prophetgoddess

i don't wanna die in here

workin on a dating simulator

podcast

blog

music

My scrobbles

lastfm thingy via here



i posted yesterday about how i have been teaching myself to dj. someone asked me what resources i've been using, and the answer is i haven't been using any, because they're all awful. i was able to figure it out because i have almost a decade of experience with digital audio workstation software and music production. here's my attempt at a guide that doesn't suck for people who don't have years of experience producing music.

1: history and technology

probably the most useful information i had before i started trying to learn how to dj was a history of how the current state of technology and practice came to be.

djing is a practical solution to a problem with vinyl records: when a song stops, you have to take the entire disc off the platter, find a new one, put it down, locate the next song, and then press play. this takes a while and can kill the vibe. djs realized that what you could do is have two turntables, linked together by a mixer. then, while one song is playing, you can find the next. the mixer has a little slider, called a crossfader, which allows you to switch which turntable the audience is hearing.

another important innovation is the notion of cueing and monitoring. dj mixers allow you to plug in headphones so you can hear both turntables while the audience only hears one. this opened the door for beatmatching. the idea is that you take two songs that are at roughly the same tempo and use your headphones to line up the beats from both. then, when you crossfade, you can seamlessly go from one song to the next without ever interrupting the groove.

the final piece of the puzzle for classic djing is effects. the most common and useful effects are filters and equalization. a classic dj style filter is a knob that, when you turn it right, becomes a high pass filter, and when you turn it left, becomes a low pass filter. at 12 o'clock, it does nothing. "high pass" means that it removes frequencies below a certain cutoff frequency. "low pass" means that it removes frequencies above a certain cutoff frequency. so, when you turn the knob to the right, it raises the cutoff frequency from 0hz to 20khz (or thereabouts, often it's much lower and even with the filter knob turned all the way right some sound will still leak through). filters also all have something called a "slope." it's basically impossible to simply remove all frequencies above or below a specific cutoff. instead, filters are built to reduce the volume of frequencies past the cutoff by a certain amount for each octave above the cutoff. a common filter slope is 24db/oct. this means that an octave from your cutoff frequency (which is twice your cutoff frequency, don't ask why it's called an "octave" when it's a doubling and not an octupling), the volume will have been reduced by 24 decibels. half an octave from your cutoff frequency, the volume will have been reduced by 12db. two octaves, the volume will have been reduced by 48 decibels.

equalization is a lot like filtering. equalizers come in all shapes and sizes, but the typical dj style equalizer is what's called a three channel EQ. you have three knobs, labeled "high," "mid," and "low." each of these is set to a specific frequency range. turn the knob to the left, and the volume of sounds in that range will be reduced. turn it to the right, and the volume of sounds in that range will get boosted. sound closer to the center of the range will get boosted or reduced more than sounds near the edge of the range.

the main purpose of filters and equalization is to allow you to cut out elements of songs so they blend better with other songs. if you use a high pass filter, you can remove the bass and kick drum from a song and just be left with the snare, hi hats, and any high frequency melodic elements like vocals, guitars, piano, synthesizers, etc. if you use a low pass filter, you can remove all of that other stuff and just be left with the bass and kick drum. this allows you to say, mash up the bass and kick from one track with the melodic elements from another, or add a vocal line from one song onto another.

modern dj technique was largely invented for use with disco music, which has a number of features that make it particularly well suited to mixing like this. disco music has a regular "four on the floor" beat, which means that the kick drum hits every quarter note, and the snare drum hits on the 2 and the 4. since every song in the genre has this same beat structure, you can easily use filters and the crossfader to seamlessly transition between songs. disco music also usually all sat in the same narrow range of tempos. when disco music started to vanish, other styles of club music, like house, electro, and techno, kept these same elements.

in the 1980s came the development of the compact disc. this led to the development of CDJs, which are devices used to allow djs to manipulate the audio on a CD in the same way that they would a vinyl record. this was a huge revelation: a CD could hold a lot more music than a vinyl record, and you could easily burn your own CDs at home with the right equipment, so DJs could put together a couple of discs with their whole setlist, rather than having to lug around heavy crates of vinyl. you no longer had to fuck around with getting the disc off the turntable, putting it back in the record sleeve, finding the next record, and getting it set up all in the short time the other turntable was playing. now, all you had to do was select the next track. the transition to digital audio also would allow some extremely important innovations in technology.

the most significant of these is automated beatmatching. it's relatively straightforward to write a computer program to detect the tempo of a song, and automatically synchronize it with another song. manual beatmatching is a difficult skill to learn, and depending on the material, matching beats in a live club environment could be almost impossible. along with automated beatmaching came digital time-stretching, allowing you to slow down or speed up tracks so they play at the exact same tempo. this meant that two songs at different tempos could be beatmatched for arbitrarily lengths of time. with vinyl, if the tempos of two records were even slightly different, they would gradually slip out of phase with each other. some turntables had manual sliders to allow you to adjust the turntable speed manually, but this was a fiddly process.

