Pushing the Faders in a DAW

[electraplay]

Hi all...
I use Sonar 5 producer and am beginning to mix a project which averages about 20 tracks per song. I seem to be making a habit of pushing the individual track faders up to a point where, even though they are not clipping, the master bus is going over by 3-5 or even more db's.

First, I'm not sure I understand how if none of the individual tracks are close to clipping, why the master is?

And more importantly for my purposes...what strategy should I employ here? I have just been cutting the master fader by 3-6db's to the point where I won't clip but it feels like I might be sacrificing resolution or overall sonic quality....How important is it to keep the master bus around unity gain (0dbs)?

thanks for any input here!

[Lagerfeldt]

This is an often misunderstood subject with much voodoo talk.

I'll try to write as clear as possible.

First the explanation: your master bus is overloading (clipping) because it is the sum of all your channels. Let' say, if your master bus can only contain a digit as large as 10, and all of your 16 faders each containing a value of 1, your master bus will be overloading with a value of 6. Get it?

As for how to avoid this situation, or if avoiding it is necessary at all:

The quick version: to be on the safe side, do not clip your master bus (or your channels if you require fidelity in your sound). This is done by lowering all your channel faders until no clipping occurs in your master bus.

The longer version:
In a floating point sequencer (this is how it calculates data) there is no - and I repeat - no difference in sound quality between A. raising your faders (clipping the master bus) and then lowering the master fader until it stops clipping, and B. lowering your faders and raising your master fader (or leaving it at unity). This is easily proven with a null test (a phase flip test where identical files will null out when played simultaneously).

But the story doesn't end there. Because that's rarely how we mix in a computer. We like output to busses and sum compress (e.g. drums) or in general add plug-ins to the bus and/or master bus. And then you have a problem if you're coming in too hot, even though the fader doesn't show any clipping going on.

That's because now you're overloading the input side of the plug-in, and that's another story. You can test this by inserting a limiter plug-in such as the Waves L1/L2 on a bus or master. Make sure the master is actually overloading but you've lowered the fader so it doesn't show. Notice how the limiter is already active without you having adjusted the threshold.

Finally there's no reason to aim for such hot mixes in 24 bits. A peak of around -6 to -3 dBFS is more than adequate. Most mastering engineers prefer a mix that doesn't peak any higher than that.

[JesseJ]

Well put Lagerfeld.


- Jesse -

[theblue1]

Couldn't have said it better myself... but it probably would have taken me about three times as many words and I probably would have worked in a story or two from my childhood.




Our DAW mixers work just about like a 3DW mixer for these purposes...

You don't want to clip your output converters or any fixed point audio file -- and many or perhaps most plugs are optimized for the "normal" sub-0 dBfs range so you've got to watch out for them along the way.

Reminds me of a time... musta been back around '64 or so... when...

[electraplay]

thanks so much...i actually understand now!!!

[Mosby]

I agree with the above. I would also restate that mixing super hot isn't necessary. If "everything" needs to be louder -than leave it until mastering. Concentrate on getting the mix right.

[zaczac]

Originally Posted by theblue1

Couldn't have said it better myself... but it probably would have taken me about three times as many words and I probably would have worked in a story or two from my childhood.

Ha, nice.

[jerdude]

Originally Posted by Mosby

I agree with the above. I would also restate that mixing super hot isn't necessary. If "everything" needs to be louder -than leave it until mastering. Concentrate on getting the mix right.

uh. except that when loudness... in your face compression... is equated with IMPACT. if your mix is peaking at -6db... no mastering engineer is going to be able to get it to commercial level without drastically changing your mix. the drums will be softer... the lead vocal softer... and everything else much louder.

don't worry about mastering... don't worry about where you are peaking... unless you are clipping and you hear unpleasant distortion... which is usually the case when clipping the master buss of a DAW... well.. most DAW's. if your vibe is audioslave... then mix it loud... if your vibe is alison krauss... mix it soft, warm and intimate.

