A Guide to Louder Mixes
One of the first surprises for home-recording musicians is how quiet their mixes sound compared to professional CDs. In the last 15 years, the major studios have gotten into an arms race of sorts when it comes to how “hot” they record CDs. Listen to a CD recorded in the early ’90s and one recorded this year, and there will be a huge difference – the new one will be heavily compressed and limited (“squashed”) and devoid of almost all dynamics.
There are many reasons given for why studios churn out over-compressed CDs: the average listener thinks it sounds better; it plays better on the radio; their CD must be louder than the rival studio’s CD, and so on. One thing is for sure though – a squashed, flat track ensures that the guy driving his stock 1978 Chevy truck with one speaker blown and the other praying for deliverance has the same experience as the guy driving his shiny new Acura TL with its 8-speaker, 5.1 DVD-Audio surround sound.
For the home studio mortal there are a lot of techniques for making a hotter mix, but keep in mind that in the loudness war big-label mastering engineers allow as much digital clipping as they think they can sneak by the casual listener. The following methods will ignore that practice. Also, while mixing with these techniques, do not solo any tracks. Listen to the whole mix when adjusting EQ and volume. If it sounds good soloed, you probably have a lot of work to do. Make sure to use RMS and peak meters on your output bus. RMS is a measure of apparent loudness and is a quick way to see whether things are headed in the right direction, and as a mortal, you need to make sure your peak volume never goes above 0 dB.
The first thing to do is hunt down bass frequencies and remove them. Bass frequencies rob the mix of headroom – they are mostly felt rather than heard, and the brain will “fill in” the missing frequencies anyway. If someone wants deafening, window-shattering bass, they will have two pairs of 12-inch subs in the trunk for that purpose. So cut everything below 40 Hz with a high-pass filter, no matter what. On everything but the bass guitar and kick drum, you can start the high-pass around 150 Hz and go up until things start to sound thin. This cleans up the low-end frequencies for the bass and kick and results in a “punchier” sound that’s easier on people’s stereos (and their ears).
Your next task is to hunt down the fundamental and resonant frequencies of each instrument in the mix. This depends on the instrument (and to some extent the notes that are played), but the fundamental frequencies can be discovered using a frequency analysis plug-in, and the resonant frequencies are usually a multiple of the fundamental where the multiple is a power of 2. For instance, a guitar’s fundamental frequency is about 100 Hz, and its resonant frequency is 800 Hz (100 x 2 = 200; 200 x 2 = 400; 400 x 2 = 800, so 100 Hz x 8 = 800 Hz). The same frequencies apply to male vocals. A peak cut at both those frequencies with a Q of 1.0-1.4 will clean things up nicely in the mix. When in doubt, the resonant frequency is usually the one that makes an instrument sound “cheap” (or recorded in a concrete bomb shelter).
How much you cut the lows, fundamentals and resonant frequencies should depend on how the track sounds in the mix – once it starts getting too “thin”-sounding, back up a notch and leave it until your next mixing session. If it still sounds thin, then back off the cut a little more, and so on. If you have some solo work that sounds too thin, you can always clone that track and run a less aggressive EQ on it. Cross-fade the two tracks during the solo, and have the best of both worlds.
While all instruments share frequencies, your brain can only hear one at a time. If your tracks are competing with each other (turning up one track drowns out another), use complimentary EQ to fix it. This means you have to find out which track is “masking” the other and at what frequency. Listen hard for whatever it is about the masked track that you stop hearing when the other track is turned up. When you know what you are looking for, solo the masked track and sweep around with a narrow peak EQ. Start with a Q around 2 or 3 and a cut of 10 dB just so it’s obvious when you hear that sound disappearing. Once you’ve found the masked frequency, cut it on the offending track to taste, while listening to the whole mix.
Use highpass filters on tracks that don’t need higher frequencies, usually guitars (electrics especially, acoustic guitars less so), kick drums, bass guitars, toms and even vocals. Set the high-pass at 15 kHz or so and work your way down until it starts affecting the mix. Another dose of reality – you probably don’t need much above 15 kHz unless your target audience includes dogs. You probably can’t hear them anyway, and too many highs can lead to harshness and high-end distortion.
Once you’ve got some basic equalization taken care of, it’s time to check out dynamics. As a rule each raw track should be consistent in volume throughout the song. The section where the vocalist is breathy should be approximately the same volume as where they are screaming their lungs out. The brain thinks the breathy section is quieter, but it’s really just different frequencies and dynamics (lower RMS). Your sequencer software should have a “clip gain” automation envelope that you can apply to the individual clips. (They may also be called pre-fader envelopes.) Use these to smooth out tracks before you go for a compressor plug-in. The end result should be that each note is the same volume, but the articulation doesn’t suffer from being over-compressed.
You will need to use compression to keep the kick and snare in line. Use a slow attack (~10ms) to keep the punch, and watch your release times – too long and you’ll cut the next hit. Bass guitar compression is pretty much a given, but you can be more liberal with it. Acoustic guitars need compression, but remember when strummed they are very percussive and need a generous attack time (several milliseconds) and a careful release time to avoid cutting down the next beat.
If there’s something in particular you want to bring out, like a solo, turn down the other non-bass/drums tracks a little bit and turn up the solo track a little bit. For example, once the vocalist starts singing, you can turn the instrument bed (but not the bass guitar or drums!) 1 or 2 dB. No one will notice, and you’ve already played that killer guitar riff 8 times, so it’s in people’s heads. They don’t really need to hear it again until the next verse.
To check your progress, put a single band compressor on the master bus. Turn up your tracks (or busses if you use them) and listen for rude instruments. The kick drum, bass guitar and chugging electric guitars are likely to be culprits here. You should also listen to the mix in a car, if at all possible. Car stereos will bring out muddy bass, clangy midrange and piercing highs in a hurry. If the RMS level is high enough and you still like the sound, then good for you. If not, start at the beginning and cut those frequencies some more. Only boost a frequency if there’s something specific you are trying to bring out, and a little bit goes a long way.
At this point you need to decide whether to master the mix yourself or not. My advice on this is if the target media is MP3/WMA, then you should master the track yourself. If the target media is a professional CD, then get a professional to master it. (Hint: if you have to ask, it’s not you.) For a demo CD, it’s a toss up. A professional master is going to sound better to someone going through stacks of demo CDs, but it will cost you, and the mix quality will make or break the track anyway.
If you are going to “master” the mix yourself, there’s no better place than right in your sequencer. A multi-band compressor is a good start, and your frequency analyzer should go right in front of it so you can see what frequencies might be getting out of control (probably the low ones). At this point you should only concern yourself with RMS vs. an over-compressed sound and the general color of the song.
If you send your mix out to be mastered, be aware that they can’t do much with a mix that’s too hot. You also need to get rid of that compressor on the master bus. If the mastering guy doesn’t tell you exactly what he wants volume-wise, you should make sure your peaks are no more than -4 dB and your RMS should probably go no higher than -14 dB.