digital sound also allowed for the introduction of granular processing, which lets you play a track at a different speed without changing its pitch. if you speed up a vinyl record, the track will play at a higher pitch; slow it down, and it will play at a lower pitch. granular audio processing takes tiny tiny slices of a song, and plays them either less or more frequently, interpolating between samples, to give the illusion that the track is playing faster or slower without adjusting the pitch. this kind of approach has flaws that become obvious with large tempo shifts, but for small adjustments (+- 10 BPM), it works pretty flawlessly.

by the mid 2000s, computer hard drives had gotten big enough that CDJs were being replaced by DJ controllers. a DJ controller is a piece of hardware you connect to your computer via USB. it reads songs off your hard drive (or, these days, SSD), instead of off a CD. most CDJs were used like turntables; you (or more likely the club you were playing at) would have two of them, and conect them with a separate mixer. most DJ controllers have two turntables and the mixer all integrated. they also usually have additional buttons and functions that let you do things like play sounds (such as the ubiquitous airhorn), or put markers in a track so you can quickly jump to the section of track you want to start playing rather than having to search for it.

the vestiges of the turntable platters remain in the form of the jog wheels, which are used for navigating through tracks to find the section that you want. you can use them to imitate vinyl record effects like scratching, but most of the time you just use them as giant glorified knobs.

anyway, now that you know all of that, let's get into the actual tutorial.

2. hardware

the first thing you'll need to get started is a dj controller. my dj controller is this one. there are a lot of options out there. for a basic, entry level controller, you can expect to pay between $100 and $200. most controllers are built to be used with specific software. the djcontrol inpulse 200 is built for use with hercules' own DJUCED software. other popular options include the more expensive traktor kontrol, built for use with native instruments traktor, and the various pioneer and numark machines built for use with serato. i went with the inpulse 200 because it was among the cheapest, and it didn't have any features that seemed like stupid gimmicks to me. i don't regret the purchase, i would recommend it.

let us consider the anatomy of the inpulse 200. if you have a different controller, it probably has the same features.

  1. jog wheels. used for seeking through tracks.
  2. equalization. the inpulse 200 only has knobs for high and low frequencies, no mid. honestly, i rarely use these.
  3. gain. different tracks will be mastered at different volume levels. gain can be used to boost or reduce the volume of a track so it's the same average volume as everything else and it doesn't clip.
  4. filter.
  5. browser knob. used for scrolling through your songs and selecting the next one.
  6. load buttons. push the button for the deck you want and it will load the song you have selected into that deck. djuced is nice enough to not let you accidentally load a track into a deck that's still playing. i imagine most dj software is similar.
  7. monitor switches. push the headphone button on a deck to have its output sent to your headphones, regardless of the position of the crossfader. the "master" button allows you to toggle whether you hear the main output in your headphones or not.
  8. faders. they control the volume of each track. unlike gain, which can boost volume, the faders can only attenuate it. i like to keep these in the middle by default, and raise or lower them depending on my mix needs.
  9. crossfader. we'll talk about this more in a minute.
  10. tempo sliders. these are a little weird with automated beatmatching, since the computer will automatically adjust the tempo of each track to be the same but can't move the sliders for you. moving them down lets you speed up the record, moving them up will slow it down.
  11. play/pause. plays or pauses the deck.
  12. cue. this is really only used when you're not using automated beatmatching, so i'll ignore it.
  13. sync. enables or disables automated beatmatching.
  14. pads. these do four different things. "hot cue" allows you to set points in a track that you want to be able to jump to quickly. "roll" lets you loop a short section of track for stutter effects, useful for some transitions. FX lets you enable or disable global effects like reverb, delay, phaser, etc. sampler lets you set and play sounds.
  15. loop buttons. "in" sets the start point of the loop. "out" sets the end point of the loop. shift+out allows you to double the length of the loop.

3. software

every controller needs software to work with. in addition to djuced, which i use, there are lots of other options, including serato, rekordbox, mixxx, and traktor. honestly, they're all basically the same. djuced is fine, it has some weird UI issues that crop up from time to time, but i like it more than mixxx and rekordbox, the other free options.

let us consider the anatomy of djuced. again, your software probably has the same features, though they might be in different places.

