MIXING IT: Remixing with Mr Motiv8 - Steve Rodway on the art of the remix

Published in SOS August 1994
Printer-friendly version

mixingit

remixing with
Mr MOTIV8

"'Rockin' for Myself' was picked up by Warners, so I first worked with Andrew on remixes of that song. As my studio is down in Kent and I live only ten minutes or so from Peer in London, I've been doing my pre-production work and programming in the programming room here rather than at my own place. When I can move my studio up to London, then I'll do a lot more work at home; I reckon you can do about 80% of the work for this kind of music in a small home studio.

"I'm currently working on a remix for Positiva (EMI) Records, for a group called Mosaic, with mixes for New Atlantic and N-Trance. I don't just take anything that I'm offered -- there's no point in doing that. I'm only interested in taking on tracks that I feel I can really enhance and contribute to in a positive way. If you don't, you'll end up doing third-rate mixes and third-rate records, and obviously that's no good for your reputation if you want to be around long-term."

What's the history behind the Doobie's remix? Their material doesn't strike me as being immediately obvious for remixing into dance tracks.

"I liked the song; it had a good melody and an uplifting feel, but I didn't think it would be as hard to do as it actually turned out to be. The other Doobie's remix, 'Long Train Running', had the guitar riff, which is the key to that record, while 'Listen to the Music' doesn't really have any outstanding features other than the vocals. Also, the tempo was considerably slower than 'Long Train Running', which was already at about 114 bpm, and that made the job of getting it up to a House tempo more difficult. The original was around 107 to 108 bpm, and trying to get that up to 124bpm presented certain technical problems. I found, for example, that when using the S1100 for time stretching the vocals and guitars, anything above about 8% was unacceptably glitchy and the acoustic guitars start to sound shimmery or grainy. In the end, I used the S1100 to time stretch by about 6%, and then pushed everything up a semitone to arrive at the final tempo."

What material did you have access to? Did you simply work from the single or were you provided with copies of the original multitrack?

"I had a copy of the original 16-track tape dubbed onto 24-track and sent to me by Warners, California. The quality of the recording, which was made in 1972, was outstanding. When you put the faders up, it sounded incredible, especially the drums.

"We took all the vocals, and some of the electric and acoustic rhythm guitar parts. We did sample other stuff, like drums and percussion, just in case it was needed, but we ended up not using it. In order to get a modern sound, you invariably have to change the bass and the drums. In fact that's the key to this kind of remix -- you leave the top half of it pretty much as it was, and change the bottom part."

Once you'd sampled the various elements and got them to the right tempo, what was the first step in rebuilding the song?

"One of the hardest things was deciding on the direction of the song, because the bass line on the original was a fairly simple, supporting line, whereas dance bass lines are very up-front and persistent. I went through three or four options before I found one that fitted -- and, remember, this had to sit in with the original guitar and vocal parts. In the end, I opted for an updated version of the original bass line, which is a supporting line that doesn't stand out too much -- if it had, it would have detracted from the melody. The sound we used for the bass was actually a sample taken from a MiniMoog."

Did you stick with traditional Dance drum sounds?

"Pretty much, though the bass and snare sounds are 909 sounds to which other sounds have been added. The bass drum, for example, is a mixture of TR909 plus a fairly hard, middly kick sound -- not unlike the old Linn. If you get the kick wrong, you're in trouble when you come to build the rest of the track. The snare was TR909 with a very metallic, almost piccolo-like snare mixed in. I started out as a drummer, so I loved big snare sounds with gated reverbs, but then I realised that people don't dance to a snare drum, they dance to a bass drum. Now the fashion is to have fairly small snare drums, which helps mixing because they don't take up so much space. Smaller, boxier snares tend to work in a functional way rather than being a main feature of the track.

"The hi-hats were classic 909 samples; there was a shaker, a tambourine (sampled from a Yamaha RX11 and repitched) and a long crash cymbal. I also used a sampled drum loop to sit behind the programmed drums -- just enough to give it a lift. I'm a great lover of loops, but you have to be careful with the top end, as it can sound phasey when added to other drum parts. Often you have to EQ it down a bit."

How do you approach EQ when working on this type of music?

"I'd say that we don't tend to over-EQ most sounds; if you EQ, say a hi-hat, you're not just changing the sound of the hi-hat, but also the way in which the rest of the mix is perceived. If you make the hi-hat brighter, then the ear is going to be less susceptible to those frequencies in the rest of the mix -- it's like a form of masking. I think EQ has to be applied very carefully, especially as the SSL EQ is pretty harsh. I do occasionally use drastic EQ on certain things, but I think that only works when most of the mix has been sparingly EQ'd.

