19/02/2015

klystrack tutorial - melodies and textures

Alright, this time I'm going to write something that's actually fun to write about. Instead of the super technical bullshit of describing precise settings and whatnot, I'm just going to ramble on for a while about various tips, tricks and techniques I've picked up over the years.

making subtle from obvious
As contradictory as it may first appear, making chiptunes in a tracker is all about being subtle. Even if the aesthetics of chip music make it sound "raw" and "simple", the really great songs out there are all great because of the detail work, the fluff that often goes unnoticed until it's removed from the music. Subtlety becomes that much more important when your basic tools and elements are clunky oldschool digital crap. When your palette is so limited, subtle differences become more important. If you have 500 oscillator shapes to choose from, you're going to put a 50/50 Pulse in the same category as a 25/75 Pulse. But when your selection of shapes is cut down to less than 10, then a 50/50 Pulse is one thing, and a 25/75 is something entirely different. Hell, they're two different instruments. They are as different from each other as a guitar is from a piano, for your needs.

In Klystrack, the Pulse wave as 128 discreet increments between 0/100 and 50/50. That means 127 different possible textures that are all different. Sure a Pulse of 7FF is really really similar to a Pulse of 7EF, but what I mean is... there's 128 fucking possibilities, not 3. So even though a Pulse wave is something very very simple, you can get a LOT of mileage out of it if you train your ear for it.

more than the sum of the parts
Something you learn the hard way, over many years, at least in my case, I'd assume, is that the overuse, and general abuse, of commas in a sentence, will piss off most readers. One other thing you learn over many a long year spent making music is that it doesn't fucking matter if your bass/guitar/whatever riff sounds like the universe is exploding in an orgasmic firework of tits and beer. If it turns to mush when you add in the other instruments, or if its punchy character gets lost in the mix, all your hard work to get that one specific instrument sound like the bomb will have been in vain. 

Have you ever listened to an acapella? Or had the chance to listen to a professionally recorded song track by track? If so, then you know how ugly everything sounds when any one track is soloed. One classic example is the guitar solo in Pink Floyd's Time, off of DSotM. This is one of the cleanest ever prod ever on any album ever, ever. But if you watch the making-of documentary, you'll get a part where they solo that solo (hah) and you hear how fucking dirty and trashy the guitar sound is. I mean it's all kinds of distorted with handling noise and fret buzz and families of rats eating at the pickups while the guitar is repeatedly drilled through, underwater, on fire, in New Jersey. Sounds like a retarded three year old (with an extreme talent for melody, and the capacity to lift a heavy mid-60s stratocaster) was playing. Yet when the rest of the song is put back on, the whole thing just kicks your balls in.

Many times you'll need something like...I don't know, strings. You want some string chords, or melody or whatever. You fire up Klystrack and realize that after two weeks of hard work without any sleep, your "strings" still sound like a motherfucking beep in a motherfucking synth (starring Samuel L. Chipson). Well you know what? It will always sound like a beep. Deal with it.

What important is not that this one particular instrument sounds so much like that thing you want it to sound like. You know why? Because, at least as far as I know, people don't listen to your project's original tracks in a studio, they listen to the final mix, on their stupid computer. To them it doesn't matter how much your sound sounds like the sound you want your sound to sound. They want a song. Not data.

As for you, you should not be trying to communicate through sound how clever an instrument is made unless the purpose of your music is actually about technical wizardry (hey... it's valid too). But for more musical purposes, if your idea was to have strings play a particular part, then all you need to do is convey the idea of strings

Well how the jolly fuckpoop do you do that, I hear none of you ask? Easy: as long as it kinda sounds like strings when the rest of the song is playing, you've got strings. You can fiddle around with the subtleties (see previous) of the settings until you get a better sounding instrument, but once the basic set up of the instrument is made, the bulk of the editing should be done while other sounds are playing. It doesn't need to be constant, you can edit, then play, then edit, then play. But you need to check how it fucking sounds with the rest. I personally like to edit the basics of a sound, then make a simple pattern that fits well with the song I'm composing and then loop that pattern over and over. I'll tell solo the instrument I'm editing and start choosing other channels to bring into the mix as I edit my sound. For example:

1) Make basic instrument, create pattern, loop pattern.

2) I'll unmute the drums first. Drums are usually the loud points in a song, so I'll use them to set the general coarse volume of my instrument. If it's glaringly too soft or loud, that's where I fix it.

3) Unmute the bass next. Again I check for volume. I adjust my instrument's volume until it sits well with the bass.

4) Start editing the instrument until the texture sounds kickass with the rest playing.

5) Unmute another channel and go back to number 4 until everything is unmuted.

