How I approach Gain Staging in DAW Production
by Jeff Evans
When we look at how we managed gain staging in the days of mixers and reel to reel multitrack analog tape recorders, there were conventions that were designed for a good reason and they actually worked!
In mixers we aimed for a uniform level flowing down the channel strip once the correct amount of front end gain was either added or subtracted to the input signal. The levels arriving at mixer inputs can be varied and over a very wide range of levels. Eg from Mic levels -60 dBu on a quiet day, to -40 dBu up to 20 dBu for some synths, then into domestic line levels such as -10 dBV or -8 dBu. 0 dBu and +4 dBu are also popular line level reference levels. The reference level travelling down a mixer channel strip may have varied somewhat too but at the end of the day the required gain was added to bring the final output up to a reference standard such as 0 dBu or +4 dBu.
Our DAW track, buss and stereo mix levels can also be at a reference level. We can apply this technique across to DAW operation. With a decent hardware mixer, an average level of +4 dBu output coming from the main outputs should be able to reach at least +24dBu before the onset of clipping. Indicating 20 dB of headroom. Headroom in mixers is built into the design. Headroom can be built into the DAW by making the reference level somewhere well down and away from 0 dB FS such as -18 or -20.
With these we tended to record everything on each track at the nominal reference level which is when the VU meter read around 0 dB VU on the scale. Going In and coming OUT. Apart from percussion and fast transient sounds, the average rms level on each multi track was the same. The ref or 0dB VU. Headroom existed due to the very nature of tape itself. It had built in headroom. 15 dB or so. Before tape machines start to sound compressed and crunchy. They added hiss into the equation.
DAW operation simply means we can aim to record tracks at a given reference level as well using the VU meter as a guide to most input sources. The headroom is also built in because of the chosen reference level. There is no need to record tracks to the point of clipping or distortion or loud. In fact it is one of the finest things about digital multi track recording. The tracks can easily be noiseless and have zero distortion. Exposing the recording itself making it as naked as it can be. Better had record things nice then!
Types of Complex Signals
Any complex signal had two parts to it. The attack transient is followed by the meat of the sound for much longer usually and this part of the signal tends to sit at a reference level for a longer time. Up to seconds in fact. Anything over 100 mS is considered a long time in terms of rms measurement. Peak meters are good for the initial fast attack transient as they are fast enough to show even the shortest transient eg 5 mS and the rms meter (VU) is good for the rest of the signal. The VU is slow and takes 300 mS to reach its full scale deflection. The VU shows the solid part of the signal and for longer. It is better to monitor a complex signal with both types of meters especially when the signals are in the percussion and drums area. But for a great majority of signal types the rms meter works well alone.
Gain Staging in your DAW
One thing that works is that if concern yourself with mostly the rms levels being consistent in several parts of your production signal flow, then you will be safe in the knowledge that the available headroom is coping very easily with not only average rms levels but the peaks too. Without them crashing into 0dB FS. The peaks are varying all over the place in relation to the rms components of the signal. But none of them are really crashing into 0dB FS so there is no clipping mostly everywhere so no red clip lights are coming on either. An ideal situation. I only see a red clipping light on the odd occasion. On the master buss a limiter can be set there to nothing other than limit the output peaks to some value below 0 dB FS eg -0.5 dB FS. That will keep your master buss from clipping as well. Elegantly that is. The overall mix is more likely to cause an overload clip light.
Keeping an eye on rms
It’s about keeping your rms levels at a comfortable level on the digital scale. And combining that with a decent amount of headroom above to take care of those wildly varying peaks. Without them showing any clipping. Keeping everything super clean in fact. Taking full advantage of the digital recording medium. Because even when you select a reference level like -20 dB FS as being 0 dB VU for example or the K System -20 reference, if you are working initially at 24 bit resolution then the theoretical digital noise floor is -124 dB below -20. Even if you shave off 6 dB for real world performance, then that still puts the digital noise floor around -118 dB (below -20 that is!) which, is a long way down or damn quiet!
No need to worry about having to record at loud digital volumes now. Think, lower the average rms levels everywhere in your system. And put a VU meter set for the desired reference whenever you want. So you need to think about what reference level you are going lock 0 dB VU to.
Monitoring the VU rms level before and after any important plugin is a great thing to do. Before, during and after complete plugin chains is also a good idea. By keeping rms levels at a nice ref level like this you will never clip a plugin again either. Many plugins are calibrated to work at -18. Using a -20 ref level is no biggie here either. It is just dropping the level by 2dB which is increasing the headroom in reality.
