Slowly but surely I am making steady progress on writing and recording what I hope will be a full record of the best guitar music I can currently make. This is something I've wanted to do almost since I started playing the guitar. Now I have the tools and the knowledge, only the precious commodity of time is holding me back. Still, with some strategically stolen late nights and weekends I finished another track, this one titled That Little Yellow Car.
Instrumental music's non-lyrical nature doesn't lend itself well to obvious song names. There is no singable chorus that you can lift a line from. So when I come up with a song and get the first snippet or demo down, I usually name it based on the first phrase that pops into my head. Originally this song had the name Like A Walk In The Park, but as it evolved I wanted to write this song more as an ode to my car.
Yes, when I read this back it sounds kind of lame and self-absorbed (though that accusation is often leveled all guitar instrumental music, isn't it?) Not to get too deep into it, but that little yellow car has become an important powerful symbol to me. So I thought, hey, it deserves a song named after it.
The Song-Writing Process
These days I capture most of my ideas using Apple's Music Memos. When I'm in idea-capture mode I'm not trying to compose anything in the sense of editing at this stage. My only goal in these moments is to just capture whatever idea pops into my head or I've stumbled upon while noodling around (always keep your phone handy).
This tune came together from two separate ideas I'd captured. The first was the chorus melody (it's the singable part of the tune). It was one of those moments where the chords and melody just materialized in my head all at once. When that happens I'm usually clumsily scrambling around trying to find my phone so I can record it before I forget it. Fortunately I caught this one before I forgot it.
The second part was the little E-F# chord vamp. It's an old Alex Lifeson trick of taking your basic E cowboy chord and moving the four lowest notes up a whole-step, leaving the B and E strings ringing. But instead of bashing the chords out à la Farewell To Kings, I invented this somewhat complicated picking pattern for it. It probably sounds like it's played as a simple up-and-down arpeggio in a sweep-picking style, but I actually alternate-pick (or "cross-pick" as the bluegrass dudes might say) it. I just found I liked the sound and the feel a little better.
As for the rhythm of the track, it was going to be a shuffle from the very start. Somewhere deep in my DNA, my ancestors must have passed on to me a deep, structural love of all things triplet. They consistently show up in my playing and writing in many, many forms. But a shuffle can be a tricky thing to pull off convincingly as it can sound kind of jokey and dated. I knew when it came time to mix this sucker that I needed some earnest ass-kicking behind that tried-and-true shuffle beat.
So here I had these two basic parts. It was a good start. Hell, lots of successful songs have been written with two parts (verse, chorus) that just alternate. But somewhere I felt the spirit of Guthrie Govan lurking over my shoulder telling me, "you should add some more fiddly bits to this one". So after a two verse-chorus pairs it was time to take the song in a totally different direction.
For that, it was time to head into the "noodle zone", which is basically sitting in my studio/office and experimenting…a lot. Eventually I started playing around with some really thick chord voicings that I like with distorted guitar (power chords are so troglodyte), and came up with this:
I feel like there's some deep territory still to be mined in the area of interesting chord-voicings that work with a distorted signal. Power chords are boring but the thickness of distortion makes busting out five- or six-voice jazz chord a strain for the listener. But I think guys like Eric Johnson find some really cool voicings that work well with a blazing Marshall. Anyway…
Once I got on that kick I came up with three different variations. Normally I would play a sort of Survivor Island style game and vote all but one off the island. But I really liked all of these ideas so I just strung them out one after another.
This was good. The song was headed in a new direction. It had a good energy, but I needed some kind of resolution. So again it was back to the laboratory to see if I could drag the song in another direction. This time I consciously started mining riff-territory as a counterpoint to the chordal stuff in the earlier section. I'll be honest, I set out to come up with the most muso-nerd thing I could. It had to be angular, it had to feel sort odd-time-signature-y (even if it wasn't) and by God it had to be in E. So I came up with this:
At this point I still wasn't quite sure how I was going to end it, but I trusted the great gods of guitar song-writing and laid the tracks down (after practicing that damn riff for about two days). Then it came time to put a bass-part down. I knew I couldn't play the riff on the bass (Billy Sheehan I am not), so I needed some kind of counter-point. Somewhere in various experiments my inner-John Taylor surfaced and that little chromatic octave-pattern came out of nowhere.
