Looking for an effective way to expand your palette of sounds? How does a 31-voice, multitimbral, 1U-high, rack-mount PCM expander sound?

LOOKING TO INCORPORATE sampled sounds into your music but unwilling to create your own samples? Put off by the disk loading times of samplers? Fancy the idea of a collection of preset sampled sounds which includes acoustic and electric pianos, vibraphones, bells, ensemble strings, choirs, acoustic and electric guitars, organs, saxes and various types of bass? If the answer to any of these question is yes, Roland may have an answer for you, in the shape of their U110 PCM Sound Module. If the answer is no, perhaps you ought to go make the tea.

You see, the U110 (introduced at the '88 BMF as the T110, you may recall) doesn't have any sampling capacity of its own. Nor does it have any RAM memory for samples, which in turn means you can't load in samples via MIDI or from disk. In fact, the U110 doesn't even have a disk drive. Wot, no disk drive?

Let me explain. The U110 has 99 ROM based samples (called Tones) stored within its slender 19", 1U-high frame, and can access further samples off Roland PCM ROM cards. You can plug up to four of these cards at a time into slots on the U110's front panel; if this sounds familiar, perhaps it's because Korg's DRM1 drum expander adopted the same approach.

You don't actually have to load card-based samples into the U110. All you do is select a card sample and it's instantly available on the keyboard, just as if you'd selected a patch on a synth (well, it should be; despite many phone calls to Roland UK on my part, and many faxes to Roland Japan on their part, we weren't able to get hold of any cards in time for this review).

Sample editing in the familiar sense doesn't exist on the U110. In fact, there's not too much you can do with the samples except play then,. Unlike Korg's M1, which allows you to use in samples as material for sound synthesis, the U110 occupies sample territory exclusively.

So you get the general idea. No tedious disk-loading, samples instantly available on power-up, no struggling to find decent loop-points - in short, all the benefits of sampled sounds without the hassle normally associated with making and using them. The U110 also has two other features in its favour: it's six-part multitimbral and - good news for home and professional recordists alike - it has six polyphonic individual audio outs in addition to the usual stereo pair.

Special mention should go to the four onboard demo sequences which are accessed via the ROM Play mode (fast becoming a regular feature of Roland's synths and samplers). Not only do they very effectively show off the capability of the U110, they're enjoyable pieces in their own right (particularly 'T-Jazz #1', an excellent recreation of bop-era jazz).

Samples

APPARENTLY THE U110'S samples are derived from Roland's S50 sample library, which means 12 bit sample resolution and a sample rate of 30kHz or 15kHz. This does mean in practice that they're not altogether free of noise, but its rarely obtrusive (anyway, why should sounds always have to be squeaky-clean?).

In addition the samples have been "compacted" by, for instance, being given shorter loops. And I'd say Roland are using sample interpolation, as introduced on their S330 sampler, as a substitute for multisampling - another useful way of cutting down on the amount of sample data, even if the results are sometimes a bit odd (as with the shakuhachi, whose characteristic "pitch warble" attack is much slowed in low registers).

The 99 internal Tones are divided into Piano, Vibraphone, Bell, Marimba, Guitar, Bass, Choir, Strings, Wind and Drum groups. As another means of saving on sample memory, there aren't actually 99 different samples - within most groups, single-sample Tones have also been doubled and detuned to provide further Tones.

"The U110 is essentially a preset instrument, but the ability to combine, detune, and put sounds through digital chorus gives considerable flexibility."

The acoustic pianos are especially good, though their tone tends to thin out on sustained notes - a consequence of looping compromises. You get mellow and bright versions, and a great honky-tonk piano, while the electric pianos include both bright, hard-edged and warm, chorused versions. The vibes and marimbas are suitably chunky but not particularly warm, while the bells are thin-sounding rather than deep and resonant.

The basses consist of slapped, fingered, picked, fretless, synth and acoustic varieties. I particularly like the warm, rounded acoustic bass and the slurred fretless, while the plentiful slapped variations are suitably snappy and funky. Among the brass, particularly good is the soft trumpet, which captures the characteristic attack very effectively, while the saxes are fairly realistic (again, the attack is good) if a bit limited in variety and character.

The skinbashers' section is taken care of by 37 drum and percussion samples, which have been mapped across a five octave note range and collectively assigned to internal Tone 99. These are standard fare: several bass and snre drums; low, middle and high toms; crash, ride and china cymbals; open and closed hi-hats; handclaps, cowbell and cabasas. In general they've been kept as short as possible (the cymbals are looped for length), which does mean they don't always "breathe" as they should. On the other hand, they're gritty and punchy, and make for a good modern-sounding rhythm section.

Personally I could have done with a few less slapped bass and organ samples. But, all in all, I'd say Roland have come up with a healthy range of basic, familiar sounds. It's a shame that effective use hasn't been made of velocity so that you could introduce, say, fingerboard slap into a plucked acoustic bass performance, or overblown notes into a sax solo (in other words, something more akin to the discontinuity of sound that you can get out of some acoustic instruments). Velocity-switching is "built into" a few selected Tones (it's not under your control), but hasn't really been taken lull advantage of.

Hopefully the PCM cards will expand on the internal sounds by being more adventurous, rather than just providing more variations on familiar themes. We'll just have to wait and see.

Organisation

YOU CALL UP the U110's samples by selecting from 64 onboard Patches. A Patch consists of six Parts, each of which is assigned one internal or one card Tone. A Part (rather than the sample assigned to it) has associated programmable output assignment, MIDI channel, note range, level, pitch and LFO parameters. By assigning all six Parts to the same MIDI channel and giving them separate or overlapping note ranges you can create all manner of keyboard textures, while assigning separate MIDI channels allows up to six instruments to be played independently from a sequencer. Of course, you can also create any combination of these two approaches (say, split double-bass, piano and sax played together on one channel, drums on another channel and strings on a third). If you don't want some Parts to be active within a Patch, all you have to do is set their output assignment to "off".

