Dance music is based on repetition — and if you are repeating
something, you've got to make sure it's something that's not just
likeable. You want to have something that gives
you goosebumps, that raises the hairs on the back of your neck. The
most important and significant parts, for me, are the chord structure
and the melodic elements. It's got to be something that people get
straight away, something that has that right feeling from the word go.
Sometimes you get an element of that, but the melody can be overly
complicated. If you're doing dance music, it's got to be quite
straightforward. So take your mix and play it to people. They're not
necessarily the best judges, but a good uplifting piece of music will
generally get a good reaction from a lot of people on first listen.
Arrangement is all about anticipation.
You can listen to a record you've never heard before, and still
instinctively feel when certain things are going to happen, because the
music gives you clues. So getting that right is the key. In the club
context you're doing it in terms of a seven or eight-minute 12-inch mix,
and from a radio mentality point of view it's within a
three-and-a-half-minute framework. I usually start by thinking about the
12-inch arrangement: if it doesn't work as a 12-inch, it just doesn't
work.
It's about the different elements you can put in.
You start off slow with simple elements, and you build up. If you want
more emphasis going from one section into another, you might drop just
the bass in a fill, or the kick and the bass, and when it kicks back in
again, there's a little bit more emphasis. Snare rolls are famous for
obvious build-ups, as is taking a piece of white noise and filtering it
to create a 'whoosh', or using sound effects. You can create your own or
use libraries. You can filter sounds. For example, I've got a big, warm
Moog pad here, and I'll open the filter to suggest there's something
coming up, or when there's a verse playing and I just want it more
thumping it might be filtered down. And then there's obviously the
natural rhythmic things that a drummer would play, like a crash on the
first beat of the bar.
When you've got those things in the right order,
it's just a case of how long you extend the track out for before it
feels right to bring in a breakdown, or how long it feels right to have a
breakdown for, or how long a build-up is right for. That is purely
based on listening to the record and thinking what feels right. It's
very hard to have a rule of thumb.
In trance records you usually get one breakdown or
two. The classic 'hands in the air' moment in trance is usually the
breakdown, and then it builds and builds and builds. House music is more
consistent, it's about playing with the constituent parts and bringing
them in and out. It still drops basses and kicks, but it's not so much
about big anthemic drops and huge builds, it's more subtle. It's very
much about chopping up certain elements. So not only might you have a
four-bar phrase, but you'd take the first two bars and repeat that, and
screw with the mix of that, and then create a build by using filters and
using additional drums and then kicking in again.
Fresh stuff and fresh sounds help you write melodies and create ideas.
Go down to a music shop, play around on all the keyboards and listen to
all the new stuff, and you'll find yourself inspired to do something.
The most important thing is to have a degree of originality.
I get inundated with records that are basically the same thing. We need
to be making the most of technology to get the most effective results,
and the great thing about it is that you can experiment with it and
create sounds and noises and effects that set your record apart, that
give it something fresh and new. On top of having your melodies right
and your arrangement and production, you've got to come up with
something fresh too.
I'll start playing around with different sounds, but
in the context of an idea I have in my head. The normal rule is, I've
got the melody or I've got the chords, I've got an idea. I would either
have found the sound first, in line with those chords, or I'll have the
chords and I'll think to myself 'Right, I need an angle here.' For
example, my current record is 'The Sun Is Shining Down On Me'. I had the
idea of doing a trance record with a soulful male vocal. The regular
thing is white female vocals, but I don't like doing the regular thing, I
like to do something a bit different. And I later reinvented the record
by bringing in instrumentalists, live bass and everything else, and
turned it into a house record. But I had the idea to do that in the
first place, and it worked.
If I'm doing a trance record or a Trance track, I'll have a template set up in Logic.
I'll have a set of basic MIDI drum patterns which gets me into the flow
of things. They're the basic building blocks. It'll be your standard
909-ish kick, your standard closed hats, the same kind of kit that
everyone is using. If it's house, I've got a slightly different set of
samples that I like — different snares, more realistic samples, more
heavily compressed sounds to give it a bit more bite and edge. That's
when I'll start doing the chords and things and get into the flow of it
with the idea in mind.
Then I start building up the arrangement. What I
tend to do is work on a four- or eight-bar section originally, rather
than just go 'Oh, I like that,' and lay it out for eight minutes. I'll
just layer down the chords and the sounds, and then start finding the
right sounds and putting together elements of track that I like. I keep
building the track up and tweaking the different elements. I get a
little section up and running and sounding good, and then I'll probably
start putting the arrangement together at that stage.
The point about being a cutting-edge dance producer is not just producing something, it's producing it really, really well.
It's making everything sound fantastic. I can mix down a dance track
and make it sound like a pop track and it will sound crap in a club. It
will sound very nice and pleasant, but it won't sound thumping and
banging, and the sounds won't jump out at me, and it won't be doing
enough — but with that kind of record, you need it to do that. Whereas
if I produce a pop-house track, it's much more of a toned-down sound,
there's not such an emphasis on blasting certain frequencies and
creating noise at certain levels. It's about creating an effective radio
record, which is not so extreme musically.
