I like to experiment with music, and I feel like breaking the rules today.  I’m going to not only give you an alternate, instrumental version…but I’m also going to let you hear the individual performances as they were recorded for this remix in the spring of 2009.  The piano was performed by me, and the violin was hired out.

Enjoy this look beneath the surface!

Ableton Live 8

What you’re watching is four hours worth of composition, stemming, mixing, “bussing”, and mastering reduced to three minutes.  You’re listening to the final result while you watch it being created!  I thought that would be cool.  It’s probably really stupid.  (The video is at the bottom of this posting, or you can skip right to YouTube to watch it in your own domain.)

In the first few seconds of the time lapse, you will see me working in Ableton Live 8.  I am “stemming” (bouncing/rendering) the parts to 32-bit WAV files.  For example, I am separating the piano part from the guitar part and so on.

Adobe Audition 3

The rest of the video is spent in Adobe Audition 3, my baby.  Some musicians and audiophiles are probably repulsed that I would use Audition for all my mixing and mastering.  I have a few words to say about that, as it turns out.  Most importantly, I think that there is no “best” multi-track suite or DAW software.  Rather, I think how you use the tools are your disposal is what matters.  I could never turn around the same quality product in Pro Tools, for example, because I have little to no experience using it.  On the other hand, I do have an extensive background in radio, where Audition tends to dominate the landscape.

I’ve easily spent tens of thousands of hours of my life in front of Audition.  It’s at that point that you get into an interesting zone where you’re able to “think” in Audition, almost as if it’s a language.  Setting levels, EQ, mixing, etc…they all become second nature.  I don’t know the psychology behind it or if it’s just muscle memory, but when you tear down the need to think about the menial details of production, you become a fluid producer.  What’s more, I’ve stuck with Audition simply because it allows me to do 90% of what I require as a musician, remixer, and mastering engineer without needing third party plug-ins.  That is to say…out of the box, Audition is up to the task of almost everything I need to do.   (This is true of any competent DAW or multi-track editing software.)

As you read on about my work flow or process, you’ll realize that I have very little need on a day-to-day basis to supplement Audition with third party plug-ins.  Even when I do require a plug-in, I’m pretty religious about the ones I will use, and it’s a short list.  It’s a combination I’ve become so comfortable with.  A comfort level to the point where I’m free to be artistic within Audition instead of procedural.  It’s able to keep up with me, it’s flexible enough to let me change my mind all the time, and the details of its inner-workings don’t get in my way as a creator.

This is probably how YOU feel about your DAW, be it Pro Tools or whatever.  My only point is that it’s not the tool, but how you use it.  Finding that synergy is absolutely critical to being able to “create” rather than “screw around with settings”.  Long story short, Audition is that tool for me.  (Rant complete.)

You’ll see my process for setting up “first draft” masters of songs.  First, I begin to tinker and arrange the stems and parts and samples to get a feel for how I want things laid out.  You see me change my mind a few times near the beginning.  Then, once the overall layout and flow of the song takes shape and is to my satisfaction, I begin to work on the sound of each specific track or instrument.

C4 Multiband Compressor

It’s at this point that you will see me using many different tools.  I rely heavily on the built-in plug-ins provided in Audition (reverb, channel EQ, etc) and also the Waves plug-ins.  Specifically, you will notice that I fall back on the C4 compressor on almost every track.

Once I’m happy with the sound of each instrument or track, I get a little more serious about the overall mix itself.  I will create a bus for each logical group to simplify the mixing/mastering part.  In audio terminology, a bus is just a group.  For example, instead of having all the percussion and drum tracks go straight to the “Master” channel, I can route them to a bus called “Drums”.  (These show up as yellow faders on screen.)

This lets me make changes to all the constituent tracks at once.  This is handy for two reasons.  First, if you’re going to apply the same effects to all your drum tracks, you can do it in one place (the bus) versus each individual track.  This saves you a TON of processing power (and time).  Second, it helps with the final stage of the process…mastering.

Towards the end, once the bus for each group has been set up and the mix is “okay”, you will see me open the automation lanes for each bus.  This allows me to basically “draw” the volume and set keyframes over time.  As an example, the loudness of the piano and guitar varies depending on the density of the mix and what else is going on at the time.  In this case, I fade them down gradually whenever the cello and violin solos kick in.  When your tracks are properly “bussed”, it makes mastering much faster and pleasant.

L2 UltraMaximizer

The audiophiles and other musicians out there are wondering what I’m using on the master channel for the “final mastering”.  The answer is the L2 from Waves.  About 99% of my tracks are mastered with the L2.  The other 1% are usually the L3 MultiMaximizer.  But, I usually prefer the “loud and proud” tendencies of the L2.  (Waves products and plug-ins have a reputation of being “pricey”.  While this is true to some extent, I’ve personally never lamented a purchase.  The proof is in the sound for me.  What’s more, an investment in the best tools can never be a mistake.)

