So I get that I should not go over 0db when recording, so I plan to record at around -10 or -6. What about after recording. When you are mixing should they stay under 0db and not reach the red zone in the track meters?

Sorry if this is a simple question

Well Jeff, dcbs is 'supposed to be' dbs ... hahahaha!   My grammar can be atrocious ... especially since I'm a podiatrist.  I'll fix that in future posts.

Everything I referred to is peak and not RMS (as I indicated in the 1st couple of examples).

I thought the Op was looking for recommendations on how mixers refrain from going above -6 dbs on most tracks, "exceptions", etc. ... with their recorded tracks by using envelopes. 

Or he could have meant 'live' recording's or such ... which is done outside of Sonar.  If thats the case my thread is pointless.

Of course I agree with your thoughts that the peak numbers are dependent on many variables ... and may even be as neurotic as the music itself. 

But, Bat got me thinking about drums being dominant.  So I thought it dutiful to validate his generous peaks for drums.

I think you might agree that one of the greatest challenges is getting the right levels for kick and bass ... on say, a rock mix.

Hi Phillip. Two signals can have a very close rms value but the peak values may be quite different. That is why I go on about rms and peak levels. eg a signal may have an rms value and be at -14 db FS. The peak value might only be 1 or 2 db higher. (eg the attack transient reached -12 db FS). But overall the signal is at -14db FS. Another signal may have the sustained part of its sound also at -14 db FS but a peak that is 8 db higher than the first signal. (Its attack transient reaches -6db FS)

Peak wise they will show something different. (one will show say -12 db FS and the other will show -6 db FS) But as the rms values are the same and rms defines the real power in the signal they will both sound equally as loud. (A VU meter will show both of those to be the same value, that is why VU's are so great. Because a VU meter needs 300ms or more to reach FSD so both attack transients wont effect them too much)

I guess it is OK to recommend mix levels but if you do then it is also important to indicate at what levels the tracks are recorded at and if a ref level is being used then what that was at the time. You can only do it if say all your tracks are tracked at say K-14 or the rms values of all your tracks are similar and very close to each other.

Kicks and snares can also be metered with a VU meter. You can still see the power in the signal and a decent kick and snare can easily make the VU reach FSD. It does depend a bit on how much sustained portion of the sound is left behind. A very tight damped kick may not move a VU a lot but that is where our peak meters come into their own. You need both I believe. (peak and rms metering) If you only rely on one then you are stabbing in the dark at times.

sharpdion23


So I get that I should not go over 0db when recording, so I plan to record at around -10 or -6. What about after recording. When you are mixing should they stay under 0db and not reach the red zone in the track meters?

Sorry if this is a simple question

During the actual mixing part there's no reason to actually go above those levels either. At the end, when you have the mix like you want, you can bring it up to final levels. You can either just bring up the levels based on your current peak so that it now peaks a fraction of a dB below 0. Or you can bring it up and use some sort of mastering tool with a limiter to take it up even higher (don't abuse it.)
 
You can of course also just export it as is and use some sort of external mastering tool as well to bring it the rest of the way up.
 
Of, if you are going to send it out for mastering, just leave it as is and don't try to bring it up to 0dBFS.
 

Jeff,

I realize that the newer mastering modules use the K system meters based on Katz criteria.  Sonar has used it on the track levels, iirc in vertical and horizontal meters.

And yes, I can see that I don't look at my RMS values enough!

Agreed that the peak values alone are quite 'shaky' to recommend.  My thoughts were penned in haste ... probably because I struggle greatly keeping the peaks from 'summation' clipping.

'Summation' clipping becomes more a problem in final stages of a mix ... when I struggle to get the kick and snare dominant while preserving transients.  'Live drums' yield stray transients that clip if some rule is not laid down.  I suppose transparent limiters would help with drum samplers that record themselves on the fly (another topic).

