soundscaper]

Plec IS SO RIGHT!!!! :hotjump:

Forget all this bitfucking, you'll end in no headroomland soon enough, and your gtrs will starting to sound as if there's some extra (UUUUGLYYYY!!!) distortion on top of them.

I learned it the hard way too... We're talking digital, there's NO reason to record loud in rockmusic! Crank your amps, not your AD-DA nor mixbus. 1 bit is not 6dB btw

Soundscaper.

PS stay away from floating point if you're so scared to loose information in your tracks. Floating point makes you loose EVERYTIME when you go over or under the line BITFUCKED again

 

[Soundscaper]

If you don't believe the previous post, give yourself one mix/song/whatever where you try it. It doesn't take that much time, your mix will fir way sooner btw. Then master the missing volume thru the ceiling, you'll get LOUD without digidistortion.

Cheers,

Soundscaper

 

[SPLASTiK]

Plec IS SO RIGHT!!!! :hotjump:

Forget all this bitfucking, you'll end in no headroomland soon enough, and your gtrs will starting to sound as if there's some extra (UUUUGLYYYY!!!) distortion on top of them.

I learned it the hard way too... We're talking digital, there's NO reason to record loud in rockmusic! Crank your amps, not your AD-DA nor mixbus. 1 bit is not 6dB btw

1bit is 6dbfs, that's basic digital theory buddy.
24 x 6 =144, and 144db is the dynamic range of a 24 bit system.

And we're talking peak levels, we're not saying the average recording level should be -6dbfs here. You're average should be around unity which is generally -12 to -18dbfs depending on what you're using.

I never have any Digidistortion problems, but then again the iZ RADAR's AD-DA's I'm using aren't shitty and I don't overload the mixbus by slamming all the faders up.

 

[Soundscaper]

i learnt to use my ears a while ago, no matter what AD-DA you use, a lot of cd's are too loud for domestic and even procdplayers AND ARE digitally distorted. What's even worse is that clients i had sometimes like this
Adding limiters often worses this process btw. Clipping can sound sexier than a lot of plugin or hardware limiters. This again depends on your ad-da's used of course.
You're right about 6 dB = 1 bit when talking track per track dynamics...
and it seems your on my side already since you state "we're talking peaklevel here".
A mastered modernday cd has no real peaks anymore and looks like RMS till -0,03 dB, innit?
However a cd IS 16 bits and by recording your tracks a few BITS softer you don't loose anything, on the contrary, headroom is won and makes mixing easier to 16/44,1, just my 0,02 euro

If your RADAR ad-da is superior, mo power to you!
To they use some compression algorithm?

I tried them out a few years ago and was not especially impressed by them, but that could have been influenced by the (then) lack of fast editing posibilities, and i didn't push the AD-DA's to limits as modern day productions seem to demand back then (as test i mean).
I could be a very strange person when it comes to what i like sounding good, as to me not so many PT setups sound decent, but that's me
Whatever system you use, the pitfalls/benefits only show after numberous projects and examining on various monitoring systems. Splastik, on what do you listen?

 

[SPLASTiK]

If your RADAR ad-da is superior, mo power to you!
To they use some compression algorithm?

I tried them out a few years ago and was not especially impressed by them, but that could have been influenced by the (then) lack of fast editing posibilities, and i didn't push the AD-DA's to limits as modern day productions seem to demand back then (as test i mean).
I could be a very strange person when it comes to what i like sounding good, as to me not so many PT setups sound decent, but that's me
Whatever system you use, the pitfalls/benefits only show after numberous projects and examining on various monitoring systems. Splastik, on what do you listen?

Supposeably the RADAR's AD's are comparable to high end Apogee's. They're pretty clean sounding to me. I usually track on it and do a few basic editing stuff then export to BWF for mixing in ProTools. I don't really like the A/D's in ProTools much.

For monitors good ol' NS-10's and some Genelecs (not positive on the model number but supposeably they're pretty good, 5.1 setup) but even then most of the time I'm using the NS-10's cause I'm goofy.