When you combine Audition with each bus with the L2, it makes mastering almost magical and fun.  You negate the need to worry about tedious details and allow yourself time to experiment with the mix of your track…the best favor you can do for yourself as a musician or engineer.

Enough jargon!  On to the fun part…the video!  Be sure to watch in HD if you’ve got broadband!

I’ve also included the audio on its own, in case you’d like listen without video or to keep a personal copy (which I’m fine with).  Just remember the copyright warning about using it for anything else. 

A lot of work went into this one.  Specifically, making sure the saxophone, guitars, and other instruments have a perfectly-crafted vintage sound.  There is also convolution reverb in play to model a jazz club.  A lot under the hood, and a new vocalist to boot!

I mastered in Adobe Audition 3 as usual, and used the Waves plug-ins to craft my vintage sounds.  The final master is the ever-reliable L2.

Mellow out, and enjoy.  It was a pleasure to deviate and produce some downtempo jazz.

Okay, so.  Let’s talk about vocals for music.  And a little about vocals in general.  You radio production people (like I am) probably need to hear this.  If you’re one of these people that “hard limits” or otherwise drastically compresses vocals, please stop.  Otherwise, I’m going to have to start hunting you down and showing up at your studio to stop you.

People, I don’t have that kind of time.

When you hard limit or overly compress any audio, you ruin any potential for its dynamic range.  This is especially true of vocals.  A giant “square” peaking at -1 dBu isn’t going to be very expressive.  And it’s going to really piss off anyone who masters your audio.  What’s more, realize that your material — be it radio production or music — is going to be mastered again in the audio chain.  Mastering often purposefully removes dynamic range to maximize the loudness.  Anyone who’s listened to the radio knows what I’m talking about.  TV and CD’s, too.  Realize that when you over do it, your audio’s going to sound like hammered awfulness on the other side.

The key to compressing and processing vocals is in the subtlety.  If they’re not loud enough, bring down everything else in the mix.  Don’t just pump up the vocals.  I realize clipping in the 32-bit real is impossible, but still…dynamic range is your friend.

On the topic of loudness, let’s talk about standing waves.  Ever been somewhere and noticed that the audio had a ground hum?  That’s a standing wave that’s easy to detect, but what about when it’s in your audio you don’t even know it?

I can’t tell you how many times I’ve recorded vocals only to realize later than there’s a sub-audible hum in the audio.  Probably from the damn air conditioner or some outside noise.  Since the human ear can’t usefully determine loudness or direction of extremely low sounds, you may not even notice it’s there.  Similarly, you may not realize how loud it is, and how much of your sonic palette it’s stealing.

If there’s “junk” in your audio, it’s going to hold back how loud you can make the mix…and how clean it’s going to be.  It’s not uncommon to eliminate unnecessary noise, hums, and standing waves and be able to make that element 6 to 9 dBu louder overall.

Not only is this a good habit for details work like garbage, but it’s also a useful philosophy for mixing audio in general.  Does EVERY element in your mix need the FULL sonic range of 20 to 20,000?  Does that bass drum need to be using the high end?  Do your vocals need to be dipping down into the 50 hertz range and messing with your bass line?  Does the violin need to be hitting the subwoofer?

No, no, and no.

Get in the habit of notching, shelving, or equalizing your audio to account for this.  Budget the frequencies in your mix more.

There are number of ways to go about this.  The quickest way to get rid of standing waves in your audio is to use a scientific filter.  You want a “high pass” filter, which means that you designate a frequency (120 hertz in the example to the right), and everything AT or BELOW that threshold is increasingly more quiet as the frequency decreases.  You can change the slope of the shelf by adjusting the “order”.  The higher the order, the more steep the shelf.

For clean vocals in music, this is the exact preset I use.  Unless you’re using a booming voice guy like Lonnie Perkins, Chris Corley, or Don Lafontaine, shelf the crap out of vocals.

If you’re a little more technically savvy, you can play with the parametric equalizer in your DAW.  I’ve displayed a screen shot of Adobe Audition 2.0′s version to the left.  With more time and tweaking, you can make parametric equalization sound much better than a simple shelf.

Retreating into basics is common for beginning audio editors and engineers.  And when I say “basics” I mean “graphic equalizer”.  It’s a lot less cool, but you can still create an effective vocal shelf (shown at right).  On a long term basis, you’re much better off taking the time and learning to use the parametric EQ function in your DAW.  It’s much more precise, and it also forces you to think about your audio in frequency space terms.

If you take anything away from this…let it be to stop over-doing it with vocals.  Let the radio processing or your mastering engineer take care of that.  Preserve as much dynamic range and color in the original as you possibly can.