[PettyCash]

Originally Posted by jerdude

uh. except that when loudness... in your face compression... is equated with IMPACT. if your mix is peaking at -6db... no mastering engineer is going to be able to get it to commercial level without drastically changing your mix. the drums will be softer... the lead vocal softer... and everything else much louder.

don't worry about mastering... don't worry about where you are peaking... unless you are clipping and you hear unpleasant distortion... which is usually the case when clipping the master buss of a DAW... well.. most DAW's. if your vibe is audioslave... then mix it loud... if your vibe is alison krauss... mix it soft, warm and intimate.

My personal aim has always been to try and get the mix to sound exactly like how I want the finish product to sound. If it means my peak is going to be higher than -6db without actually clipping, then so be it.

[Lagerfeldt]

Originally Posted by jerdude

If your mix is peaking at -6db... no mastering engineer is going to be able to get it to commercial level without drastically changing your mix. the drums will be softer... the lead vocal softer... and everything else much louder.

That's not correct. You seem to fundementally misunderstand how things work.

Delivering a 24 bit mix that peaks at e.g. -12 dBFS, -6 dBFS or -3 dBFS will basically have 0% influence on the final sound or loudness of the mastered track. What's relevant is how the song is mixed i.e how the different parts in the song are mixed (and compressed/limited) in respect to each other. Not having the final mix song peaking near 0 dBFS.

If anything, delivering a mix that peaks at -3 dBFS or above could have a detrimental effect on the final outcome, as you risk the mastering engineer having to attenuate the whole signal digitally (depending on what chain he's using).

Even assuming he's only cutting (not boosting) at the first level of EQ, phase changes will likely cause the mix to overload when printed very hot, thus introducing an extra processing step of unnecessary attenuation across the whole spectrum.

[gproper]

Originally Posted by Lagerfeldt

That's not correct. You seem to fundementally misunderstand how things work.

Delivering a 24 bit mix that peaks at e.g. -12 dBFS, -6 dBFS or -3 dBFS will basically have 0% influence on the final sound or loudness of the mastered track. What's relevant is how the song is mixed i.e how the different parts in the song are mixed (and compressed/limited) in respect to each other. Not having the final mix song peaking near 0 dBFS.

If anything, delivering a mix that peaks at -3 dBFS or above could have a detrimental effect on the final outcome, as you risk the mastering engineer having to attenuate the whole signal digitally (depending on what chain you're using).

Even assuming he's only cutting (not boosting) at the first level of EQ, phase changes will likely cause the mix to overload when printed very hot, thus introducing an extra processing step of unnecessary attenuation across the whole spectrum.

Your Good.

[jerdude]

Originally Posted by Lagerfeldt

That's not correct. You seem to fundementally misunderstand how things work.

Delivering a 24 bit mix that peaks at e.g. -12 dBFS, -6 dBFS or -3 dBFS will basically have 0% influence on the final sound or loudness of the mastered track. What's relevant is how the song is mixed i.e how the different parts in the song are mixed (and compressed/limited) in respect to each other. Not having the final mix song peaking near 0 dBFS.

If anything, delivering a mix that peaks at -3 dBFS or above could have a detrimental effect on the final outcome, as you risk the mastering engineer having to attenuate the whole signal digitally (depending on what chain he's using).

Even assuming he's only cutting (not boosting) at the first level of EQ, phase changes will likely cause the mix to overload when printed very hot, thus introducing an extra processing step of unnecessary attenuation across the whole spectrum.

um. this explains nothing. you've already said all this.

i've seen and heard alot of pre-mastered mixes done by tops guys on SSL's. i've never seen or heard a mix peaking at -3db... let alone -12db. that's just ludicrous. most had peaks near 0db and already had a RMS around -12 to -14 db. i'm talking about guys people have heard.

granted if a mix is slammed with compression (which i don't recommend) but is only peaking at -12db... bringing it up to commercial level won't be a problem.

no one should ever mix a song thinking about the mastering engineer.

My personal aim has always been to try and get the mix to sound exactly like how I want the finish product to sound. If it means my peak is going to be higher than -6db without actually clipping, then so be it.

exactly.

