How I Write Songs
I wanted to share my process for writing a song from scratch. I’ve gone pretty deep into this with my 30 day Ableton Production Journal but those of you who haven’t purchased it are still left in the dark. Plus I figure if I could give a breakdown of my process, it might help people save some time making their own songs. Some of this might be a bit technical and understanding it might push you to research further in Ableton or google. I’m not purposely trying to confuse anyone but rather wanting to go through my basic process in an unedited way. I’ll do my best to be clear.
This is pretty much how it all starts. For me, I rarely have “go to” sounds or templates set up. I probably should in some cases, but I usually can’t be bothered. Besides, it doesn’t take long to set things up. I’ll usually drag a reverb and a delay onto the 2 default return tracks. I’ll add another return track for Compression. This will be used for Parrallel or “New York” compression, which is basically mixing the fully compressed sound with the original. This has the advantage of adding punch and fatness without losing the transients. You also have control over how much you want to add to any of your tracks.
After adding the return tracks, I’ll usually drop a Drum Racks kit into a track to make a basic beat. I dial the sounds in as I go but I usually want to find a groove with a good set of sounds. Again, although I do have custom kits, I don’t have any “go to” sounds. I just keep testing sounds until something catches my attention. Once I have a working beat, I’ll dial in the volumes, attack, decay and release on each sound. I’ll use compression to try to get the most out of kicks and snares. I’ll also drag in a reverb as a return within drumracks. What is great about Ableton is that I’m able to do anything with each sound within drumracks as I would be able to do on any regular track (and that is ALOT).
I also have Drum Racks kits that are strictly hi hats, or snares, or kicks. I may find myself dragging several more tracks or drums to layer with the original kit. Usually 2-3 sounds make up a kick or snare that I’m happy with, typically choosing which sound has the best low end and which have the snap or high end. I’ll then EQ out the low end of the sample that doesn’t need it. Sometimes I’ll use a low pass filter as well to purposely take off some of the high end frequencies, giving the drums a darker or more dubby sound. Other times I’ll enhance the the highs for a more agressive sound. It all depends.
Next I’ll dive into bass. Sometimes I know exactly what I’m looking for, but usually I haven’t got a clue and I need to find the right sound before I know what to do with it. Depending on whether I want a more analog or digital tone I may look toward a virtual Moog, Juno or Ableton’s own Operator or Analog. Since I know Operator pretty well, it’s easier for me to dial in a sound from scratch than with most other synths. With the other synths, I rely a bit on presets as a starter point and the tweak to taste. Another great bass tool is Spectrosonic’s Trilogy. A huge range of great acoustic and electronic sounds. With some groove quantizing and good programming, you can get some pretty believable sounds. If I am not already using a simple sine wave, I’ll usually add an Operator just for the purpose of Sub bass. Then I’ll back off the low frequencies of the other synth. I may also back off the highs of the bass so it’s not too aggressive. If i later need more “bite” I may duplicate the synth and pitch it up an octave, knock out most of the lows and mids and add a bit of distortion or saturation. I usually can keep this layer quite low and it’ll still have an impact on the overall sound. You can duplicate again and add another layer pitched up 3 or 7 semitones to create a harmonic if you want. I’ve found this to give me a far more exciting sound at times.
For drums and bass, I’ll usually create a group for each and add a subtle compression (2:1 or 2.5:1) just to make things gel together. I rarely add reverb to my bass or the kick drum, but certainly use it lightly on other drum sounds if it helps to add depth. With Drums, I find it important to have a combination or wet and dry sounds. The dryer sounds being the ones you want to stand out while the others play more of a support role. I’ll also try to remove any “mud” in my bass or drum sounds. I use a highpass filter to knock out anything unneccesary. I also will put a dip in the 350hz-650hz range to sharpen some sounds. This range tends to by where “mud” lives. A little subtractive EQ goes a long way.