Check the level you sending into a reverb for example. Important. It is easy to over drive some reverbs and not know it at first.
Many plugins have output controls and these are excellent places to fine tune and reset levels that are flowing out of plugins or plugin chains.
Monitoring in your Studio
The system will go quieter as you lower ref levels. But as a counter measure we turn up our monitoring system now in our control room. K System refers to digital reference levels as well as monitoring SPL levels in your working environment. And the relationship between them being linked.
Monitoring how loud you are listening to everything you do is super important. When people are monitoring too low they are pushing everything higher in their systems level wise and then in many parts of your DAW production you are clipping and rms levels are way too high. If you are serious about this type of work it is imperative that you have the ability to monitor at a wide range of SPL levels. I use 85 for a lot of long term general work operations but 95 and 105 are excellent for checking mixes and listening for many other things. Not to mention but everything sounds great at 105! I find having a permanent SPL meter in operation well positioned too for accuracy is a great thing to have. It keeps you very honest too! I also listen super quiet too eg 65 dB SPL through a single small mono speaker and put the mix into mono while doing this. This is also great for revealing many other things about your mix too.
It is not hard to track incoming signals and aim for a lot of 0 dB VU recording levels. There are other sources and virtual instruments are another commonly used source. Once you get savvy at using a VU meter in all the vital points in your production, you start to find that when you monitor the outputs from virtual instruments, they are all over the place in terms of a ref set VU meter. Some will blow it off the scale others will make it barely move. For example my Korg Wavestation synth plugin sounds amazing but is quite a low level. I have to add the +6 dB output boost just to get it to start moving the VU nicely. And even add some gain sometimes on a quiet patch. The volume controls on some virtual instruments need to be turned up that extra mile, often fully clockwise.
On another synth plugin, you load a patch and play it and the VU blasts over and pins the poor needle. Well it’s time to turn these types way down. Some output levels need to be almost turned off to get the VU to move nicely up to and around 0 dB VU. Over the range of your playing aim for the VU showing 0 dB most of the time.
Check the dynamics of your playing eg midi velocities. These are easy to look at, edit and change overall if needed. On some patches the velocity changes make a huge difference.
VU meters have other benefits
As soon as you start inserting dynamics processing over a source signal, the VU ballistics of the needle will change. Real VU’s, and quality ones at that, are a dream to watch ballistics wise. It is sublime how they move to the music. Very well mixed or mastered material has a very definite ballistic over the whole mix for example. When your own mixes are not as good, the ballistic looks and feels a bit wrong. Once you start doing the right things in a mix the ballistic looks right. The VU is good for showing up a rogue track that has too much something anyway, dynamics usually. Once that one track is fixed the overall stereo mix ballistic starts to change for the better usually.
VU meter plugins are doing their best to copy the ballistic of a quality VU meter. Some are better than others. Right now one of the closest is the Klanghelm meter. I have devised a series of complex tests for testing VU meter ballistics. Real and virtual. I can sit the virtual meters (at the bottom of a screen that is sitting on top of the real VU meter box) right above the real ones and compare how they both move together.
Presonus VU Meters
Presonus have kindly given away their VU meters. The range of ref levels is limited unfortunately. I am hoping they improve this over time because making the meter adjustable to any ref level is obviously the smarter way to go. Because I work at -14 as one of my ref levels a lot, the Presonus -12 setting is the closest setting to this ref level of -14. By fiddling with the sensitivity knob it is possible to get the needed FSD (Full Scale deflection ie up to 0 dB VU that is) showing mainly a normal VU swing up to and around 0 dB VU. Only falling a tiny bit short. A -18 or -20 ref level does not work so well with the Presonus meters because then they are always falling short of 0 dB VU about 4 to 6 dB which is significant. Making this meter adjustable to -18 or -20 would be an excellent addition.
What is great about the Presonus meters are their ability to open up and show themselves on tracks and everywhere which is pretty handy for some quick ref level checks. It’s great to see so many VU meters moving at once!
VU meters can be inserted on the input channels right before the signal even hits a track for recording. The Studio One level meter is also excellent for this type of work. Simply put into K System -14 or -20 mode shows them swinging nice right up to 0 dB VU. Although a bar graph, it moves rather slow which is true of an rms meter by design. As long as you are aiming for that incoming level to be hovering up to and around 0 dB VU then you will be good to go for tracking.
