Track Meters Question

[Rachel Elliott]

Ok, this may be more of a "newbie" question but I can't find the answer in the manual.

Say I'm mixing a session and I have 1 track whose clipping indicator is lighting up, but the master volume is set low enough that the master meter is within acceptable level.  Will the track with the lit indicator be distorted?  This is a 32-bit 96khz session.  I have tried a listening test, taking one track and maxing it out while keeping the master level low enough not to clip, but I can't tell anything is changing or not.

It depends on whether the indicator is lit because the track was clipped when it was created or not.  But I don't think you mean that because you already know that garbage in = garbage out.  So let's assume it's clipped because 1) post-recording processing of some kind shot it over zero or 2) the track volume you've set raises the gain so peaks are over zero.  But since you're working with a 32-bit float file (which covers the former scenario) and Audition mixes internally in float and not integer (which covers the latter scenario), you're covered.  No distortion.

Just don't try that in Pro Tools.

[Rachel Elliott]

Fantastic!  This is just what I was hoping to hear.     Thanks for the reply!

So let's assume it's clipped because 1) post-recording processing of some kind shot it over zero

Boosting some frequencies with EQ would be the main one, right?

or 2) the track volume you've set raises the gain so peaks are over zero.  But since you're working with a 32-bit float file (which covers the former scenario) and Audition mixes internally in float and not integer (which covers the latter scenario), you're covered.  No distortion.

This one makes me uneasy.  Wouldn't it be safer to lower the 'clipping' track to a safe level, drop all the other tracks by a similar amount to keep the mix right, and then boost the master gain?

I don't think so, because there's nothing inherently unsafe here.  It's either clipped, or it's not, and placing a 100Hz tone on a track, raising it past clipping, lowering the master fader, and mixing down will show you that AA is trustworthy.  I'm not saying this is good engineeering -- what you've described is much better -- but good technology does allow for a certain amount of sloppiness in certain areas.

Yes, channel EQ is a great example of how to take a loud track and to boost it over "0," as would potentially any other "additive" effect.

I'm not saying this is good engineeering -- what you've described is much better -- but good technology does allow for a certain amount of sloppiness in certain areas.

The only problem with running your mixes that hot is that you are still monitoring via a soundcard - and this, whatever it's being fed, is still operating in integer mode, as all hardware does. So although your signal isn't distorted per se, it might still sound distorted when you play it back in the mix, because you've hit the hard 0dB point of the converter - which means that you aren't really listening to what you will finally get in the way of sound. Basically this means that the FP signal has asked the soundcard output to do something that it's inherently not capable of, even though the signal itself is.

So as a safeguard, and a way of preserving resolution FP mixes are fine - but when it comes to monitoring, dobro's approach is better - but not for the reason he's suggesting. When it comes to the actual levels, Audition really doesn't care what they are - you want to boost all your levels by 100dB and have a mix that's a solid block of green? Well that's fine - you can normalise it back without loss. But if you try to play it at this level in MV, then... well, I don't fancy being your monitors!

What it comes down to when monitoring a mix is that you can have the channel levels set to whatever you like - as long as the balance is correct. And if it's hot, then reducing the master level should let it sound okay during monitoring, regardless.

Yeah, I understand that if the original recorded track wasn't distorted, then processing gain that makes it clip in MV won't damage the original file, it'll just 'damage' the sound and queer the mixing process.  But if you then did a mixdown of that multitrack session at the same master gain level, then the distortion you heard in MV would also be heard in the mixdown, right? 

[Rachel Elliott]

Yeah, I understand that if the original recorded track wasn't distorted, then processing gain that makes it clip in MV won't damage the original file, it'll just 'damage' the sound and queer the mixing process.  But if you then did a mixdown of that multitrack session at the same master gain level, then the distortion you heard in MV would also be heard in the mixdown, right? 

Try it out, I did.  I took a recording of a bass drum (incidentally it is this instrument that caused me to raise this question) and cranked the tracks volume, then backed off the master volume and mixed down.  Then I lowered the volume of the track and increased the master volume to make up for the change, then mixed down.  Listening to both files, I can't hear a major difference between the two.  Although I must say that I did this test very fast and did not spend a lot of time listening.  Visually, they seem to be the same.  But I thought I would ask here for a quick answer.

Audition didn't always have individual track meters, certainly not the way it is now.

The above answers are all correct, but one thing to watch out for is whether the extra hot signals are going to be fed to an effect, especially Dynamics Processing. There are at least two ways an "over the top" signal could find its way to such an effect, in which case it may not work quite the way you expect.

Steve and Alan's points are very important.  Soloing tracks when building a mix is important to me, and soloing one with over-0 peaks would result in inaccurate (clipped) monitoring; and yes, we all use dynamics processing at some point or another.  The concern would apply to anything level-dependant.  I suppose the workaround there is to insert the effects prefader?  That can have undesirable behavior too.

But purity is unquestioned.  Take Despised7's scenario.  In fact, take any session.  Raise your track levels by a uniform amount that induces clipping -- say 20 dB.  Lower your master 20dB.  Mix.  Revert to previous sane levels.  Mix.  Invert paste the two mixes.  Unless there's a level-dependent effect in your mix, they should be exactly identical and cancel out perfectly.