The same goes for FX editing. Test it against your track, every time! Sure as balls, every single solo instrument in the world sounds better with chorus and reverb and a bit of compression, and maybe a little tape sat here, some quiet slapback, eq the fret buzz out, noise print the shit out of it and use a reducer, aaaaaaaaaaaaaaaaaaaaaand there we are, sounds like a ton of bricks. Treat every track this way and when they're all playing together, you get a bombastic mess of shit that sounds too blown up. Some shit needs to be in mono. Some shit needs to NOT have a five minute delay with infinite feedback and gnomes. So when you add effects to your track, just like when you add instruments to your song, repeatedly check them with the track playing, and do the bulk of your editing there. 

Treat your main melody like a singer
No that doesn't mean to wait until the show is over and then try to have sex with your main melody backstage. It means to put some life into your shit. And I don't mean growing mushrooms on a turd. What was I talking about? 

A good singer doesn't just hit the notes as they are written. Same goes for instruments where the performance is more than just speed and volume. The guitar is similar to a singer in that notes can be bent willy nilly. Good singers, and good guitar players, put feeling into their craft. And as dumb as it sounds, in pure technical empirical data, feeling in music pretty much translates directly to going off key / out of tune. 

So your main melody is, let's say, C E F F G. Pretty boring melody and I'm not convinced you should continue this song, but who am I to judge your art? You want to put more life into it. Some pzaaz, or hutzpah, or other vaguely jewish-sounding words without proper dictionary entries. Some tricks:

1) Bend down at the end of sustained notes. It's that simple. Once in a while you'll have a note in your melody where this little trick will make the notes pop out a little more. At the very end of a note, about 1 to 4 rows depending on your speed, you can add a little 2XX effect to pitch the sound down. You can do subtle pitch bends (210 or 220) to add a touch of guitar-like playing. You can add more pronounced portamentos up to 2FF to make more extreme bends. The subtle stuff works best to conserve the melodic aspect of the song. More extreme version often add a distinct chiptune feel. Try different stuff out. You previous melody could bend down between the two F notes, breaking the repetitiveness of two consecutive notes.

2) Bend up to your note. Directly inspired by vocals and guitars, bending UP to a note is usually a very fast bend from a note that's not too far down. In most cases you'll find that one to three notes down the scale will be just perfect for the effect you're shooting for, or the next note down in a chord. This is done by inputing two notes back to back, and making the second one a legato note with a slide effect. Let's say the last note of your melody, the G, is on row 16, and the key of your song is C. You can slide up from C, D, E or F, up to G and it should sound pretty good. Any more distance than that and the portamento will take longer and might not reach the note in time. Could be cool for effects though! Let's say you choose D, because you want the D. On row 16, instead of having G, you put D. Your G now goes on row 17. And on that G note, you add the L ans S effect in the effect column by tapping the "1" key when the cursor is on the first two digits of the effect column. There ya go.

3) If you listen closely to a vocal performance, or guitar solo, you'll notice that a LOT of the notes in the melody have vibrato added to them. Sometimes subtle, sometimes not so subtle, vibrato adds a little life to otherwise stale and mechanical melodies. There are two basic ways to do vibrato: you either use the vibrato function built into the instrument itself, or you do it all manually in the Command column. If you do it directly from the instrument itself, you'll have a little bit less control over how the vibrato is acting, but it will also be much simpler to use later on. If you do everything manually, you'll get perfect control over everything, but it takes about 800 years to edit it all by hand. The best solution in my opinion is to switch between one and the other according to what the song needs at any given time. I usually reserve the instrument's vibrato for subtle / musical effects, and I keep the manual work for when I need a more extreme or complex vibrato.

4) Let the fucking silence speak a bit. Too many newcomers to music fall into the habit of layering 600 instruments together while they all play 16 notes per beat. This is especially true in electronic music where the "emptiness" between notes can be seen on your screen. Seeing these holes, the amateur musician almost invariably thinks "dammit! must fill everything!" and then ruins a perfectly serviceable song. Don't be afraid to let the song breathe a bit. When a particular part of your song isn't clicking right, sometimes fixing it isn't about adding new stuff, but removing the extra bullshit. In music, less is almost always more. Do you really NEED all those notes? I mean sure you can fit them in, keep them in key, and make something that sounds okay. But was it necessary to convey the idea? Can you achieve the same effect with less? Yes? Then do so.

5) And if you must speak, it's okay to repeat yourself. A lot of people just starting out in music develop the reflex to try and avoid repeating a note twice or more times in a row. The thought process behind it is usually that you don't want the audience to think you couldn't come up with an actual real melody. But that's bullshit. In fact, using the same note over and over again creates tension and anticipation for the next change and can be used to great effect. So y'know... don't be afraid to have C-C-C-C-C-D#-F as a melody if that's what kicks the song into second gear.

Well that's it. Those were just random thoughts about music and tracking in general. Next time, we'll take a long look at all the stuff you can use in the Command column.