Page 1 of 2

Gain Structure

Posted: Sat May 03, 2008 10:08 am
by Lo-Key Fu
I notice lately that there has been a fair bit of talk about track clipping, master levels and track levels in various threads dotted throughout the forum so thought I'd start a thread on Gain Structure to explain my own methods and see how others are approaching this often neglected (but very important!) part of mixing.

I organise my own live sets like this:

All individual tracks are levelled to around -6dB output when their master fader is at unity (or if you prefer, 0dB at +6dB on the fader). This is to leave a bit of headroom to play with if required. Volume of the source signal (eg. soft synths, recorded audio, etc.) is adjusted accordingly to achieve this result, and the individual tracks are mixed by adjusting the individual track faders.

If adding FX chains to any part of my mix, I aim to make sure the input volume matches the output. Generally speaking, I don't use FX chains to add gain, but to affect the source signal in some way. I find that markedly changing volume during this process can affect your perception of the changes you are applying; so volume-wise, I like to keep any A-B comparison on an equal footing.

All my individual tracks are then mixed down to one of (6) BUSes which are routed to the Master. I find this to be a convenient way for me to apply effects to groups of sounds (eg. Drum Kit, side-chaining) and render my finished studio tunes into a format I can easily translate to a performance environment.

All of these BUSes are set to unity (0dB) and never adjusted higher or lower; the adjustment is made from the source tracks as noted above. This way, if any are accidentally peaking, I can immediately see the problem as the metering flashes red.

I never use any effects on the master BUS other than perhaps a spectrum analyzer or very occasionally an EQ to remove sub 32Hz bass. This most certainly includes a limiter or compressor as I believe that problems with volume peaks should be addressed in the relevant channel. IMHO, compressing or limiting your master bus should be the sole privilege of the Mastering engineer.

That's about it from me. How are you guys tackling this task? Care to share your approach?

Posted: Sat May 03, 2008 4:46 pm
by roach808
my gain structure normally depends on what program I'm working with.

When I start a track, i keep most of the fader at around -5db and then the master fader around -2.5 or -3.5 depending on the style of music. This way i have plenty of head-room for when the actual mixing of the track occurs. Once, i have all the song written, then i export all the tracks individually and load everything into Logic Pro. When everything is in Logic, all the fader normally get bumped down to -5 as well, but then the master fader gets set at 0, with a EQ to roll off anything below 30hz (i.e. sub mud). This way also have plenty of headroom to play around with when i start compressing diffent tracks and or stem-mixes

Posted: Sun May 04, 2008 3:13 am
by leedsquietman
Your approach is good. Headroom is vital in getting a clean but solid mix, especially when it is going to be pumped up for mastering. -5/-6 seems to be quite a popular threshold.

Although I sometimes use SUBTLE compression with peak limiting on the master bus, but mostly for quick demos. A lot of producers and mix engineers are known to use stuff like the SSL Master buss compressor (or good button) on the master out but too much defeats the purpose of mastering it properly, has to be subtle if at all.

As I am the mastering engineer too on my mixes (can't afford a pro ME and given I've been recording, engineering for 20 years and done courses on mastering and have a reasonable monitoring environment, I ought to be able to make some positive difference) I am at least aware of having used any FX on the master buss and can compensate for it somewhat if I need to.

Posted: Sun May 04, 2008 7:07 pm
by Lo-Key Fu
Cheers for the replies people.

Like you, I often end up being the one to run my own demos through a quick mastering process, so I understand where you're coming from there.

By stating that Master Bus limiting should be reserved for the Mastering Engineer, I guess I was alluding to the fact that I like to keep the mix and the master as entirely separate processes - even if it is me that is carrying out both tasks. If I can end up with a smooth, controlled mix that still has a wide dynamic range without simply slapping on a heavy-handed limiter to solve any peaking problems I find that it makes the mastering process a helluva lot easier for myself too.

I can understand why you might choose to limit your master though, as there is certainly occasion where this process would be useful; but in an ideal audio world I try to avoid it wherever possible and fix any issues in the mixing stage.

Out of curiousity, what are others doing? Are you limiting your master BUS or fixing peaks in the mix?

