Small Box, Big Impact

Many people have been through a lot of gear changes and many people who have started exploring gear in recent times may find it hard to find that one piece of gear they really connect with.

Having so many choices can both be a blessing and a curse.

Those who have been swapping out gear for a while will also know that some devices were passed on with feelings of regret.

This happens to all of us.

You have this box that you really gel with and then you see an ad for a new product and you too hastily make a decision to sell your trusty workhorse, only realising later that you should have kept it and the new box wasn’t much of an improvement.

Luckily we can often repurchase it on an ever growing second hand market, but for some reason it feels like rekindling with an ex lover, it’s just not the same anymore.

The only thing that remains are the sweet memories.


I have those sweet memories of one particular piece of gear.

It wasn’t perhaps so much that particular device, but the time it entered my life.

But before this becomes too psychological, let’s discuss this particular device, as it wasn’t just important to me, but for the technological developments in general.


Of all the pieces of gear that I have ever owned, which is a large number and too many to list here, this one box always gives me the same warm feelings that you would get when you think back of your high-school flame.

That piece of gear was the Roland MT-32.

It’s perhaps also the most primitive box I have ever owned.

How could such a primitive device leave so many indelible memories?

Well, like I said, it’s all about Time & Place.



It was a cold winters day in February of 1988 and I almost didn’t get my MT-32 at Kick Music in Hilversum.

Many other customers had rung the shop that day, as it had just arrived, but I happened to get there first. 

This little box, as low-spect as it was, came as a huge liberation for many, especially for a young bloke like me who had little dosh to spend.

I still vividly remember the sheer bliss I felt from holding that box, thinking about all the musical ideas and applications I had in my head.

There wasn’t a more exciting feeling imaginable than to spend the cold and wet Dutch winter inside my room creating music.


Multi Timbre

Multitimbrality was a feature that had been around for some years, but the MT-32 definitely took it to a new realm.

Before the MT-32 most synths were either mono-timbral or bi-timbral.

When a synth is Mono-timbral it can simply only produce one sound at a time.

The DX-7 for example was mono-timbral, which meant that you could record a piano part into the sequencer and perhaps a bass part as well, but they would both play the same one sound on the DX.

If a synth was Bi-timbral, like the D-50, the engine could be split into two MIDI channels, both using a different sound, thus one track on the sequencer could play back the piano and the other track the bass sound.

If you had a Multi-timbral synth, you could have up to 6 or 8 different sounds assigned to different sequencer tracks, thus  allowing you to produce an entire arrangement.

Because each sound had its own MIDI-Channel, there was unlimited control over timbre and the MIDI editing of it. 


To get a better understanding of how the Multi-Timbral synth evolved, let’s have a brief look at what the options on the market were at that stage.


Fairlight CMI (1979)

The first ever multi-timbral instrument was the Fairlight CMI from 1979.

Holding 8 individual voice-cards, each card could have its own sampled sound loaded and then be recorded on separate tracks into the sequencer.

Typically this is what this would look like: 

Track-1 > Kick

Track-2 > Snare

Track-3 > Hihat

Track-4 > Cowbell

Track-5 > Toms

Track-6 > Bass

Track-7 > Flute

Track-8 > Marimba

A major limitation though was, that each sound could only be recorded monophonically.

If you wanted to record the Marimba polyphonically, it had to be loaded onto more tracks.

You would then record each note separately on its dedicated sequencer track.

Nevertheless, this instrument was a huge leap forward and in many respects way ahead of its time.

The reason why it wasn’t a commercial success was, that at the time of release the CMI was roughly priced the same as a house.


Kurzweil K-250 (1984)

This instrument was also a major league contestant, costing in the tens of thousands.

An extremely advanced machine, with many outstanding features.

Its 12-voices could be split into 6 parts for orchestration purposes.

The system could also sample at 16-Bit / 50kHz through the use of an Apple computer.

Price, size and availability made it accessible to the rich and famous only.


Sequential Six-trak (1984)

The Sequential Sixtrak from 1984 was one of the first affordable synth that could split its 6-voices and assign a different sound to each. 