  1. track visualizers. in djuced, bluer colors mean more trebel, redder colors mean more bass. this is useful for finding where drum or bass parts kick in without having to hear. otherwise, this shows the current location in the track and scrolls along with the playback. the lines indicate where the beats are, and the slightly thicker line with the red triangles on the ends is the first beat of each measure.
  2. decks. these are your two playing tracks and all the stuff you can do with them
  3. pads. these are the same as the pads on the controller itself. note there are more of them here, my controller only has four pads, the higher end hercules controllers have eight.
  4. bpm. will show the current playing bpm of the track, not the track's original bpm.
  5. key. we will talk about this in great detail later. also features arrow buttons for pitch shifting.
  6. beat guides. indicates which beat the song is on. the bottom, red indicator shows the deck it's in, the yellow indicator shows the "master" deck. the other deck will by tempo synced to the master deck when sync is turned on.
  7. mixer. same controls as on the controller.
  8. song library. allows you to browse through your songs. features playlists, which let you pick out and sequence tracks, rather than having to go through your whole library. you can sort by tempo, key, genre, artist, etc. most dj software also allows you to use songs from streaming services like soundcloud and tidal, but i play tracks from my collection.

4. tunes

in order to dj, you need music. again, dj software often lets you use tracks from various streaming services, but i'm going to focus on playing out of your own collection. if you don't have a collection, start one! bandcamp is a great place to get drm-free music for your sets. here are some albums i think are good for mixing:

you can, ultimately, mix anything you want. the set i streamed the other day features some typical club music, but also detours into indie rock, hip hop, and hyperpop. i wouldn't worry too much about style or genre, just play music you like. as you play more, you'll get a sense of what kinds of tunes go well together.

5. your first mix

ok, now you have some tunes. figure out how to load them into your software (sorry, this isn't that kind of tutorial), and then pick two. you'll want to pick two tracks that are at around the same tempo (+- 10 BPM), and in the same key. we'll talk more about mixing songs in different keys in a bit.

one song is going to be the one you start with. the other, you'll mix into. in the track you'll mix into, you want to find a section that can loop perfectly. you can often identify these by eye in the track visualizer. often with dance music, there will be an intro loop for 16-32 bars at the beginning of the track. other times you'll have to scroll around a bit. play around with the loop feature to get a sense of where these loops are and what they sound like. here are some examples of perfect loops in tracks i've mixed with before:

you can hear them here:

in order: hand crushed by a mallet, your favorite sidekick, toothless, yellow calx, rainbow six, yellow calx/rainbow six mashup.

(note that i've used the hot cue feature to store the locations of all these loops. djuced stores both the cue location and the in and out position of whatever loop you have set up when you press the hot cue button.)

it's worth noting that you don't need to be able to find a perfect loop to mix into a song. you can also transition using stutter effects, or by pushing play precisely when the first track ends or when you crossfade so the transition is seamless, or you can just gradually fade out one track during the intro of another track. sometimes a track will have an intro that's more of a build so there won't be a clean section that starts exactly the same as it ends. sometimes "good enough" is fine. one of the most fundamental rules of music is repetition legitimizes (s/o adam neely). if you grab a short section of a track and loop it over and and over again, it can start to sound intentional over time, especially when combined with filters and other effects to blend it into another track.

here's an example of a stutter transition (baby i'm bleeding by JPEGMAFIA to kick of a lifetime by leroy):

and here's an example of just hitting play at exactly the right moment (toothless by 100 gecs to copyright strike my fucking nuts by leroy):

once you have your loop, get ready to crossfade. turn on sync for both decks, and start playing both tracks with the crossfader slid all the way over to the track you're starting with. something that surprised me about the crossfader was the exact curve with how it works. i expected it to look like this:

but actually, it looks like this (red represents the volume of the left deck, blue represents the volume of the right deck):

(modern dj software makes these curves highly configurable, but this is the default and i wound up understanding why)

what this means is that you only have to move the crossfader a little bit away from the furthest left or right position for both tracks to be almost fully audible. the difference between having the crossfader at 50% and having it at 75% is very minimal. the crossfader acts more like a three-way switch with smoothing between to prevent pops and clicks than it does a smooth transition between two sounds. the reason for this is to reduce the amount of precision necessary for transitioning. if you want that smooth volume transition, what you'd do instead is move the volume fader all the way down, move the crossfader to the middle, and then slowly fade in the volume.

what i like to do is set the filter all the way closed for the looping track. for most material, i'll turn it all the way right into high pass mode, but for some i'll prefer to fade in from a low pass. then, i'll move the crossfader to the middle, and slowly open up the filter, turning it to 12 o'clock. sometimes, at the same time, i'll close the filter for the playing track. other times, i want the two tracks to play simultaneously for a while, and then i'll fade out the old track. sometimes i'll open the filter all the way, most of the time i'll only open it a little bit so we get the high end of the looping track without any of its drums or bass.

however you want to transition, before you do so, make sure that the beats are aligned; not just that the hits are happening at the same time, but that the first beat of each measure is happening at the same time for both tracks. use the monitor buttons to preview the transition before you do it, and then.... do it. now do it again. find another track at around the same tempo in the same key, find a good transition point, loop it, then transition.

something that is difficult to convey is what a good transition point is. generally, you want to bring in the new track when the old track is transitioning to a new section. some material will benefit from being brought in during a transition to a more chill section, some will work best if you bring it in during a drop or a hook. it takes practice, i'm definitely still not perfect at it. as a general rule, though, you might have the instinct to very slowly fade in the new material, but this is rarely the best option. i find that, in general, both in djing and in music production, you should add things quickly and take them away slowly (though there are, as always, exceptions).