"I learnt a lot about EQ from having my own studio, and in the early days, I'd use quite a lot of EQ, but things wouldn't sit right in the mix. At first, I couldn't understand why, on my Soundtracs desk, the monitor mix always sounded fatter than the main mix. Then I realised why -- because though I might have wanted to EQ a bass drum on the monitor mix, the limited EQ (high and low only) meant that I couldn't really do much to it. As soon as I got the mix up on the main channels and used the more powerful EQ, the sound actually suffered. At the time, I thought that professional mixes sounded better because of the more sophisticated equipment used, but that really isn't the case. As it turns out, some records can actually sound better when made on cheaper equipment."

Did you use any external EQ on the Doobie remix or did you rely entirely on the SSL EQ?

"We hired in some Focusrite EQ, which adds a nice warmth to vocals, and I'm also a fan of Massemburg EQ. And, talking of outboard, we also used a Neve compressor. We compressed the whole mix, but you have to have the compressor in as you mix rather than applying it after the mix is set up, because adding overall compression changes the apparent balance of certain instruments. The drum loop was a classic example of this -- without the compressor in, it was inaudible, but as soon the compressor went in, it brought it up -- so you have to set your balance with the compressor in.

"I'm a producer who has come from being a songwriter, I'm not a DJ..."

"The lead vocals were treated with both the Neve compressor and the SSL desk's compressor. When you think that more compression will probably be applied at the cut, and again when the music is played on the radio, you realise that it's compression all the way down the line really.

"SSL compressors were used over the stereo rhythm guitars and a dbx 160 compressor was used on the bass. The bass drum had some SSL compression on it, and we also tried gating it, but it lost all its power, because a lot of the low-frequency energy was in the decay of the sound, which was being cut off. In the end, we didn't gate it, so there's probably a bit of noise in there, but you can't hear it in the mix."

How did you get rhythm guitar part to work? Did you sample the whole two-bar sequence or did you have to sample individual chords?

"That was actually quite complex, because of the timing of the original, but it worked out well, and I think it's fair to say that the record now has a tightness about it. When I started, I thought it would be enough to take the two-bar riff, but I ran into the classic problem of the human-played instrument not locking in with the metronomic tempo of the drum machine. I ended up having to first pick the two-bar section that sounded tightest, and then subdivide that further. Also, there are two guitar parts, and I tried to find two bars where I could cut them in different places so that as one part was being retriggered, the other part was playing to cover the join.

"Timing is something that I am particularly meticulous about, and I often used to advance bass sounds to compensate for triggering delays or for resonant filter settings that slowed the sound down. I used to put an FSK code on tape and then bring it back through a digital delay so that I could adjust the timing of the sync'ed sound relative to what was on tape. These days, I use Notator with Unitor [SMPTE sync unit], and of course Notator allows you to adjust the timing of each track independently. I don't think the type of sequencer you use is important, so long as it allows you to do what you want to do, and Notator just happens to be what I've grown up with. The arrange mode on Cubase looks more flexible, but it's down to what you feel comfortable with."

How do you approach quantising when creating dance mixes? Is everything quantised to hard 16s?

"I like to use a slight degree of swing, something like setting 16A, rather than sticking rigidly to 16s, which can sound a bit wooden. The only thing I like to have a lot of swing on is the tambourine, because that's how humans tend to play them. Sometimes I even sample a two-bar loop of somebody actually playing one, just to retain the feel. It's all very personal, but I like it to be very loose, and I'll often use a very heavy swing on a sequenced tambourine, for example 16C, which, on its own, may sound completely out, but when the rest of the mix is in, it works."

"The Doobie mix was very much a hands-on mix, because even though we used the SSL desk, we didn't use the automation at all. All the sounds were run from the MIDI sequencer -- there was no multitrack recorder involved -- and the flanging was just brought in and out on a fader. The drum and bass sounds tend to remain constant, but I wanted to put a synth sound in along with the rhythm guitars just to help because, on their own, the guitars didn't sound all that convincing. The synth was a Juno 106, and the mix between the synth and guitar was critical, and required some riding during the mix.

"There's also a section of retriggered sampled vocal, which was used to create a solo section after the middle eight. We were struggling a bit, because there was no solo on the original, so we experimented with the vocal phrase 'Get it on', which was part of a line from the verse, but it was in the wrong key, so it had to be pitched up. Ideas like that often take quite a bit of time to set up, but they're usually worth it."

One of the things I've noticed about your earlier mixes is the way you manage to set the vocals really quite far back in the track, yet they're still clearly audible.

"I think it's very important in dance music to get things sitting right in the track. There used to be a time when you'd mix records with the vocals right up front, and then you'd do a different club mix with less vocal and more bottom end. Now what's tending to happen is that the radio version is really an edit of the 12-inch. It's important to use the right EQ on the vocal, and mix the backing track so that it leaves space for the vocal. That way you don't have to crank the level up too much, though you do have to pay special attention to the mid-range EQ. If the cutting engineer has to add a lot of mid, then your mix isn't that great. Another mistake people make is over-equalising the hi-hats, because that won't cut onto vinyl, and vinyl is still the medium for dance records. The Doobie's vocal is rather louder than I'd normally mix a vocal, but I think it works because of the type of song and because of where the vocal came from. You can't bury a vocal like that, but you still have to get it sitting right in the track. That's the difference between DJs making records using only samplers, and doing songs, where you get to mix all the various elements."

