Magazine Archive

Home -> Magazines -> Issues -> Articles in this issue -> View

How To Recreate A Jam & Lewis Production

'When I Think Of You' by Janet Jackson

One good means of learning your way around the equipment you own is to use it to recreate somebody else's music. That's exactly what Mike Collins did last year whilst working at Yamaha's Research & Development Studio in London. Using nothing but Yamaha gear, he set about reproducing a Janet Jackson hit in full, by himself. Here, he recalls what was involved.

One good means of learning your way around the equipment you own is to use it to recreate somebody else's music. That's exactly what Mike Collins did last year whilst working at Yamaha's Research & Development Studio in London. Using nothing but Yamaha gear, he set about reproducing a Janet Jackson hit in full, by himself. Here, he recalls what was involved.

Jimmy Jam and Terry Lewis are two of the 'hottest' producers in the Pop and Dance charts today. They use plenty of the latest MIDI technology and instruments in their work, and so I decided to recreate one of their hit records as a demonstration recording in the Yamaha R & D Studio last year. My purpose was to demonstrate how Yamaha equipment could be used in practice on this type of recording, and I also undertook the project in order to get the studio 'up and running' by producing a recording just as it would be done in a commercial studio, to make sure that the newly-installed R & D studio was fully operational.

Before I was employed by Yamaha to help plan and install their MIDI recording studio, I had worked on TV sessions for the previous two years. These sessions involved recreating 'hit' records by visiting American artists for British TV shows such as Top Of The Pops and Solid Soul. So, it was logical for me to choose a similar type of project to check out Yamaha's new R & D studio. The particular song I chose was 'When I Think Of You' by Janet Jackson, which many people will be familiar with. Here's how I went about it...


The first step was to listen to the original recording to identify the instruments used, and to decide which Yamaha instruments to use to recreate the track. The most distinctive sounds on this record were the prominent bass synth and electric piano figures which formed the backbone of the track. Immediately, I recognised these as DX-type sounds and, as soon as I started to listen through my DX5 factory programs, I found that Performance Memory 4-5 contained an Electric Bass/Electric Piano set-up which was identical to that used on the record!

The next most distinctive sound was that of the repetitive 'brass stabs' heard throughout the choruses and elsewhere on the track. I had just started using the RX5 drum machine at this time, and this contained a sampled TX816 brass sound which, with a bit of editing on the RX5, again sounded similar to the one on the record. I also decided that the RX5 would provide reasonably close approximations to the drum sounds used on the record, and that a TX81Z would provide a reasonable marimba-type sound for one of the other parts in the verses.

Near the end of the record, after the breakdown section, there is a great 'explosion' sound, followed by a quick harp-like glissando to lead in to the end choruses. For the explosion, I chose to use one patch from the DX5 and one from the TX81Z, to make a more interesting combination of sounds. Performance Memory 4-6 in the DX5 factory programs contained a sound called 'Double Harps', which proved to be very suitable for the harp gliss. There was a 'picking' guitar rhythm part also, which I chose to play on a Yamaha SG2000 guitar fed through a Yamaha guitar combo amplifier, and there were a couple of additional percussion instruments (such as a Quihada) which weren't available on the RX5, so I also played these 'live' on the re-recording.

The next step was to transcribe the arrangement from the record. I have developed a personal method for doing this which works well in practice. The first thing I do is to get some blank manuscript paper and write barlines corresponding to each bar of the music on the record. Instead of using repeat signs and so on, I always write out every bar in full, because the arrangement will often change in repeated sections, sometimes quite subtly, and writing them out in full makes everything more explicit and easier to follow. For similar reasons, I usually put a maximum of four bars to a line and normally have four bars on every line, unless there are odd numbers of bars involved somewhere. I write bar numbers at the start of each line, or even above every bar for extra clarity, and label different sections, such as verse, chorus, and so on.


If the drum pattern is very simple, as it was in this case, I programme it into the drum machine straight away, and mark the drum machine pattern numbers (and song steps, if a song structure is used in the drum machine) onto a copy of the music manuscript. If more complex drum parts are involved, I may write these out in full on the manuscript, and then programme them into the drum machine or into a MIDI sequencer, in step-time if necessary.

The next most obvious step is to record the bass line into a sequencer, using similar methods. I chose to use a QX5 sequencer, and as the bass line was quite repetitive but with some interesting improvised variations, I learned this part 'by ear' from listening to the record. I then practiced the part until I could record it into the QX5 in real time, without using quantisation, retaining as much spontaneity as possible and adding one or two improvised variations of my own around the basic pattern but without losing the 'feel' of the original.