6. harmonic mixing

so, you're probably wondering: what if the songs i like aren't all in the same key? good news: you can mix songs that aren't in the same key. let me introduce you to your new best friend, the circle of fifths:

the circle of fifths is a wormhole of music theory. it is monstrously useful for a huge variety of musical tasks. for us, we're concerned about which keys are going to sound good together. if you're not very musical, you might think that keys with adjacent letters will sounds good together: for instance, you might think that C major and B major would sound good together. that is not correct.

what matters for mixing is how many notes two keys share, not how close their root notes are. consider again C major and B major. C major has the following notes: C, D, E, F, G, A, B. B major, in contrast has B, C#, D#, E, F#, G#, A#. the only notes they have in common are B and E. now consider G major, which looks to be quite far from C major. G major has the notes G, A, B, C, D, E, F#. almost all the notes are the same! they're just in a different order. the circle of fifths is a way of visualizing this relationship.

it's called the circle of fifths because, going clockwise, each key is a fifth above the previous key. note that G is the fifth note of C major. going counter-clockwise, each key is a fourth below the previous key. note that F is the fourth note of C major. the circle of fifths is also sometimes called the "circle of fourths" for this reason. i love this wikipedia diagram because it also shows the key signatures for each key. you can see that going clockwise, you add one sharp for each key. going counterclockwise, you add one flat for each key. in music education, the circle of fifths is used as a mnemonic tool for learning and understanding the relationships between different keys and key signatures. the underlying reason why the sharps and flats are added in the order and location that they are is beyond the scope of this post. also note that the keys at the bottom of the scale have two names, one flat and one sharp. which one you want to use is a matter of personal preference and the specifics of the music you're playing or composing1.

you'l also notice the smaller inner ring of lowercase letters. these are the minor keys. A minor has the same notes as C major, just in a different order! two keys that have the same notes in a different order are called "relative."

as a general rule of thumb, you can mix two songs if their keys are adjacent in the circle of fifths. you can mix from C major to G or F major, or to A minor. some people say you shouldn't mix from C major to E or D minor, but i find it generally sounds fine. you can often go further than that: in my set i linked to earlier, i mixied straight from C major (came to my show by 100 gecs) to D major (shashshashshash by leroy) and it sounded fine. it depends on the particular characteristics of the melodies you're mixing: if a D major melody hits that C# and F# a lot, it'll sound bad with a melody in C major, but if it sticks to the other notes, it'll sound perfectly fine. use your ears to determine if something sounds good enough to mix.

dj software also usually allows you to pitch shift tracks one semitone at a time. you can, for instance, pitch shift a track one semitone down from Bb to A, to get it to mix with a track in A major, D major, E major, or F# minor. doing this wildly opens up mixing possibilities, but some material may sound strange when pitch shifted, especially anything with vocals. again, use your ears to decide if a pitch shifted track sounds good enough to use.

you might also wonder how to mix songs that are at very different tempos. there are a few things you can do here, but what i like to do is try and bridge the gap. to get from a slower song to a faster song, you can just pick songs that are faster than your current song by 5-10 BPM, adjust your tempo upwards to match the new song, and keep going until you're at your target BPM.

7. conclusion

and that is all there is to it. watch other people's dj sets, like the ones i linked earlier (people have linked lots of other good ones in the comments). listen broadly, mix fearlessly, and keep practicing!


  1. you'll note that all the other keys have one name. the key of Bb will never be referred to by a jazz or classical musician as A#. however, most dj software does use the name A# for that key, and most digital audio workstation software exclusively uses sharps to name notes in general. they're technically the same key, but the name "A#" sounds really weird to most musicians with formal training. this is because of the rules for naming notes in a key. you have to use every letter in order, you have to use them all exactly once, and you can't mix sharps and flats. major keys are constructed using a pattern of whole step, whole step, half step, whole step, whole step, whole step. if you call the first note Bb, this is fine: you get Bb, C, D, Eb, F, G, A. if you call the first note A#, however, you run into a problem: a half step up from A# is C, but you can't use C because you haven't used B yet, so you have to write the very weird note B#. then D is even weirder, you have to write it as some kind of C, so you have to write C##. yes, that's C double sharp. Eb becomes D#, F becomes E#, G becomes F##, and A becomes G##. that is an absurd number of sharps, and a comical number of double sharps. someone tell the programmers who make audio software about the existence of flats so i can stop having a headache.


You must log in to comment.