Approaches to Better Gain Staging Need Help

I usually start with putting the free PSP Vintage Meter on the 2-bus. I then solo the kick and set it to about -10 VU and then solo the bass and set it to about -10 VU as well, this often yields a good starting point for me which only require a little tweaking. After that I raise the snare and then other drum elements followed by percussion to set the foundation for the mix.

After that I sometimes raise the vox to put them so they sound good in comparison or the lead/most important instrument. After that I usually work my way from the lower register instruments up to the brighter ones. Also remember that if everything is mixed loud the result will be a flat mix where nothing is loud instead since there is no relative loudness between the different tracks. Some things have to be more on the quiet side for others to be percieved as loud and you can think of the mix as a roller coaster; if you only go up to the top and stay there with no slopes, valleys or hills it won't be a very interesting ride now will it? (For a minimalistic mix all tracks can be loud if there is enough silence to act as valleys and let the transients smash through to grab attention.)

This way there is often pretty much headroom left for final adjustments when everything is roughly in order and works well for me.

I usually do this both with a more or less unprocessed mix to see where I'm at (and bounce that out for reference) and then again when I'm doing the final mix and ad comp, EQ, FX etc.

ShadDOH


I find this works well?

It indeed does and is very important as well. I you for some reason find yourself pushing the faders constantly and getting clipping, or close to clipping, try boosting the volume knob on you speakers so you are listening at a greater loudness to start with. In that case maybe you won't feel the urge to up the faders in your DAW as much as it could cause you a headache.

Oh, one important notice as well is that cranking the gain on you speaker controller works well even if it doesn't go to eleven even if popular belief claims otherwise (but if you can afford one that goes to eleven you should off course hurry out and buy it immedeatly ).

This helps a lot.
I guess I was needing to back down my peak bus level down to the -10 in order for everything else to sit on top of the mix without driving the bus into the red.

I am sorry this post is so long but I have been meaning to post some information about reference levels in the digital system for some time now and this thread seems to be a great opportunity.

I believe in order to set up any recording system correctly you need to think about reference levels. Many people go wrong here and don’t have any reference level system.

Back in the old analog days we had reference levels of 0db VU and this worked well. Our input signal had enough gain added by the mic pre so 0db flowed down the channel. Then we controlled the level with our channel faders to get a group level of all the tracks onto our main mix buss of 0db as well. We had VU meters to tell us when things were correct.

You can apply the same approach to digital except that as 0db FS is our loudest signal we have to think about new ref levels now that are well below that.

Any signal has two components to it. The RMS value of the signal or the meaty part of the wave if you like and the peak level of the transients. Back in the analog days we concerned ourselves with the RMS part of the signal and the transients took care of themselves to a certain extent. But in digital now we need to be aware of both parts of the signals. A synth pad might have very similar peak and RMS levels but a snare drum hit is a completely different story. Know the difference!

Peak level metering is great for the transient parts and VU meters (real ones) are the best way to monitor average levels.

So where do we put our reference levels? Minus what? Well Bob Katz explains this very well in his book on mastering and I agree with it totally. There are three ref levels that one can apply in digital. The first is -20 db and this is good for very high fidelity recordings where you want maximum transient performance. 24 bit is better in this mode although 16 bit can do it as well. There is still 70 db of range of signal below -20 db even in a 16 bit world and that is still a lot. But I like 24 bit in this mode. The level of -20db refers to the RMS part of the signal and that means we have got a full 20 db of headroom to handle a transient if it is required. The next level up from that is -14db This is great because it is 6 db louder than 20 db but still allows for a decent transient. The third ref level is -12 db. This is 8db louder now than our -20db level but still allows for a good level and still maintain a reasonable transient.

You need to get a program like Adobe Audition to make three reference level recordings for alignment. A 1000Hz sine wave at -20, -14 and -12 db is a good start. Stereo ref tracks are fine. Now you can load in a ref level track wave. You need to decide now what ref level you are going to work at!