[Mosby]

Originally Posted by jerdude

uh. except that when loudness... in your face compression... is equated with IMPACT. if your mix is peaking at -6db... no mastering engineer is going to be able to get it to commercial level without drastically changing your mix. the drums will be softer... the lead vocal softer... and everything else much louder.

don't worry about mastering... don't worry about where you are peaking... unless you are clipping and you hear unpleasant distortion... which is usually the case when clipping the master buss of a DAW... well.. most DAW's. if your vibe is audioslave... then mix it loud... if your vibe is alison krauss... mix it soft, warm and intimate.

WEll, I guess I should have been more specific. Getting the mix right is more important than getting it loud was my point. "Right" is a sunjective term - I was just saying that if *all* you need is louder, than there are other ways to do that (limiters, or just turning everything up at the end (not that those two are at all the same thing)). My point is that getting the mix right should be a bigger deal than how much head room you have on the master.

[jerdude]

Originally Posted by Mosby

WEll, I guess I should have been more specific. Getting the mix right is more important than getting it loud was my point. "Right" is a sunjective term - I was just saying that if *all* you need is louder, than there are other ways to do that (limiters, or just turning everything up at the end (not that those two are at all the same thing)). My point is that getting the mix right should be a bigger deal than how much head room you have on the master.

i couldn't agree more.

[George Necola]

and just keep in mind, that digital levels are not analogue levels.

if you have -3dB digital peaking, you are counting samples, not measuring analogue signal. that is why things like the sonalksis free G or Inspector XXL are showing diffrent results than your sequencer.

cheers

[H-Rezz]

Originally Posted by jerdude

um. this explains nothing. you've already said all this.

i've seen and heard alot of pre-mastered mixes done by tops guys on SSL's. i've never seen or heard a mix peaking at -3db... let alone -12db. that's just ludicrous. most had peaks near 0db and already had a RMS around -12 to -14 db. i'm talking about guys people have heard.

granted if a mix is slammed with compression (which i don't recommend) but is only peaking at -12db... bringing it up to commercial level won't be a problem.

no one should never mix a song thinking about the mastering engineer.



exactly.

Man you got this all wrong , never heard a mix engineer send me back a master at near 0db, nor do i do that myself , they would only do this for pre-mixes handed to record companies to bring up listening levels , but not something they would send off to be mastered ....

And of course you have to take the mastering process into consideration , mixing is about balance,seperation,soundscape and anything else you can think of but not about maximum level, this is simply not the stage it's done at ! The mastering dude needs at least 2-3db to do his thing and give you a great sounding product ! At near 0db he has no more headroom to work with and would somehow have to find a way to drop the level !

If you are wondering where your mix should sit , call a mastering house and ask them how hot should you print your mix for them , this might clear things up ...

[Lagerfeldt]

Originally Posted by jerdude

um. this explains nothing. you've already said all this.

i've seen and heard alot of pre-mastered mixes done by tops guys on SSL's. i've never seen or heard a mix peaking at -3db... let alone -12db. that's just ludicrous. most had peaks near 0db and already had a RMS around -12 to -14 db. i'm talking about guys people have heard.

Selling more than 6 million copies I'd say a few people have heard me too :-)

Well, you need to get out more then if you think -3 dbFS or -12 dBFS is ludicrous. And talk to some more mastering engineers too. Again, I'm talking about peak levels, not RMS.

You obviously still don't understand how things work but seem intent on insisting on something that's simply not correct without providing arguments (which don't exist of course). There's actually a lot of good information on this subject for you to read, both on Gearslutz and perhaps using my link list here: Popmusic.dk - Links - The Ultimative Link List - Articles - Free Plug-ins - Gear - Equipment - Links Overview

I'm not saying some people (professional or not) aren't delivering mixes that peak at -3dBFS or above, I'm just saying what's best and what most people deliver to mastering. And I'm providing you with a techincal founded explanation to why it's so.