These sounds could then be recorded into the internal sequencer.

The 6-voices could also be stacked. 

Sequential also made a MIDI interface for the Commodore-64 computer, including a software sequencer to work in conjunction with the Sixtrak and Drumtraks.

This way one would have a budget system similar to a Fairlight.


MKS-7 (1986)

Roland’s first Multitimbral synth, with built in drums.

Based on the Juno-106 sound engine, it had a bass, chord and lead section.

All together there were only 8-voices though.


Ensoniq ESQ-1 (1986)

This synth had almost reached workstation status, with 8-voices and a comprehensive onboard sequencer.

Unfortunately, there were no onboard drum sounds or effects.


Yamaha TX-81Z & FB-01 (1986)

These two boxes were quite popular and cheaply priced.

Again, the only problem was that they had 8-voices and no drum sounds, so one could not create entire arrangements with them.



By 1986 there were cheaper sampling units on the market that would also give you multitimbrality, like the Roland S-50 and Akai S-900, but the one main thing still remained an issue: Polyphony.


The ultimate goal of multitimbrality was to be able to create an entire piece of music using the one unit.

It needed to have a large array of different (sampled) instrument sounds, have a drum bank, on board effects and enough polyphony to produce an entire piece of music.


A new benchmark

Instead of the 8-voices that the previously mentioned synths had, the MT-32 had…… guessed it…….32-voices.

Well, they were actually 32-partials, not strictly voices.

Roland used the term Partials, because they didn’t quite work like oscillators.

Like on the D-50, there was a sample portion (start) and a sustained synthesized portion of the sound.

The D-50 had two sets of those, whereas the MT-32 only had one set.

If a sound was made up of 2-partials, it would use up 2-voices.

Meaning that if you played 3-notes, 6 of the 32-voices would be used up, leaving you with another 26.

Even though this still seems meagre by today’s standards, it was surprisingly sufficient for its day.

Voices were also dynamically allocated, meaning that whenever a sound was not being used, these voices would become available for another sound.

However, sounds with a longer release time, such as bells and cymbals, needed to be put in the lower parts.

For example, if you put a bell sound on part-8, the parts before it would get priority, thus possibly cutting off the sustain portion of the bell sound.

These were the type of things one had to be aware of back then, which is unthinkable of in this day and age with unlimited polyphony on VST’s.


The MT-32 was 9-part multitimbral, with one part being dedicated to the drums.

The way these parts were organised may seem a bit convoluted, but here is how it works:

Parts 1-5 were accessible through the front panel.

Parts 6, 7 & 8 had to be accessed via the computer, keyboard synth or hardware sequencer.

You would have to perform a program-select/change to assign sounds to parts 6, 7 & 8.

Each part had a fixed MIDI-channel, but…..

It didn’t start with MIDI-channel-1, it started with channel-2.


The only thing I can think of is that Roland considered that people using a keyboard synth would also want to put this in the MIDI chain.

Most budget synths would operate under MIDI channel-1, so it was convenient if the MIDI channels on the MT-32 were shifted up.

However clever, if you didn’t like that, you could easily set the MT-32 to start at MIDI channel-1, by holding the Master volume button and the Part-5 button.

This would however reset itself after the unit was turned off.



Then there was the Rhythm part, which held the Drum kit and was fixed to MIDI channel-10.

The drum bank in particular was a significant new step up in sound module design that quickly became the norm.

Instead of having a dedicated drummachine to supply the drum sounds, you now had them available in the same module.

This feature was not available yet on any other affordable synth, but it seemed to make more sense as there was a huge shift in the market towards software sequencing.



It’s become expected on synths these days to have built in effects.

The first synth to feature built-in effects was the D-50, so it was no surprise that the MT-32 featured reverb as well.

If you reckon the reverb on the D-50 was of inferior quality, wait till you hear the reverb on the MT-32.

Not only was it extremely poor sounding, the Dry/Wet signal could only be set from 0 to 5.

There were no size, Decay time or EQ settings, there was just one reverb type.

Whatever you set the reverb to would apply to the entire unit, so every sound would go through the same reverb.