VU Meters and why we need them!  Now I believe you need real VU meters on the stereo mix buss of Sonar. I don’t know how people manage without them, I can’t. Once you have them and they are set up properly you will never go back! (not only are they fantastic for levels but they can also tell you a lot about a sound eg if it has too much bass in it for example. The VU's will go mad a bit. Also a great mix will always make the VU's dance just the right way.) If we are working at a reference level eg -14db then the Sonar track and buss meters although they will be working fine are hard to read as they are moving around a low level most of the time. Best to leave Sonar metering in Peak mode as well not VU by the way. Use a real VU meter to show VU. Pretend VU plugins are crap too. They bear no relationship to a real VU meter! While we may be working at say -14db we need to see something that is showing 0db eg a VU meter. Seeing a VU move up to a healthy 0db VU gives us a sense that things are good (in a psychological sense too) and levels are right even though on the track and busses this is all happening at -14 db instead! They stop you from recording too high.

Let’s say -14db which is a good compromise between level and transients. But we now have another issue here. The VU’s are set to read 0 db VU while the track (and the main output of Sonar) is playing at the ref level. The VU’s need to be switchable because of you work at -20 db and have your VU’s showing 0db then you decide to work at -12 db then your VU’s are going to read 8 db higher and pin over and that is no good so the VU’s also need to be re calibrated every time you change your ref level. Once all this is setup tracking, mixing and mastering is all much easier.

Tracking. Say we are working at -14db. Simply set the incoming signal to the audio interface (and that involves its gain stages obviously) so that the VU reads around 0db. Now you know that you are averaging around -14 and you have got plenty of headroom for a transient. Recording as loud as you can in 16 bit mode is not cool and just plain wrong. What ref level are you talking about? Loud as you can means the slightest transient will blast through the 0db ceiling and that is not good.

Mixing If you have got a bunch of tracks all working around -14 db then when you are doing your mix, firstly the track faders will all be up around 0db or just a bit lower to allow for the cumulative effect of the signal getting louder on the main busses. Your main buss faders will also be around 0db but the average signal level will still be around -14db. The whole mix now will peak the VU’s to around 0db VU as well. This means you end up with a mix that is also RMS averaging around -14 db and still plenty of transients. Same as the tracks.

Mastering Well this is another story and obviously the plan is to now get the average or RMS levels much higher to compete with commercial CD’s and there are various stages that are employed to do it and that is a complete topic for another discussion but the same principal applies here too. As you move the RMS levels up then the available room for transients decreases. I tend to push things up to about -6db (RMS) in mastering situations but it does depend on the music etc and the client. This also means altering the VU response to allow for it. Eg I drop my VU’s down another 6dv from -12 in order to see what commercial levels are like and working with your own mastering levels.

You can work at different levels as well within a project eg Tracking at -20db ref level but mixing at around -14db with some mix buss compression to just bring this up to -12db. But the only thing with this is that you have to alter your VU sensitivity as you do these various phases of a production. Not a bid deal though.

Another thing to note is Calibrated Monitor Gain and this is the concept of setting the levels in your room while you are working with these three ref levels. Band limited pink noise is a good test (eg 500 to 2Khz) and you also need three test signals (-20, -14, -12) of those as well. With those playing you can adjust you monitors to 83 db SPL (individually too) CJ once said in other thread there is no ref level with monitoring and anybody can be monitoring at any level and that is true to a point but there are also very precise reference levels for recording and monitoring. Start thinking about calibrating your system!!! Many people don’t and they have no reference levels for recording, mixing and mastering and monitoring and they wonder why they have problems getting it all together.

I have found that once you start adhering to the things I have mentioned above your productions will just sound better, track better, mix and master better. I never bring on any clip lights anywhere, I never overload any plugins and even after mastering I get a big loud fat sound (with plenty of transients still present) still with no distortion or peak lights coming on anywhere! Everything is better! Digital is a bit tricky becaause we have to think about everything being below 0db FS in all stages of our production. But once you get around this concept of exact ref levels then things just start falling into place.