Determing average level of a track

[e_rock]

So I have a newbie question here. Often I'll read comments about tracking at a certain level, averaging around for example, -12 dB, to ensure there is enough headroom for mixing and later mastering.

I am not quite sure how to determine the average of a track though. I use Reaper and it's quite easy to see the peak of a track, but I don't know about average. Is this done solely by watching the meters and making a decisions about where it more or less hovers and using that as an average? Or do folks tend to use software to process the true average of the track?

Any advice would be greatly appreciated!

 

[Farview]

When you are recording, it's pretty easy to determine the average level of the track.

All you have to do is have the musician hold a note on the instrument (or voice) and set the level to -18dbfs. It is the level of the held note that is important when setting record levels. A lot of instruments have large peaks at the beginning of the note, these peaks don't matter (as long as they don't clip) so you just deal with the decay of the note.

For things like drums and other percussive instruments that have a very short decay and a large transient, just set the levels so that the peaks end up around -6dbfs or lower.

The average level of a mix is not as important and, if you recorded at proper levels, will tend to work itself out naturally. (with the exception of productions with 48 or more tracks)

If you still feel you need to know, there are vst plugin meters that will display average levels.

 

[e_rock]

Thanks for the reply, Farview. Could you clarify a bit for me though? Let's take guitars for example and see if I follow you correctly.


Strummed Acoustic Guitar
Have the musician strum a single chord and as it rings out set the level to -18dbfs? I ignore the initial attack (being conscious of not clipping) and try to have the bulk of the decaying sound average about -18dbfs?

Fingerpicked Acoustic Guitar
Have the musician pluck a few notes and again ignore the initial attack and try to have the bulk of the sound average at -18dbfs?


Strummed/Fingerpicked Acoustic Guitar
Let's say the player starts a song by fingerpicking the first few measures and then starts to strum? I'd guess one would have to set the average level for the strummed portion and either record the parts separately or accept that the finger picked portion is quieter during tracking?


Electric Guitar Rhythm
Have the player hit a power chord and have the bulk of the sustained note hover about -18 dbfs?

 

[mixsit]

All those examples are fine starting points. Basically you're in the ball park with the -18'. All that's left is to adjust from there where needed. Like if it's 'mostly light strumming' you clould say that's sorft of the avrage', but even if there's just one point where it get's slammed to full scale then you record it a little lower if need be- increase you headroom. Or, if you know it's comming ride the input or maybe a compressor are the other options.
If it's sustained' only you can track it higher, but there's no particular advantage in it.

 

[Farview]

Strummed Acoustic Guitar
Have the musician strum a single chord and as it rings out set the level to -18dbfs? I ignore the initial attack (being conscious of not clipping) and try to have the bulk of the decaying sound average about -18dbfs?

Yes

Fingerpicked Acoustic Guitar
Have the musician pluck a few notes and again ignore the initial attack and try to have the bulk of the sound average at -18dbfs?

Yes

Strummed/Fingerpicked Acoustic Guitar
Let's say the player starts a song by fingerpicking the first few measures and then starts to strum? I'd guess one would have to set the average level for the strummed portion and either record the parts separately or accept that the finger picked portion is quieter during tracking?

yes

Electric Guitar Rhythm
Have the player hit a power chord and have the bulk of the sustained note hover about -18 dbfs?

Yes.

Just remember that you are just trying to get in the -18dbfs ballpark, you don't have to be right on it at all times. Close is more than good enough.

 

[Massive Master]

And lower is a safer bet than hotter. If in doubt, turn it down and you almost can't go wrong.

 

[e_rock]

Appreciate the advice from all of you. I think I have definitely been tracking way too low. (Certainly better than too hot though.) I was tracking where my peaks hit -12dbfs, resulting with the average being quite low. I should be doing a little recording this weekend. I look forward to hearing the difference.

 

[Massive Master]

I rarely, if ever hit anything above -12dBFS at the input. If I did, I slap myself for being too hot. You're confusing "quite low" with "perfectly typical."

 

[Farview]

It all depends on what you are recording. Drums will have much higher peaks than a violin. The peak level is irrelevant as long as it doesn't clip. The average level where most of the signals power sits is the only thing that is really important. Running that 6 or 12db lower than it needs to be is just as silly (although not as destructive) as running it 12db too hot.

20 years ago, nobody had peak meters so no one really knew what their peak level was. Life was good, everything worked out fine. Nothing has changed except the meter we are looking at, the same rules apply.