The Bass and kick drum would then also go through the reverb.

There wasn’t an extra set of outputs on the unit either, so often you would just end up using the dry signal.


Sound editing

Editing the sounds was not possible from the unit itself, but if you had a sound-editor for your computer you could load different sounds into the MT-32 via system-exclusive (sysex). 

Again, once the unit had been switched off, it would return back to its factory settings.


Noisy bugger

If you have read my U-220 blog, you would have learned about the U-110 as well and how it had mad hiss and crackling issues.

The MT-32 rose to quick infamy for exactly the same reasons.

When the unit is stationary, a distinct background hiss can be heard that resembles the sound of an airplane in parking mode.

Add to that the anti-aliasing hiss and crackles of the 8-bit samples and “lo-fi” will get a whole new meaning.


Secret fame

Superseded by the more professional D-110 late 1988, the MT-32 had gained moderate success but rose to greater fame in an unexpected corner of the market: Computer-Games.

The music that was made for computer games was usually relying on the computer’s built-in 8-bit sound cards.

Some sound cards could only produce one note and some could do 3-notes.

Its size, MIDI connectivity to a PCI-interface also designed by Roland, MIDI mapping and the great array of sounds, were seen as a perfect match for Game music.

A massive list of IBM computer games were compatible with the MT-32, so it became the de-facto sound-card, even though it technically wasn’t a sound card, but MIDI was starting to open up more doors than expected.

The MIDI mapping structure of the MT-32 was really the precursor to another influential format that is still being used to this day.  


General MIDI

As all the sounds and the mapping of them on the MT-32 were factory set, it kind of unintentionally created a standard, as game music composers just had to work with it.

For this reason other manufacturers had to accommodate the gaming industry as well.

As newer modules were being released, deviating from this format would have caused gamers to hear the music differently on each unit.

Kawai had released the K1M synth module around 1989, which was also multi-timbral and comparable to the MT-32.

However, it had a completely different structure and set of sounds.

If a piece of game music was made on an MT-32, someone using a K1M would hear the music completely scrambled: drums playing the piano part, the bass playing the melody, etc…

Because the MT-32 had cemented itself firmly in the gaming world, its mapping format was largely used as an example for the development of General-MIDI.

This new format would force everyone to stick to the sounds that were programmed to a fixed mapping.

Each device’s acoustic guitar sample would be in the same place as another device’s, so when an acoustic guitar was used in a track, it would always match the acoustic guitar of another sound module.

This consequently lead to another new format, the MIDI-File, which we won’t get in to here.



Roland made a whole host of products that used the MT-32’s sound engine.

The MT-100 was a combination of the MT-32 and the PR-100 sequencer.

Then there was of course the slicker 19-inch rack version the D-110, on which more in a bit.

With the Gaming world embracing the MT-32 as their favourite sound processor, Roland were clever enough to design a series of devices for the computer market with the MT-32 built-in and later also adding sounds of the U-110.

This resulted in the CM series of devices, on which you can find more information if you do a search.

By 1988 Roland also tapped into the Arranger-Keyboard market, with the E-series.

These also had the MT-32 sound engine built in, albeit a slightly improved version of it.



As mentioned before, the MT-32 was superseded by the D-110, which became much more successful due to its professional features, such as being able to edit the sounds from the front panel, individual outputs on the back, more internal sounds, trendier drum samples, more effects, performance patches and a card-expansion slot.

The DAC’s were also improved.

It even had its own hardware programmer, the PG-10.

The D-110 arrived in the late summer of 1988.

I remember going to a Roland Benelux product presentation in The de Lamar theatre in Amsterdam, September 1988.

When they announced the D-110 it really got me spewing, but I also realised that I had been having a great time with the MT-32 in the previous 6-months. 

You can clearly hear the MT-32’s guts inside the D-110 and most sounds are even the same.


So why did I buy it?

At home we only had a Roland RD-300, which was my dad’s.

He had a DX-7 and Prophet-600 as well, which he used a lot for gigging.

Whenever I had the attic to myself, I would compose songs on the Atari, using the RD-300 and my TR-505.