Another key thing is knowing when it doesn't work. I
start umpteen tracks where people go 'That's really good, I really like
that' — and I'll still scrap them. For me, they're still not good
enough, because all those elements aren't quite right.
What you must do is listen to your gut instinct.
In the old days I never used to be able to mix properly, because I had
no acoustic treatment, and it was like monitoring in a cave. I used to
do the arrangements at home and then hire other studios to mix. And what
I found was that I was hiring these studios and listening to the
engineers, and they were going 'Right, you want that, and you shouldn't
be doing that...' and I'd take my track away and think 'How come this
sounds even worse than when I was doing it in my stinking setup at
home?' I started realising that when I was in these studios, my gut
instinct wasn't necessarily agreeing with what the engineer was saying,
and I started saying 'No, this is how this track has got to sound.' And
that made all the difference. As you listen to it in a correct
acoustically treated environment, you start to develop that ear that
lets you recognise when something sounds great or doesn't. When I
started hearing things in a proper environment was when I realised I
could do this myself.
I don't want to hear any rubbish going on in the room, I just want to hear what's coming out of the speakers.
I did all the treatment in here myself, and originally there weren't
enough panels on the wall, and I kept putting more and more on. I like
the sound quite dead. But because it's slightly coffin-shaped it's quite
good, because you get the sub-bass sitting at the back like in a normal
studio, and you can get quite a realistic impression of the bass
response. If I really want to hear the lowest frequencies I've got my
Quested 3208s. I don't need the sub, they're so heavy as they are. You
can sit at the back and listen and it'll sound fine on the NS10s,
because they only register so low, but you listen to the 3208s and
suddenly you'll hear this huge frequency that you wouldn't have heard.
You won't have realised how boomy it is until you hear it on a monitor
that has that bandwidth.
What you have to do is get your sounds right in the first place, and then take it from there.
I remember going to Turnkey years ago and saying 'I can't get my mixes
to sound like this,' and they said 'Oh, what you want is a Finalizer.'
That's quite possibly the worst piece of advice I could have been given.
I know a lot of engineers who'll go 'I've got to record with this kind
of mic,' or 'I've got to get that,' but it's not much bloody good doing
that if you then set it 2dB too quiet in your mix, or you use the wrong
effect. I've heard recordings made on really good microphones and they
do sound lovely, and there is probably a slight tonality difference, but
there's bugger all you can hear in a dance record. There's too much
stuff going on. There's more important things to get right. Say, for
example, I have a pad, I'm thinking about how heavy the bass that I've
chosen is, how the EQ of that works with my pad, and how that fits into
the whole scheme with the drums in there as well.
If you want a challenge as an engineer, nothing's harder to mix than a trance record.
The whole way trance works is that the big noises sound anthemic
because they have ridiculously long reverbs, or long delays, or a
combination of both. You need to use a lot of reverb, and if you're
using a lot of reverb on a lot of sounds it clutters up the mix. In one
instance I used a reverb which was 30 seconds long. As time went on the
reverb times got shorter, but for trance it's still six or seven seconds
on lead sounds. But you get a very nice balance if you combine that
with some very dry elements, because you get the difference between
sounds like the drums being up front, and the sound of something that's
massive behind, and the overall effect is very big. The reverb is
cluttery, but the reverb belongs to the frequency of the instrument
that's playing it. So the kick drum ain't producing reverb (unless
you're a certain kind of producer, in which case it can, but that's a
whole different ball game).
There are certain synths that are infamous for
trance sounds — the Roland JP8000 and JP8080, obviously the Access
Virus, and I like the Novation Nova. Some of these sounds have been used
to death, and combined with certain effects, they make what is now
considered a standard in trance production. But you can still do
something with them. If you combine the Roland with something that's
heavy and wet, and then put that against a nice plug-in soft synth
that's very dry, maybe with a delay, each sound still has a lot of
separation, even if they are in the same frequency range.
Sometimes I'll do something totally bizarre and use orchestral instruments in a dance track, like a 'cello or a French horn.
If I blatantly put it in, people will think 'What the hell are you
doing?', but quietly in the track it works a treat. As long as their
dynamic and their frequency range works in combination with the other
sounds that are present, and they don't clutter the mix up too much,
it's great.
You can do a lot of work on a kick drum.
Sometimes I've taken a kick drum off a record, but what I then do is
put it in an audio editor and find another element of a kick drum that I
like, like the attack, and line it up, and on the audio channel I'll
crossfade between them to create my definitive-sounding kick. And then
if it needs more sub-bass, I'll add the lowest frequency from a heavier
kick. On top of that I might layer a really thin snare to give it a bit
more bite. And then on top of that, I might have the Pultec EQ emulator
running. I'll listen to a kick drum and think, bearing in mind where
club speakers have their pronounced point — generally around 60Hz to
100Hz — I know that for this particular mix that I'm working on, I
really want it thumping at around 100Hz, that's the really chunky part
of the sound, whereas the really low part just creates an unwanted
'mmmph' that isn't necessary, so I may roll off the really low sub
frequencies. On top of that I might then compress it to make it a little
bit tougher.