Don’t forget the Peak Meters
Once the input sources start getting fast and transient eg drums then get those peak meters out and just make sure they are rising to decent enough level but not hitting 0 dB FS. It only takes a slight increase in power from the player of a snare hit to slam right into 0dB FS. Keep the input gain down low as to prevent this happening. You can always lower the dynamics and raise the overall rms level of even transient sounds later in production. Once you start sending a bunch of fast attack sounds to a buss, then the rms level on that buss will rise. A normal VU meter on a drum buss will start moving nicely in fact. Toms can still be open and the kick drum too so quite high rms components in some drum sounds can still be possible.
It is not a bad idea to get some test signals flowing around your DAW either from the test tone generator or by playing back test tones on tracks. Ensure mono test signals are on mono tracks and stereo ones on stereo tracks too.
For the test tone generator in Studio One select a sinewave for say 400 Hz and the output level is adjusted for the desired level. Eg -14 or -18 or -20 etc.
Why different reference levels ?
It depends a bit on what the final outcome will be. If a client wants a very loud master then starting all your work at -14 means not too far to go master for those rms levels to come up. A decent limiter will add 4 dB of rms level over a mix without hesitation and it should do it at high quality and sound great as well. PSP Xenon does this very well. That means you are now creating a -10 dB rms master which is damn loud but still has some transients and punch. You can master higher of course up to levels such as -8 and even -6 but as well know these sound horrible. Like many I am hoping for a time where -14 and even -20 masters become the norm. Once you turn up loud a mix like that it is hard to beat.
Starting at -20 means you have further to go mastering wise if you want a very loud master at the end of the day. But the -18 and -20 ref levels have superb dynamics and the most lovely of transients because the transients are basically being left alone and not being altered. ie Pristine is a word that comes to mind. So no harm in having a super clean -18 or -20 pre mastered mix that is. I treat mastering as a separate process anyway so for me the premastered mix is important.
Some Mix Tips
Staring with all your tracks at an even ref rms level puts you in a great place prior to mixing. You will never need to add or subtract gain from a channel or track. Use the clip gain handle to fine tune your individual audio events too. It makes a big difference. Studio One shows the waveform height changes as you change the clip gain handle. This is good because you can after a while become quite skilled at matching rms levels by eye.
Groups of tracks can be sent to buss with the buss fader at unity or near 0 dB. The tracks themselves will only have to be a little lower than unity on their faders in order for a buss to end up at the ref level again. I like the channel faders reflecting the mix visually. It was also how it was done mixing on large analog consoles too. The aim was to get the final mix at the ref level. On the console VU meters.
Groups of busses can be sent to the main stereo mix buss. The buss levels then can be tweaked down slightly. I find they only need be at slightly lower then unity for the final mix to end up nicely at the ref level as well. It’s OK to fine tune buss levels too. That is what the main buss faders are for! If you have got all your horns going to horn buss then fine tune the level of that total horn sound to fit into the mix.
Keep an eye on the main stereo VU meters at all times too, You can almost mix into VU’s. Get the drum buss pumping for example and add that to the main mix but only up to say -3 or -4 dB VU. As you slowly bring in the other remaining buses which will complete the picture, you find yourself with the whole mix just peaking up real nice to 0 dB VU once again.
If you pre master 10 tracks for an album say all at the same rms ref level then you are in an ideal place for mastering all those tracks too. They will all be at the same level loudness wise overall so you won’t have to spend any time matching rms levels in mastering on your tracks. The mastering compressor is going to be pretty right for all of them!
If someone sends you a multitrack mix to mix for them the best thing you can do is choose an rms ref level and adjust each and every track in an editing program to conform to it. Then once again you will be in a perfect place to start a mix. You will amazed at how wildly the track levels will vary in a situation like this. Many have no idea how to track really well and consistently. Or how a multitrack mix should be checked and edited before it is sent out to the mix engineer. Some basic things there.
Get all this right at the source and the rest just falls into place.
I am not saying this is the only approach. It is how I like to work most of the time. I have inherited these techniques over from the good old analog days of mixer and tape operations. It worked very well then and it seems to be so adaptable into our modern DAW world now. We have got better tools than ever now for recording and we have the benefit of very low noise floors and low distortion. Not to mention the ability to put back into our mixes so much like console and tape emulation. Fun. Happy gain staging!
Jeff Evans is a composer, teacher and recording engineer living in Ballarat , Austrailia