Luckily the RD-300 was 16-voices, so I could create bass parts, chords and a melody using the same sound.

As soon as my dad came back from gigging, I would take the DX and Prophet upstairs and assign them to the tracks I had made the night before using the RD piano.

Then I would quickly record them to cassette tape before my dad needed his synths back.

Even when I had the two synths and the piano, I was still limited to 3-sounds, so I would perform a lot of program changes to switch between different sounds, creating the illusion that more synths were used.

The DX-7 had bass sounds that could also be used as clavinets or leads, so the creative mind was always challenged.

However, the MT-32 provided so much more flexibility. 


Final words

The MT-32 was like a poor sod’s Fairlight, without the sampling option and sequencer of course.

It really felt that way, because by then everyone was using an Atari-ST computer, including myself.

I was playing in a band and 4-guys in that band had an Atari.

Before the MT-32 there wasn’t really an affordable multitimbral device that featured tons of sounds, including a drum bank and a reverb effect.

Imagine the ripple effect this suddenly created, especially at that ridiculously affordable price, which in hindsight doesn’t seem so affordable at all, compared to todays offerings and prices, but back then it broke all price barriers. 

Even though I had set my sights on a D-50, when I found out about the MT-32 it was a no-brainer.


The MT-32 was an absolute blessing, arriving at exactly the right time in my life.

I could suddenly create new ideas and arrange entire tracks, which before the MT-32 required you to have many more synths and a larger mixer.

Of all the devices I have ever owned, my fondest memories are of the MT-32.

Because it suddenly gave me so many opportunities to create music, my creativity surged to new heights.

Making music suddenly became addictive.


During the second half of the 80’s there was a second coming of the Jazz-Fusion genre.

I had grown up listening to Fusion during the late 70’s, with records by Weather Report and The Jan Hammer Group, but as a teenager my ears veered more towards the slick pop tunes of the early-mid 80’s.

The music school I was in noticed the Fusion revival too and decided to put together a Fusion ensemble, with me on drums.

We continued on with that band for the next 5-years.

Fusion music gave me so much inspiration, that the advent of the MIDI studio gave me the chance to really explore Fusion.

Before the MIDI studio, trying to explain your musical ideas to other musicians was a difficult process.

I was now able to create the entire idea using MIDI and record it onto cassette tape, which everyone could then take home and listen and practice to.

The MT-32 played an important role in that process, having constant access to its many sounds.

This creative spurt unfortunately also had an adverse effect.

I inflicted my ideas at a rampant pace, which caused the band to pull the brakes.

This left me frustrated and consequently became one of the reasons for me to quit the band.

The MIDI studio became my new band and I abandoned live playing for many years.


The 80’s hedonistically marched on and gear was getting cheaper.

My MIDI skills and ear for detail had also developed and the MT-32 just didn’t quite fit the bill anymore.

The MT-32 was an amazing solution for a penny scrambling young bloke like me, but the sounds soon started to appear weaker than they first appeared.

By the end of 1989 I had acquired a DX-7IIfd.

A company called “Grey Matter” had made a memory expansion board for it that also featured multitimbrality.

I sold the MT-32 to fund the E-Board for the DX-7II.


I always secretly wish I had kept my MT-32.

They are somewhat hard to find these days, which is due to people rediscovering the old computer games again, for which they would need an MT-32.

10-years ago you’d find them for around 50-euros, but nowadays they can be 3-4 times that price.

Somehow devices reappear again after many years of obsolescence.

I had never thought that the MT-32 would get a second life, as it was such a lo-fi, low-budget piece of gear.

It is therefor rather exceptional that such a low quality device had such a significant impact on many levels.

Hence this blog, as it will hopefully shine a different light on things.

The MT-32 deserves more respect and credit.

It was ahead of the curve in more than a few ways.

«   »

Add comment


Nick W
a year ago

Nice to read about this Gee! The Casio CZ synths were multitimbral, but in multitimbral mode each of the four timbres became monophonic. I'm assuming the MT-32 is a precursor of the D20 (which was multitimbral too)?