The next step I might take after getting my bass sounding good is to tune my drums to the bass. This is easier than it might sound and there are a couple approaches that work for me. The one I use most is using an EQ8 effect and sweeping the frequency with a thin Q and alot of gain until the sound rigs in harmony. Then I’ll back off the gain so that the tuning is a bit more hinted instead of in your face. I may add a 2nd or 3rd harmonic on the same sound if it seems to work. Another approach to tuning drums I learned from Ableton guy Dennis Desantis’s misuse of the Frequency shifter. Using the fine tune can give some great results without ruining the sound.
Sidechaining can be a godsend for much more than just making that dancefloor “pumping” sound. Used more subtly it can really help sort out places where the low frequencies of the kick drum and bass interfere with eachother. I typically decide whether the low end of the bass or the kick is more important and sidechain the other. By layering an instrument rack you can sidechain only the low frequencies and leave the higher frequencies untouched. This makes the effect much less noticeable but improves the clarity of the mix. I may sidechain the hi hats to the kick for that typical pumping sound as well as atmospherics and reverbs. This can really clean you mix up. Another use is in mixing. Instead of automating one part down everytime another part comes in, I’ll just let the sidechain compressor push the sound down automatically whenever it hears the other sound coming in. I may duck a pad to make a lead or vocal sound more clear. I highly recommend you look into this.
Then next step I will usually attempt is some sort of Stab sound. Something with a fast attack and decay. This is the part that will play of the drum and bass rhythm. The goal is to get a good sound that cuts through the mix and doesn’t screw up the momentum of the track. For an idea of a stab that really drives a dance song, look no further than Lil Louis – French Kiss. People have been ripping this off for ages because, quite simply, it works! If you have an arpeggiated bassline, you might not need a stab. I’ll usually create a 4 bar stab-like part just so there is a bit of variation in the pattern. Be careful not to lose the hypnotic quality of the part by changing things up too much (unless that is exactly what you are going for).
Pads, stretched tones
I’m a bit lazy when it comes to making pads and they are typically an afterthought even though they can be incredibly important. They give a track that loops alot of movement and keeps the song sounding too clean, which can often make things sound too mechanical or even unprofessional. A great way to get a good pad or drone is to stretch the hell out of a sample or play it on the low note of a sampler. Stretching vocal samples it a favorite of mine. It can sound so good so easily that it almost seems like you’re cheating 😉 I will typically add delay and long reverb to make it sound nice and wet and to smooth the edges. This is a fantastic way to add some much needed atmosphere and movement to a track. Just make sure you are in the correct key. A simple pad layered with a stretched vocal or sound effect can give you excellent results.
Sometimes chopped up pads can make a very good stab or melodic hook sound if done correctly. Sasha’s – Xpander is a perfect example.
At this point in my songwriting I usually have to make a decision. Either let the song cruise along as is without an obvious hook or lead riff, or realize that it would seem lacking without it. In my 30 Day Ableton Journal. I attempted several leads but in the end, decided it was better without them. This is a completely subjective thing as your musical goals might be different than mine. As long as you are using your best judgement and you aren’t just throwing parts in because they are “supposed” to be there, you should be in good shape.
This is where I do a bit of filling in the blanks. I’ll usually drop crashes and reverse cymbals in their logical places. Sometimes I just use these as markers to possibly be replaced by another sound. These will usually be placed at 8 or 16 bar intervals. Pretty obvious stuff, but definitely makes a difference. Some artists don’t use crashes at all. Listen to Daft Punk for example. If you can make it work for you, go for it. Personally I’ve tried, but it’s deceptively difficult to do what they do and still keep things interesting.
Once I’ve got the basics crashes and reverse sounds in place, I’ll put atmospheric sounds with a good amount of delay, reverb or both in spots throughout the mix. Especially when a build or breakdown is needed. Without overdoing it, I just filled in the gaps. I check to make sure the intro part of the track isn’t too empty. Although it’s nice to save the good part for later, you don’t want any of your track to be boring. Using cut edits of parts that play later with some delay can give good results, especially if you are lacking hi frequency content.