With the bass and drums in place, the next most obvious part to duplicate was the electric piano. I thought that this would be the simplest of the lot, but came up against a small problem when trying to identify the actual chord inversions used on the 16-bar intro section of the song. I had quickly worked out that the main chords used in the choruses were A Major9 and E7sus4. The actual inversions used were four-note chords situated around the C above middle C on the keyboard, and the actual notes in use (reading from the bottom note up) were G#, B, C#, and E for the A Maj9, and A, B, D and E for the E7sus4 chord. The problem was this: the chords sounded subtly different in the intro section, and although I could tell they were different, I could not at first identify what this difference was.

After much pondering, and asking two or three other musicians for their opinions, I realised that the chords used in the intro only employed three notes each. This made them sound a little sparser and starker but more distinctive as a consequence, and thus more suitable for the song intro where they would be heard more clearly! Nice one Jam and Lewis! The note missing from the first chord was the G#, which simplified the chord to A9; the note missing from the second chord was the D, which simplified this chord to Esus4. Of course the bass line provided the root notes for these chords, leaving plenty of musical space in-between and giving the track a light and 'airy' feel.

The marimba part heard in the verses was very simple and repetitive, and consequently easy to programme straight into the sequencer. The harp glissando was similarly easy to play with one finger on the 'white' keys, and the explosion was also easy to slot in, as it only occurred once. I inserted Program Changes into the QX5 sequencer to switch the DX5 and TX81Z to the explosion sounds, and then to switch the DX5 to the harp sound, and to switch them both back to Piano/Bass and Marimba for the end choruses. These Program Changes had to be positioned carefully in the sequences, just a few sequence 'clocks' before the new sounds were needed, so that the synthesizers would have sufficient time to respond by switching programs.

The guitar part was a muted picking style, which is often used on this type of record, and the 'live' percussion parts were very straightforward also. I wrote the percussion parts out on the manuscript, as this was easier than memorising where they came in, and then I sight read them in the studio while recording 'live' to tape. As it happens, I started my musical career as a guitarist, so it was fun to find an opportunity to pick up the instrument again, and I have always enjoyed playing percussion - ever since my Dad gave me some maraccas at the age of about five!

The final element to the song was the 'brass stabs', which were subtly different each time they came up. This made me think that they had obviously been improvised on the original recording, and they played a prominent role in defining the recognisable sound of the record. So, I wrote these parts out in full on the manuscript and then programmed them into a second RX5 drum machine which was available in the studio. Using Multi-Voice mode, I assigned different pitches to the different pads on the RX5, to cover the range of notes played by the brass.

Using Mark Of The Unicorn's Professional Composer program for the Macintosh computer, I have written out two bars of the music showing the three-note electric piano chords used on the intro, and two bars showing the four-note chords used later in the song. Underneath, I have included four bars of the bass line (ie. two 2-bar sections) showing the basic pattern first, followed by one of the variations. I have also written out eight bars (bars 5-12) showing one section of the marimba part and one section of the brass stabs beneath, though these actually occur in different places in the song.

Screen dump from Performer showing the song and track structure.

To illustrate the layout of the sequencer tracks and to show the structure of the song, I have included a screen dump from Mark Of The Unicorn's Performer sequencer, which uses the Markers window (top) to name the different sections and show bar lengths and section timings, and the Tracks window (bottom) to specify the instruments and the MIDI channels used. Performer is very good at this sort of thing (visual displays, etc), but I chose to use the QX5 sequencer for speed and convenience when I originally re-recorded the piece. This was partly because I was using Yamaha sequencers almost exclusively at that time, and could work at a faster pace with equipment which I had reason to use virtually every day of the week, and partly to demonstrate that it was quite easy to sequence a complete recording of this nature using Yamaha's humble QX5!


Once the sequences were prepared, which took about a day including the transcription, I was ready to record the backing track in the Eastlake-designed Yamaha R & D studio using the custom-built Yamaha RM3000 40:24 recording console, an Otari MTR90 24-track analogue tape machine, Dolby A noise reduction, Yamaha monitor amplifiers, and Quested monitors. As all the sequences were prepared in advance, this took about three or four hours in the studio one morning, and in the afternoon I recorded the guitar, percussion, and vocals. I often work with songwriters and vocalists doing demos for music publishers and record companies, and remembered a promising young singer called Shara who sounded very like Janet Jackson. I called her up, and luckily she was free for an hour that afternoon - in between doing her shopping or something - and she agreed to come in and sing the song.

I had the lyrics ready when she came and a copy of the original song on cassette to play to her while she rehearsed in the studio, and, amazingly, she learned the whole song and recorded both lead and backing vocals in just over an hour! The best bit of vocal had to be the crazy laugh, which came just after the explosion and harp gliss near the end of the song - Shara got this off to a 'T' first time!