Logically, that little disco walk shouldn't really work over the muso-nerd riff. But somehow it just does. I don't know why, but once I stumbled on it I knew I had a way to end this thing, which was to build that disco walk up to a frenzy with a big punch in the face at the end.
The final bits were the echoey Strat chords (which function almost like keyboards) and one last Big Guitar Solo. Boom. Song finished. Now let's record the thing.
The Recording Process
As I work out the song-structure in my head, I move things around in Logic to match. Once I've got a good idea of how things will lay out, I start adding section markers and copy-and-pasting parts so I can hear the backing tracks as one whole unit. This lets me think in terms of larger "song components" where I can just move whole sections around without having to record each and every little part. Eventually, most of those looped and copied parts get replaced by new tracks that run through the entire length of the song as I'm always wary of things sounding too chopped-up.
Since I work in the genre of guitar-nerd music, I have the great fortune of declaring great swathes of musical territory as "guitar solos". These are where I get to stretch out beyond whatever basic melody I've cooked up. For these I would loop four bars at a time and start piecing together a lead part. Where possible, I'd try to sing something to myself first so that whatever I came up with had more of a lyrical bent. But in some cases the fingers just fall into place and magic happens.
This is, admittedly, dangerous territory. Working in little four-bar chunks it's very easy to lose sight of the forest in the midst of all these wonderful trees, and there were many occasions where phrases that made sense in that looped four-bar world didn't fit in with what came before and after it. So while I was trying to nail the execution of these ideas, occasionally I had to throw them away and invent something else to fit better into the song. At times it feels a little two-steps-forward-one-step-back, but it's a workflow that gets me to a happy place in the end.
All of the guitars (as well as the bass) were recorded through the Kemper Profiling Amp. The clean tones were recorded with a Fender Telecaster with very heavy compression. The remaining tracks were a combination of my EVH "Bumblebee" and Ibanez Jem. The wah-wah was all done in the Kemper using its model of the Dunlop Cry-Baby.
Because this song had so many guitar parts and because I was often working on the same thing over multiple days, I had to be very disciplined about writing down my Kemper and guitar settings as I went. I made heavy use of Logic's Track Notes feature and always wrote down both the guitar settings as well as Kemper settings. If I customized anything in the Kemper I made sure it was saved in a place I could always get back to.
Another useful trick in Logic is designating custom icons for tracks. Logic as long had the feature of selecting from a list of stock icons (Les Paul, Strat, P-Bass, etc.). But a recent release finally added the ability to choose custom icons. When that feature came out I hit Google Images to find the best images I could of all of the guitars I own. So if I record a track with the Jem, I make sure I choose that icon for it. You won't believe how much of a difference this makes when trying to find a specific track in a sea of them
In general, I try to capture two tracks for each guitar part: one from the Kemper and a pure DI signal I can use if I ever want to re-amp anything. For the re-amp track, I plug my guitar into a Radial J48 DI box which splits the signal. One output goes to the Kemper and the other goes straight into my audio interface. Then I take the outputs from the Kemper and route those to separate inputs on my audio interface.
One of Logic's most impressive features is Drummer Tracks. I still can't believe how well these actually work in practice. You pick a drummer persona, and a basic framework for a beat and then start customizing. I've spent a lot of time lost in the minutiae of drum-programming because I have pretty strong opinions of how I want drum parts to be played. But with Logic's Drummer Tracks I learned to let go and trust the "drummer" and as a result I can get on with the creative process much more quickly.
The interface is much like a conversation you'd have with a real drummer. Can you swing this a bit more? Make this part a little softer. No, I want more toms there. I'd say 90% of the time Drummer either played exactly what I wanted, or at least got in the ballpark.
On those occasions where it doesn't quite do what I want and no amount of jiggling the Drummer controls gives me what I want, I drag the drummer section into an empty MIDI drum track and do a little selective surgery as needed.
One of the challenges I had with this track was getting the tempo and time signature setup correctly so that Drummer would do what I wanted. The basic beat of the song is a shuffle, which is typically thought of in 12/8 time. Since the basic unit-of-time in 12/8 is an eighth-note, the tempo I selected for the song that was pretty fast, but that's because we're quickly counting twelve beats per measure. I could have set the time signature to 4/4 and set the song tempo to a third of what I specified, but I found it hard to make Drummer swing with the triplet feel I was looking for.