This part is always tough for me. I know it shouldn’t be as I’ve been making songs since 1988, but structuring club music can still be challenging. One example of a song structure may look like this:
Intro: 16 Bars
Bass dropp: 16 bars
Added layers and elements: 32 bars
breakdown: 4/8/16 or 32 bars
Full on track with hook or lead: 32 bars
short break: 4/8 or 16 bars
groove with less elements: 16 bars
outro: 16 bars
Although it’s good to have a fairly solid structure, sometimes using odd times can create some nice tension and make your track less predictable. Otherwise, straying from multiples of 4 bars will just sound off. Use your own judgement and don’t be afraid to “borrow” the structure from another song. People have been doing this for ages, it’s not a big deal.
Next I go through the mixing stage and dial in the panning for each sound. I may add an auto pan or automation to some sounds to give some random movement. This works well on atmospherics while subtle panning is good for drum sounds. I like to use Ableton’s auto filter to give subtle movement as well. I may do a stereo spread of the upper frequencies of the Bass while keeping the subs mono. Keep in mind the panning is also important for sounds with the same frequencies. Often times panning is more effective than EQing to bring out a part’s clarity. Unless I’m going for an artificial sound, I try to avoid extreme EQ’ing. For a more natural sound start with panning and then attenuate unwanted frequencies instead of boosting frequencies. If you are using extreme boosting to try to bring out the bass in a sound, you probably should rework the source sound. You can’t bring out something that isn’t there to begin with.
When it comes to mixing, it’s a good practice to listen to all the other parts besides the one you are mixing. It’s very important to concentrate on how a part effects the totality of your song. If you do find yourself focusing on the part you are mixing, take it to where you think it sounds good and then drop it back about 10-15%. If you don’t get this habit under control, you’ll end up with the newbie mistake of having the first parts you mix buried by the last parts you mix.
Once I have a basic mix I like to switch to mono. In Ableton it’s easy. Just drop a utility effect on the master track and change the stereo spread to zero. Then turn off 1 of your monitors and get in front of it. I have written a whole blog post of mixing in mono, but in short, mixing in mono reveals phasing issues, if your effects are overdone or if your panning is making a part less clear than it should be. Another great thing is that in mono, there is no “sweet” spot. Wherever you are sitting, you’ll get the same results. If you can get things sounding great in mono, you can be sure it’ll sound excellent when back in stereo. Don’t forget to turn off your utility before mixdown!
Editing for groove
Once I’ve made it this far, the first question I ask myself (when making music for clubs) is “is this danceable? Does this make me want to move or just stand in one place and nod my head”. There is nothing worse than a well produced track that doesn’t move you… or your audience. If I find that it isn’t giving me the results I was hoping for I inspect what part is slowing the track down. I first make sure the drums and bass alone get me excited, then I add parts one by one until the track sounds less danceable. Once I find it, I’ll consider changing the rhythm or even pulling the whole part out. Then I’ll continue adding parts going through the same process. This can be frustrating after all the work you’ve put into each part, but it’s important not to be married to any one part. If it’s not improving your song, it’s making it worse. Also it’s not a bad idea to experiment with groove templates. They can make a pretty drastic difference in the vibe of your song.
Some of these final tweaks can make the biggest impact on whether your song sounds great or just ok. Typically at this stage I’ve made a pretty safe mix. Everything has it’s place and nothing is taking up too much attention. That’s a good place to be but not a great place to end. It’s important that you let certain parts have their chance to shine, so for each important part, find out where that is. Once you introduce it loud enough to grab the listener’s attention, they will hear it throughout the mix even when it’s lower in volume. These little moments for each instrument gives your song more dynamics and constantly keeps your attention. A little boost goes a long way, so don’t try to make ears bleed here, just make sure the listener knows when something new has arrived.
Once I’ve gone through this process, I’m usually pretty happy with my mix and I mix it down. Of course you want to make sure to burn a CD and play this on a few systems and do some a/b tests with some songs you like the production on and if necessary fine tune your mix. I’ve been known to be “happy” with my mix and still go back 3 or 4 times for final tweaks. Luckily this isn’t all that bothersome because you know you are very close to finished and that’s a great feeling.
Well folks, generally speaking, that’s my process. Everyone has their own approach and all are valid if they work for you. Hopefully this has given you some ideas which I encourage you to borrow or steal.
Happy music making,