Volume lowers when I export audio, but sounds good in editing mixing session

[Stefie47]

Hi all, I did a simple mix of a couple of tracks. The volume of the tracks sounds great while I'm editing in the working session, but once I export the mixdown, the audio volume is significantly lower. What's happening? Is it my export settings?

[SteveG_AudioMasters_]

No, not the export settings - it's all to do with Multitrack Pan laws. Here is an explanation that I wrote in another thread, a while back:

What happens depends mainly upon two things. Firstly, the settings on the Multitrack mixer panel, and secondly the pan law setting you used when setting up the multitrack session (in Edit>Preferences). If the mixer levels are set so that all tracks at at 0dB, and the pan law was set to L/R cut - Logarithmic, then all the levels will transfer between Multitrack and Waveform views correctly. But if you created the session with the multitrack pan law set to -3dB center, then the track will appear to be 3dB lower in level in Waveform view. This is because panning is effected by using equal power settings, and compensating the channel levels accordingly. What this means in practice is that the level of each channel has to be reduced in Multitrack view to take account of what happens when the pan control is advanced fully to one side - that's the point where the channel level will appear 'correctly'.

L/R cut - Logarithmic doesn't do that - it achieves panning by reducing the amount of the other channel, and leaving the first one alone. These settings only make a difference if you are actually panning a mono source across a stereo sound stage. With the -3dB center setting, the sound will remain at the same apparent level right across it, as the apparent increase you'd get in the centre (dB addition) is compensated for. With the L/R cut, the law is somewhat different, and the panning won't be as smooth at all. If you are only placing a signal into a sound field and not moving it, it makes no difference which you use - until you do what you're doing, and jumping between Waveform and Multitrack.

One 'solution' to this that works reasonably well is to use -3dB center for your setting, and advance the master fader to +3dB. The apparent level between the two playbacks should remain the same. But this means that you have to have the channel fader also set to 0dB, otherwise all bets are off as to what you hear.

Incidentally, the information about which law you used is written into the session file, and also you won't be able to get the results from any change you make until you've shut down and restarted Audition. That catches out a few people. I believe it's possible to edit your session file to make it use the other law, but we'd need Suite spot to remind us how to do that, I think...

[Stefie47]

Hi SteveG, thank you so much for your response. I saw your reply on another post regarding the -3dB center setting. That was very helpful. I switched to L/R cut logarithmic when I saw your reply. And thanks for elaborating further on different solutions. I did try the one "solution" you mentioned, but I'm not noticing a change. How can I set up my multitrack session settings so that all the levels transfer correctly? And where can I see which law I used in the session file? I would love to get this resolved before I dive into my project. 

[SteveG_AudioMasters_]

I think that the reason you won't notice a change is because you've tried to apply it to a session that already exists - that's where we need SuiteSpot's input to tell you what you need to change in the session file. I'm reasonably sure that he's already outlined this in another thread, but I couldn't find it when I looked. If I get a chance, I'll have another look later.

But in a very real sense, it doesn't matter in the slightest what your mix levels are (except that if they're too high, you'll upset your monitoring whilst actually mixing). You can have a mix that peaks at -50dB, and if you normalize it afterwards in Waveform view to -1dB, it will sound fine - that's the absolute joy of a Floating Point mix; no losses at all from mixing at the 'wrong' level, as the bits themselves don't change - only the scaling value, and you can set that to anything you like. So if you whack the master volume up 15dB on your mix, and it opens up in Waveform as a sea of green, just normalize it down to -1dB and it will sound just as it did before you did that!

So there's no such thing as a 'correct' value at all when it comes to a 32-bit Floating Point mix!

[Stefie47]

Hi SteveG, 

Thanks for your response. I've been looking into this since I posted the query. I don't know why I'm having so much trouble with this. Essentially what I'm trying to do is just boost the volume of the .wav file that was recorded in a professional studio because it's too low compared to other tracks I'm using (i.e. music beds). I'm editing in multi-tracks. I'm wondering what is the proper way to increase volume to the track without distorting? Is it hard limiter, the volume toggle, or clip gain?

Also, what is the difference between the three? Any guidance is much appreciated. Thank you so much for your support. 

[Stefie47]

I also forgot, what is pan and is that a way to increase the volume of a track? Thanks. 

[JoeF68]

I am having this same issue and tried the solution mentioned by SteveG didn't help. When "Exporting Mixdown" to an MP3, the volume of the exported file is significantly lower in the final product. Is there another way to boost the volume?

[SteveG_AudioMasters_]

It's generally difficult to control the levels of an export accurately, as how 'loud' it is depends entirely upon how the Master channel is set up - and because it's a 'virtual' system, you can't normalize the output from it - which is what you need to do in order to determine what it actually is. This is why you shouldn't really use 'export' for this purpose. The correct way to do it is to create a mixdown of your track, which then opens in Waveform view. Because it's a 'real' file at this point, you can now run Normalize on it (and set the level to about -2dB if it's an MP3) and you'll know that what you've set is the accurate peak level of the file.

This isn't really about 'boosting volume' - it's about setting the level of the file to make maximum use of the dynamics available, which is what you need to do accurately, and that's what I've described. It doesn't matter what the peak level in Multitrack is, because the mix is done in a Floating Point format, which is why you can't normalize it there - it's virtual, not real. It's only when it's real that you can set the absolute level with any certainty at all.