Most of the sounds were recorded 'flat' to tape, or with a minimal amount of EQ, and they sounded very close to the original recording straight away. So all that was needed on the mix was a small amount of additional EQ'ing, level balancing, and some reverb, delay and compression effects. The most important effect to get right was a short 'slapback' echo on the snare with a reasonably short reverb, which gave the rhythm track its distinctive 'feel'. This was added using a Yamaha Rev1 digital reverb, and SPX90s were used on the other instruments where necessary. By the end of the two days the whole track had been recreated by one musician and one vocalist, taking advantage of a complete range of Yamaha musical equipment!


Everyone who has heard the finished result has commented on how close to the original it sounds, and on how quickly it was recreated compared with how long it must have taken to record the original. It would typically take three or four days or more to record a track like this, and would usually involve at least three to four musicians. However, you have to bear in mind that when Jam and Lewis recorded it, they had to devise the parts from scratch and experiment with the mix to come up with a hit recording like this. All I had to do was to copy what they had done as quickly and efficiently and as faithfully as possible!

Anyway, now you know how it was done! Or, at least, how to begin to recreate a track like this. Obviously, one of the secrets is in being able to identify the sounds of the different instruments being used on such a recording. I have found that, with a little practice, it is relatively easy to distinguish many of the synthesizer and drum machine sounds made by different manufacturers' equipment. You should be able to recognise an analogue synthesizer as opposed to an FM synthesizer without much trouble, and tell the difference between a Roland TR707 and a Yamaha RX5 or RX11 quite easily. Some sounds can be more difficult to recognise than others however, and, in some cases, reasonable simulations may be possible using a variety of different instruments.

I hope that this brief insight will be of practical use to those of you who may wish to follow in the footsteps of Jam and Lewis, or to anyone who needs to recreate records originally recorded by other musicians. Now that the Yamaha R & D studio is fully operational I have returned to freelance work, and I am hoping to get a call from CBS Records to work with other Jam and Lewis produced artists, such as Alexander O'Neal.


I often find that when playing the same chord on a different synthesizer patch, the chord sounds 'different' due to the different harmonics emphasised in the actual synthesizer patch. An acoustic piano or a Rhodes sounds different again, for the same reason. So, when I am transcribing parts, I usually have a keyboard set up with as close a sound as possible to the original instrument that played the part which I am transcribing, so that I can check the notes by actually playing them back on the keyboard.

It is very easy to hear what you think is a four- or five-note chord played on the electric piano, only to find on closer inspection that one of these notes is actually being played on another instrument, like the bass or the guitar or another synthesizer. You should bear in mind that on many modern arrangements the piano or chordal synth pad does not necessarily need to play the root notes of the chords, because the bass will probably be playing these. This differs from solo piano arrangements, where the piano obviously does need to play the root notes.

I have a Technics SL1200 turntable with a varispeed control which allows me to match the pitch of any gramophone record to concert pitch (A=440Hz). Often, record producers use tape varispeed on the final version of a single to increase the tempo a little to give it the 'edge' they think it needs.

Sometimes records end up in other than concert pitch because all the musicians have tuned their instruments differently. This typically happens if a guitarist tunes his instrument 'to itself', and not to A440, either through laziness or because he prefers the different string tension and sound that results. All the other musicians must then follow suit, which obviously rules out the use of acoustic pianos, Hammond organs, and similar instruments. Sometimes a pitch difference can occur when tapes are copied from machine to machine and one (or both) machine is running at a slightly different tape speed - this is less likely to be the cause now that digital recording is being used more extensively.

It is often useful to copy a record onto a cassette for convenience when going over and over particular sections. Also, a dual speed reel-to-reel tape recorder can be useful as well, because you can play back fast pieces at half tempo by dropping the tape speed. This makes the faster phrases easier to transcribe, even if they are an octave lower in pitch. You could even record a whole phrase into a digital sampler if you have access to one, then loop it, and have it repeat continuously whilst you figure out what notes are being played.

Previous Article in this issue

Spectrum Synthesis

Next article in this issue

Making Tracks

Sound On Sound - Copyright: SOS Publications Ltd.
The contents of this magazine are re-published here with the kind permission of SOS Publications Ltd.


Sound On Sound - Sep 1988

Feature by Mike Collins

Previous article in this issue:

> Spectrum Synthesis

Next article in this issue:

> Making Tracks

Help Support The Things You Love

mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.

If you value this resource, you can support this project - it really helps!

Donations for July 2020
Issues donated this month: 0

New issues that have been donated or scanned for us this month.

Funds donated this month: £20.00

All donations and support are gratefully appreciated - thank you.

Please Contribute to mu:zines by supplying magazines, scanning or donating funds. Thanks!

Monetary donations go towards site running costs, and the occasional coffee for me if there's anything left over!

Small Print

Terms of usePrivacy