Posted: Mon May 05, 2008 7:08 am
by fishmonkey
yeah, in general it's (much) more time consuming but (much) better to fix rogue peaks in individual tracks...

Posted: Fri Jun 06, 2008 2:49 am
by sunrahrahrah
how heavy handed are you guys about stopping clipping in all areas? In effects for instance? I usually try to make sure nothing is clipping at all, ever, unless its intentional, but I've seen quite a few videos where people allow tons of clipping with the individual components of an (audio/instrument) track, then just make sure the volume fader doesnt clip the master. What gives?

Posted: Fri Jun 06, 2008 4:32 am
by Nick the Zombie
Watching this thread

Posted: Fri Jun 06, 2008 5:18 am
by Lo-Key Fu
sunrahrahrah wrote:how heavy handed are you guys about stopping clipping in all areas?
I follow the same approach as you sunrahrahrah in that I avoid clipping at every stage of the mix. Can't speak for those who choose to clip in their tracks, but maybe someone who works in this way could shed some light on the decision?

Posted: Fri Jun 06, 2008 7:08 am
by fishmonkey
it's good general audio practice to work so that you don't get peaks above 0 dBFS in any of your tracks...

however, with Live, there is headroom built in to individual internal tracks so that peaking over 0 dBFS does not necessarily mean the signal is clipping and distorted...

the master is different though, and peaking above 0 dBFS there will usually result in audible clipping distortion... for all i know there may actually be headroom on the master too, but since the output from there usually goes to an external device (e.g. soundcard-amp-monitors) which expects 0 dBFS to be the maximum, you will hear clipping even if Live itself isn't clipping... apologies for that torturous sentence...

Posted: Fri Jun 06, 2008 7:14 am
by three
Glad to hear you guys talking about this - I agree in general, but let me add one more thing:

When you are preparing your tracks at -6db, or at unity, or at whatever you happen to choose for a uniform level, remember to do that (if possible) in a single place, i.e., make a single gain change rather than many small ones. If possible, that place should be the synth/sampler/whatever output.

Do gain as early as possible, so that everything that comes after the first element in the chain is outputting the same levels it's receiving.



Re: Gain Structure

Posted: Fri Jun 06, 2008 7:16 am
by a1studmuffin
Lo-Key Fu wrote:That's about it from me. How are you guys tackling this task? Care to share your approach?
Absolutely spot on mate. I approach things exactly the same way.

Posted: Fri Jun 06, 2008 7:33 am
by mdk
like fishmonkey says there is plenty of headroom on the normal tracks so i dont really worry where they are, i just make sure the master doesnt clip.

but overall i generally approach the structure like mr lo-key except that most of the gain control i do on my bus tracks. if i need to automate the volume of a track i'll always use a utility device, that way the fader can still be used to control the overall level.

i never put any plugins on the master (apart from spectrum to see whats going on)

if i've got a problem with a single track thats creating massive peaks then i might use the wonderful blockfish plugin ... &subItem=5

but of course it depends on what the material is on that track, often when its snares (or claps) i dont like to smash them too much so a bit of compression and eq normally does the trick.

Posted: Fri Jun 06, 2008 7:58 am
by three
mdk wrote:i never put any plugins on the master (apart from spectrum to see whats going on)
Yes! This too!

It's soooooo easy to forget things on your master. Much better to have a sum bus that goes straight to master.


Posted: Fri Jun 06, 2008 2:11 pm
by belmiro
I usually drop the faders on individual tracks to ensure -6db on the master channel.

I never thought to use the gain on the individual clips to achieve this instead thus keeping each individual channel at 0db.

Which is the best way, or is either ok?

Posted: Fri Jun 06, 2008 2:40 pm
by andydes
Question for those of you with no limiters or anything on the master and don't send tracks out for mastering.

You render out with maximum peaks at say -3dB. Then re-import the audio file and use a limiter subtlely to bring up the overall level, correct? I've heard this is the best procedure if you're not getting tracks professionally mastered, but not entirely sure about the benefit over having a limiter on the master in the first place. Other than getting to see the wave before you add the limiter.

Or do you just not worry about the perceived volume?