What a lot of people do in trance to create snare sounds is to get white noise and compress it really heavily, even using the Logic
compressor, so it cuts in with a heavy attack and then cuts out. It
makes a blisteringly sharp clap noise, and that works really well. You
can put that in stereo and put a nice reverb on that.
With hi-hats it's about making them sound tight,
with a sharp attack. I've got an SPL Transient Designer which I don't
use much any more, but in the past I remember using it to bring in the
attack, it just trims start points so you can make things sound really
clipped. I think with this kind of stuff, drums should sound very tight
and really stick out in the mix. I tend to use short reverbs on hi-hats,
generally, but I tend to use quite a long reverb on crashes and things,
I like it lingering. If it's in a breakdown and I just want the crash
floating off into the distance I'll use automation to switch on the
delay or turn up the send level.
All club systems are mono, pretty much, so just make sure you don't do anything stupid and put your mix out of phase.
I sometimes like my drums to sound very mono if I've got a lot of other
stereo stuff going on, but sometimes it's nice when they sound really
wide when you've got less going on. Some people I know use two hi-hats
and pan them left and right, some put it centre in the mix. Both work,
it depends what effect you want.
Compression is an integral part of sounds, and an integral part of components of sounds.
Often I'll compress the reverbs on the drums on every beat, so that you
feel it pumping. You'll hear the reverb come in and out, and that's a
really nice effect. It makes the kick itself sound more defined — rather
than the reverb washing over the kick that comes on every beat of the
bar, the kick comes through and then you get a kind of 'oomph'. I don't
use the kick drum itself as the side-chain trigger — I create a MIDI
pattern so that I can control how that compressor fires against the
drums. That way I can make it compress on every kick beat, but I can
also make the track compress without the kick drum running. Often I'll
use it on other instruments too, and it creates that great bouncy,
pumping feel.
Engineers are taught to use EQ to cut, but the most natural thing is to boost. I mix as I go along, and I'll leave lots of headroom on the desk, I'll keep my levels low in Logic so
I've got room to play in terms of automation, and I'll tend to boost
things when I need to boost them. I tend to boost more high frequencies —
1kHz upwards tends to be my boost territory. If I have unwanted
sibilance or nasty noise at the top, I'll roll off some of the highest
frequencies. Usually all my drum sounds have been treated individually
with filters or EQs.
They say that in America mastering engineers
master to the bass, whereas if I do a dance mastering session here it's
all about bringing the kick out. The bass is a good thing to
duck sometimes, if you've got a running line — it gives you a little bit
more impact when the kick fires — but then again, sometimes it's really
nice just under it. It depends on what you're working on. I've done
some R&B as well, and the emphasis is different. The kick's there,
but the bass is really important.
Basses are a bit annoying if they cover a lot of
notes, because you'll often get certain notes that just physically sound
louder than others. What I do if I have a real overload on a certain
note, a real 'bludge' of sub-bass just at that certain frequency, is
automate the level or automate an EQ, and just on those notes I'll lose
that. A compressor won't always pick it up, it's something you usually
need to do with an EQ.
Vocals are a world unto themselves in terms of production.
Bringing up and balancing vocal levels is really important in dance
music, where you need the vocals sitting up there in the mix. You don't
want to just lose things. If you're working with the vocals from the
beginning, sometimes you end up doing an instrumental track and you have
to fit a vocal on. That can be quite hard, but obviously there's lots
of methods of doing it. It depends on the style of the vocal, and
whether it's a lower male voice or a slightly higher female one.
Higher
can be easier to place, but it completely depends on how much is going
on in the track. If you've got very little going on, for a house record,
it's quite straightforward, especially for empty-ish funky house
records because there's tons of space. I'll tend to compress vocals
quite heavily for dance records, I'll tend to EQ them quite brutally,
I'll effect them as required and do all sorts of crazy stuff to them.
The final discipline is knowing when to cut something to make a track work better. I think that's one of the hardest things people have to learn — sometimes, less is more.
I'll always find things wrong with a mix if I take it away thinking it's perfect.
I tend to mix vocals last, and then once that's done I still end up
going over my mix again. I have this little method that I work by, where
if I've finished a complete MIDI arrangement but haven't done the final
mix, I'll come in and really get that mix sounding as tight as humanly
possible, exactly where I want it to be, and then I'll take it away and
listen to it. I can guarantee I'll still think 'That needs sorting out,
and that needs sorting out...'
What I do sometimes in house music is that I won't
just bounce down the whole mix when I'm done, I'll bounce down the drum
part and then I'll bounce down say the strings and the bass, and do
maybe three or four bounces, and then I'll start pissing around with the
stems themselves, adding effects and stuff to each bit, be it filtering
or flanging or whatever. In one instance, when I did Angelic's 'It's My
Turn', I put the whole record through my DJM600 mixer, which has got a
wicked flanger in it, and I just used it before the record kicked in
after the breakdown.
To me, if it sounds great in my studio, and I'm happy with it there, it sounds great in a club. I was writing club music and dance music before I went clubbing.