However, the problem with a fast tempo is that sometimes Drummer would concoct some really inhuman, coked-out drum fills that sounded like Animal on a bender. For these I either twiddled the Drummer knobs until I got what I wanted, or applied the aforementioned MIDI-surgery.
I have always had a deep love for reverb, but I know I've over-used it in the past. I've always loved the lushness it adds but when mixing with several instruments, reverb is really more about psycho-acoustics and how the brain perceives "depth" in a recording than it is about tone. When everything starts getting reverb, it's a lot of sonic information stacked on top of itself and the result is usually a blurry, incoherent mess.
So these days I really try to restrain myself from using reverb. When I record my lead parts, I add a little reverb to the monitoring path (it doesn't get recorded) just as a sort of comfort-blanket. But I record all of the parts to-disk bone dry.
For the opening guitars I found that I really started to like the dry, up-close sound and left it there. When I did add reverb to guitar tracks, I setup a specific auxiliary channel with a nice lush plate reverb plugin (oh I love me some plate reverb…) and would just bus a little bit of whichever lead part I wanted onto that reverb channel.
By having a single reverb unit (instead of one per channel) I get the benefits of cutting down on CPU load as well as having a more cohesive reverb sound. That one really good reverb plugin makes those sounds gel in a way that separate reverbs just don't—at least to my ears.
On the lead tracks I'll put a little delay just to widen them up a bit and make them sound bigger. Because the lead tracks have reverb and the rhythm tracks don't, there's a tendency for the lead sounds to feel as if they're "behind" the rhythm tracks. A few volume tweaks and some delay brought the lead tracks back to the forefront.
As for EQ, I typically put a high-pass filter on all guitar tracks to cut out everything below 200 Hz. There just isn't much useful sonic information for a guitar in that range and it's an area that's very easy to build up too much, resulting in a very "woofy" sounding track.
I try to do as little post-processing on the bass as I can. Like the guitars, it's recorded bone-dry and usually stays that way. I never add reverb or delay to a bass unless it's for a special-effect of some sort. However bass does tend to get a more liberal dose of EQ and, especially, compression.
If you boil mixing down to its essence, it's really all about setting volume levels. Or at least, that's the primary focus. Even EQ and compression are, from a certain perspective, about volume levels. EQs shape volume in the dimension of frequencies, while compression shapes volume in the time-dimension.
Since I really want the drums and bass to "sit" together as a cohesive unit, when I first start the official mixing process, I just bring those tracks up and get a solid foundation laid. As I tweak one, I typically have to balance the other. So there's a little bit of back-and-forth when getting this setup. Once I start bringing the other tracks in, I typically have to fiddle with these a little more before I'm happy. Rare is the mix that doesn't involve looping back and adjusting something a few times.
I've been on a parallel-compression trick with drums lately. Also known as the "New York" method, the idea is to take a copy of your drum tracks (or just a bus send) to a separate auxiliary with an incredibly squashed compressor setting. Ratios of 12:1 are not unheard of with this method. While your normal drum tracks are playing, just start to fade in a little of the hyper-compressed version until you get the crack of the snare and click of the kick drum you want. It doesn't take very much of the hyper-compressed signal to make a difference, but oh what a difference.
Once I had the basic "static" mix setup, I had a few parts that stuck out in odd places. The mix was also a single energy-level all the way through, so the final pass was to make a handful of small adjustments using volume automation to get the parts to sit right as the listener progressed through the song.
I also used a few panning tricks to get multiple parts to play together better. For example at the end of the second intro (around the 1:27 mark) the rhythm section drops out and there's just the Tele rhythm guitar and the lead. I wanted to keep both parts, but sonically they occupied similar space. If I couldn't separate them in the frequency spectrum, I had to choose some other dimension to split them across. So I did a small panning automation to briefly push one part to the left and the other to the right, just to separate them.
If you listen to this part with headphones the effect is more pronounced, but in a pseudo-stereo environment in which we don't always get clear left-right separation, the net-effect is better sonic separation.
That's a Wrap
Every song feels like a step forward in writing, recording and mixing. Like most of the stuff I write, this comes from parts that have been rolling around in my head for some time. I still have song ideas that are over a decade old that still haven't made it out of my head. But song by song I slowly get this catalog of ideas out into the real world. Whether anyone listens to them or not is really beside the point for me. I just want to make the music that's in my head come out of my speakers so I can hear it in the real world.