So far it seems that you've done all this using relatively unsophisticated equipment, but did you resort to hard disk editing at any stage?

"The song was mixed as a radio version, which is the opposite to what usually takes place. The 12-inch was put together using Pro Tools; I didn't want to be limited to pure stereo editing, which is why we didn't use Sound Tools. What I like to do is superimpose bits of vocals or other musical sections which I have on tracks 3 and 4 while the main stereo mix is on tracks 1 and 2. It helps you avoid obvious joins, and gives you more flexibility. We simply ran off lots of sections, including remixed sections from the original multitrack, and tried to see what fitted best. You have to start off with a lot more material than you think you'll need, because when you come to mix, a lot of what you have doesn't really work. There are also a couple of stops where the two sides pan, and during the break, there's absolute silence. And breaks are very important in this type of music; you have to take people on a musical journey with ups and downs -- you have to tell them a story. I'm a producer who has come from being a songwriter, I'm not a DJ, and I don't lose sight of the melody, because I think that is the most important element of the song. I love the rhythm element of dance music, but it's interesting that when you're working with people who don't have a musical background, they often spend hours trying to perfect something that is, in all honesty, largely irrelevant to the song. Even with dance music, what really gets people's attention is a melody or a hook, and regardless of technology, that will never change."

TRIGGERED GATING

"I like to use triggered gating on pad parts, because you still keep the impression of a pad, yet as you're gating it from a rhythmic source, it helps reinforce the rhythm as well. I normally program a dedicated rhythm to trigger the gate, rather than keying it from a hi-hat or whatever. It can be triggered from an output on the TR909 or from one of the outputs on the sampler -- it doesn't really matter. It's a brilliant effect, and I'm surprised more people don't use it.

"Here's a tip for anyone who has a Roland JP-4, a synth with a built-in arpeggiator: if you feed in an external trigger and press the Poly button at the same time as the Arpeggiator Up button, any chords you play will be played back as block chords gated by the external trigger, not as a normal arpeggio. Of course, with MIDI synths, you can also fake this kind of effect by creating rhythmic bursts of Controller 7 (Volume) information in your sequencer."

EAR CANDY -- SONIC TRICKS TO MAINTAIN INTEREST

All producers use certain little sonic tricks to keep their listeners' interest, and Steve calls these 'Ear Candy' -- "stuff that isn't critical to the record overall, but it adds interest." On some of his early mixes, there are things like reversed samples in breaks taking the place of drum fills, but this is just one of the many devices he employs to keep listeners interested.

"I like the reverse effect, but I don't over-use it. It's a nice alternative to a drum fill, and I'll usually create a reversed sample rather than running a tape backwards. Triggered gating [see box] is also a nice effect, as are things like 3D sound processors, though you have to be careful with some of these, because they always don't work in mono. To be honest, I'm not entirely convinced by any of the currently available 3D processes. It's very important to check mixes for mono compatibility, not only because of the uncertain nature of club PA systems, but also because of the number of people still listening in mono.

"I like panning as an effect, but it is important to get the timing right, otherwise you can leave the listener feeling that something isn't working. On the Doobie mix, panning delays were created using a straight delay on one side and triplets on the other. This was used on the Juno synth sound, where we had 360ms on one side and 240ms on the other, with just a little feedback, so as not to clutter the mix. Delays are very important in dance music, possibly more so than reverb, because you can get all the depth without the clutter.

'For reverb, I love the Lexicon 224 -- for me, Lexicon has what I call the sound of money. The PCM70 also has that same character. A good reverb has a body to it that you just can't get from a cheap reverb. Andrew is a great fan of the Concert Wave program with all the modulation stuff taken out."

At this point I asked Steve if he'd considered the creative potential of Vocoders or Lexicon's resonant chord delay programs, as these would seem to have all the hallmarks of 'ear candy'.

"Vocoders tend to come in and out of fashion, but you have to be careful when using them, because the effect is generally very obvious -- very 'in your face'. There are less obvious ways to use them, and I'd like to put some time aside to experiment with those. As to the resonant chord programs, those are potentially useful, though I haven't used them on anything yet.

"The flanging you can hear in the mix of 'Listen to the Music' was recreated using an old MXR analogue flanger, which was brought up on a separate fader so we could bring it in and out in the mix. It worked nicely, and I suppose the flanging was one of the three elements that was retained from the original song, the others being the guitars and the vocals."

Home | Search | News | Current Issue | Digital Editions | Articles | Forum | Subscribe | Shop | Readers Ads

Advertise | Information | Links | Privacy Policy | Support