[George Necola]

Selling more than 6 million copies I'd say a few people have heard me too :-)

he did aquagirl.. you know, you were the guy I wanted to kill the most when this song was playing up and down in the radio

now, it's timbaland

[Lagerfeldt]

Originally Posted by George Necola

he did aquagirl.. you know, you were the guy I wanted to kill the most when this song was playing up and down in the radio

now, it's timbaland

If you're thinking of Barbie Girl with Aqua, actually that's Johnny Jam & Delgado (also Danish), which dwarfs my 6 million with about 12 million or so. I did do a medley for Aqua though.

[George Necola]

Originally Posted by Lagerfeldt

If you're thinking of Barbie Girl with Aqua, actually that's Johnny Jam & Delgado (also Danish), which dwarfs my 6 million with about 12 million or so. I did do a medley for Aqua though.

I hate you

[Boogle]

Originally Posted by George Necola

I hate you

+1 .

[The MPCist]

This is what I hate about digital. In the good ole analog days, I could keep piling up tracks as I go along (recording BV) and the levels pretty much stayed constant... but with digital, doing the same thing, recording BV vocals, the levels just keep increasing and you've got to keep pulling individual faders down as you go along recording...

[Lagerfeldt]

Yeah but don't forget you were substituting headroom for subjective distortion. Not necessarily a bad trade off - mayb even desirable - but not objectively better.

I actually like the hybrid workflow/sound of digital ITB mixing and analog outboard processing, the best of both worlds.

[Ashermusic]

Originally Posted by Lagerfeldt

This is an often misunderstood subject with much voodoo talk.

I'll try to write as clear as possible.

First the explanation: your master bus is overloading (clipping) because it is the sum of all your channels. Let' say, if your master bus can only contain a digit as large as 10, and all of your 16 faders each containing a value of 1, your master bus will be overloading with a value of 6. Get it?

As for how to avoid this situation, or if avoiding it is necessary at all:

The quick version: to be on the safe side, do not clip your master bus (or your channels if you require fidelity in your sound). This is done by lowering all your channel faders until no clipping occurs in your master bus.

The longer version:
In a floating point sequencer (this is how it calculates data) there is no - and I repeat - no difference in sound quality between A. raising your faders (clipping the master bus) and then lowering the master fader until it stops clipping, and B. lowering your faders and raising your master fader (or leaving it at unity). This is easily proven with a null test (a phase flip test where identical files will null out when played simultaneously).

But the story doesn't end there. Because that's rarely how we mix in a computer. We like output to busses and sum compress (e.g. drums) or in general add plug-ins to the bus and/or master bus. And then you have a problem if you're coming in too hot, even though the fader doesn't show any clipping going on.

That's because now you're overloading the input side of the plug-in, and that's another story. You can test this by inserting a limiter plug-in such as the Waves L1/L2 on a bus or master. Make sure the master is actually overloading but you've lowered the fader so it doesn't show. Notice how the limiter is already active without you having adjusted the threshold.

Finally there's no reason to aim for such hot mixes in 24 bits. A peak of around -6 to -3 dBFS is more than adequate. Most mastering engineers prefer a mix that doesn't peak any higher than that.

Very thorough and complete explanation Lagerfeldt.

[lofi]

Originally Posted by The MPCist

This is what I hate about digital. In the good ole analog days, I could keep piling up tracks as I go along (recording BV) and the levels pretty much stayed constant... but with digital, doing the same thing, recording BV vocals, the levels just keep increasing and you've got to keep pulling individual faders down as you go along recording...

do a BV submix.

treat yellow as red.

there

[bachconnelly]

Originally Posted by Lagerfeldt

This is an often misunderstood subject with much voodoo talk.

I'll try to write as clear as possible.

First the explanation: your master bus is overloading (clipping) because it is the sum of all your channels. Let' say, if your master bus can only contain a digit as large as 10, and all of your 16 faders each containing a value of 1, your master bus will be overloading with a value of 6. Get it?

Now, maybe I'm wrong here.. but I've learned that this isn't exactly correct. Doubling the amount of signal in dBFS in the box will only yield a +3dB increase on the master bus. That is why when you have two things that peak at -16dB your master bus peaks at -13dB. Am I wrong here?

D-

[kafka]

Originally Posted by Boogle