It might be good idea to talk about levels on busses. So far I have been talking about tracking using the K System. If you choose a ref level of K-14 then your tracks should all have signals recorded on them that are averaging an rms value around -14 db FS and the peaks just end up where they end up.  (with K-14 there is only 14 db of headroom but usually enough for most signal transients)

All I do regarding busses is do the same thing there. If I send a bunch of tracks to a buss then obviously if the tracks were all around unity then the combined effect of them would overload a buss bigtime. But by pulling individual tracks down accordingly then all you have to aim for is a total of K -14 on a buss. Same thing applies there. Average combined rms value will be around -14 and the peaks will still just take care of themselves. Rarely will anything clip a buss.

For the master buss I am just adjusting the busses so the combined total signal on the masterbuss is also K-14. Average rms value of the whole mix is still around K-14 and peaks will still take care of themselves and will rarely clip. Keep the volume up a bit in the room so the mix sounds powerful and big.

How do you do all these measurements? With VU meters of course.  And the good news is that the Klanghelm VST meters do a great job. (Cost under $10, you get a mono and stereo meter for tracks and busses) Just set the ref level and away you go. The meter will show FSD or 0db VU when your ref level is reached rms wise that is.

Big punchy drum sounds with lots of attack and punch are very possible on the drum buss for example. You can do all this at K-20 if you are worried about very transient signals clipping anywhere. It just means you have got further to master it up level wise to get it up to commercial final levels.

The problem I think many have is trying too hard to create a loud and powerful mix right from the start and that is where clipping starts to make itself apparent etc. People are trying too hard to get right up near 0db FS during tracking because they think it is best up there. Not so. They start overdriving busses and the final masterbuss. All this dumb talk about what happens in Sonar and other DAW's when you go over 0db FS. But you don't have to come anywhere near it!

You don't do that at all. You just create this lovely K -14 master which is not overloud of course but punchy with no clipping in sight. Then you master in a separate session and a week later as well. That is the time to start using EQ, compresssion and limiting to get those average levels well up into commercial territory. Of course I would love it to just stop there. You should hear how fantastic  a K-14 or even a K-20 master sounds turned right up loud.

But for the moment the client wont accept it like that. But I believe with some careful compression and use of a great limiter such as the PSP Xenon you CAN achieve very loud masters that are still punchy and kickass.

Jeff Evans


Hi Phillip Just a few observations. 

What is a dcbs? Is this your slang for decibels. Others like my self may be wondering what unit you are referring to. Decibels are best abbreviated to db. That is the convention.

[pedantic]
Sorry, but that's not the correct convention, either, Jeff.  The correct abbreviation for decibels is dB, not "db."
http://en.wikipedia.org/wiki/Decibel
I now return you to your discussion. 
[/pedantic]
 
 

Thanks Beagle you are correct, now that is me being slack. At least it is better than dcbs!

Jeff Evans


Thanks Beagle you are correct, now that is me being slack. At least it is better than dcbs!

lol!  agreed!
 
this is a great thread, tho.  great info being shared.

Thanks Beagle for your thoughts on the thread. It is interesting and recording and mixing levels have always been one of my interests.

If you keep all your peak levels constant which is what people tend to do and there is nothing wrong with that, the only problem is that the peak levels don't necessarily tell you much about the rms levels that are below the peak levels. They can be all over the place while your peak levels are all constant.

Then when you come to mix you may be wondering why some tracks are way too soft for example and have to gain or trim them up. It is because that track has a signal with a high peak level but low rms level. And you have metered the peak rather than look closer at the rms value.

The K system is all about keeping rms levels constant and not worrying too much about the peaks. Just as we used to do back in the analog days. Even if a DAW shows rms values, they are usually way too low down on the scale to effectively read them. The VU meter is great because it can ignore peaks and only show rms and also show your digital ref level right up at FSD or 0 dB VU which is right where you need to see it. Some other DAW meters can be put into K system rms VU mode then they show 0 dB VU with the headroom above that. (I wish more DAW's would take this on) But even if your DAW cannot do that the Klanghelm meters will do it very well. And I am pleased to say the ballistics are very close to the real deal. (with my settings of course)