Each Part within a Patch on be assigned its own volume level and velocity response scale together with amplitude envelope attack and release rates (one of the few ways of altering the character of a sample on the U110). Additionally you can transpose a Part +/- one octave in semitone steps, fine tune its pitch, and set a pitch-bend range and detune depth. Independent LFO control for each Part allows you to specify rate together with automatic depth, delay time and rise time or manual depth and rise time (rise time controls how quickly the LFO effect moves from zero to maximum). Manual control can come from the mod wheel, polyphonic aftertouch or channel aftertouch. Polyphonic aftertouch can also control pitch, while channel aftertouch can control volume (so owners of channel-aftertouch keyboard instruments can't control pitch-bend from the keyboard - surely a rather silly limitation).

Patches can be bulk dumped via MIDI SysEx, taking only a few seconds to transfer. In this way you'll be able to build up a Patch library on external disks quite easily.

Programming the U110 is accomplished from a mere six buttons on the front panel. I found that, with only a little practice, I could move around the instrument's parameters extremely quickly. If you want even more speed, you can program and activate jumps to any four LED pages of your choosing.

"All too often, an instrument's multitimbral capabilityis limited by its polyphony - thus it's good to see that the U110 gives you 31-voice polyphony."

Master tune, MIDI control channel and patch Maps (see below) together with MIDI control change, patch change, channel aftertouch, polyphonic aftertouch, pitch-bend and SysEx on/off are all specified in what's known as a Setup. There's only one Setup memory, but you can transfer Setups via MIDI SysEx, making it possible to build up a library using external software. The U110 can respond to pitch-bend, modulation, volume and sustain pedal independently for each Part (if, of course, the Parts are on different MIDI channels). This would be particularly useful for the Maps, as you'll soon realise.

One area into which Roland have clearly put some thought is that of calling up the U110's sounds via MIDI. As well as being able to select Patches via MIDI (up to six Tones at once) you can select Tones for individual Parts independently.

A MIDI patch number received on the U110's control channel (which can be any one of channels 1-16) calls up one of the 64 Patches. In this case a MIDI patch number calls up the equivalent U110 Patch number. But a MIDI patch number received on a channel assigned to one of the Parts will select a Tone for that Part (as long as the Part channel doesn't coincide with the control channel). More than this, the U110 allows you to create six patch Maps, each of which allows you to "map" U110 Tones (internal and card) onto incoming MIDI patch numbers. For instance, MIDI patch 43 could call up Tone 11 on PCM Card three. Each U110 Part can have one of the six Maps assigned to it, so that effectively you can call up any internal or card Tone via MIDI's 128 patch numbers. This approach also makes it possible to call up layered sounds on the U110 from a single MIDI patch change.

The U110 places no limitation on the combination of internal and card samples, and you can use up to six samples off any one card (though as mentioned earlier, I wasn't able to try out the card aspect of the U110). One puzzling aspect of the U110's Tone selection procedure is why it allows you to select from 31 cards on its Tone Select and Map Setup pages, when only four cards at a time can be plugged in.

All too often, an instrument's multitimbral capability can be limited to some extent by its polyphony - particularly if you like using thick chordal textures. Thus it's good to see that the U110 gives you not eight, not 16, but 31-voice polyphony. Er, 31? Roland Japan's official explanation of this (literally) odd number is that the 32nd voice is needed for the U110's ROM Play facility. But of coarse you'd realised that already, hadn't you?

In practice, the U110's polyphony depends on whether you're using Tones which require one or two voices (these are split 50:50 on the internal Tones). The two-voice Tones are dual, detuned and velocity-mixed samples (settings are determined by Roland rather than you).

Output Assignment

THIS IS ONE area of the U110 which is bound to cause some confusion, not least because the manual isn't very helpful (it's about time Roland Japan stopped providing their customers with poorly-translated versions of what the Japanese think a manual ought to be).

The U110's 31 voices are not dynamically assigned across the six individual outs. You assign each Part to one of these outputs, as you would expect, but the number of voices available for each Part is determined by which output mode you've selected. There are 50 of these output modes to choose from, each one allocating the U110's 31 voices to the outputs in a different way (Roland use the term "voice group" to refer to an individual output). Output assignments and output mode are both programmable per Patch.

For example, output mode one allocates all 31 voices to output one, meaning that no voices can be sent from the other outputs even if you've assigned Parts to them. Mode nine allocates 15:12:4 voices to outputs 1-3 respectively. If you want to use all six outputs then you're going to have a spread like mode 20, which gives you 7:8:4:4:4:4 voices. The greatest number of voices are invariably concentrated in outputs 1-3, so it makes sense to assign chordal parts to these outputs and solo parts to the higher outputs. But seeing as the least number of voices you can select for a Part is three, if one Part is a monophonic bassline and another Part in a monophonic lead line, you're wasting four voices (two if the Tones are two-voice).

"There are no less than 50 output modes to choose from, each allocating the U110's 31 voices to the outputs in a different way."

Output modes 21-50 treat outputs one and two as the same "voice group", and the number of voices allocated to them as a pair is shared dynamically between them. However, even if you assign one Part to output one and another to output two, they will both be sent from both outputs.

Half of modes 21-50 have been given a stereo audio output assignment (L and R) for outputs one and two, which allows Parts sent via these outputs to be effected by onboard digital chorus and tremolo. Rote and depth can be programmed for each effect, with settings storable for each Patch rather than for each output mode (if you select an "active effect" output mode for a Patch, its effect settings will come into play).

You can assign up to six Parts to any one audio output it you want, in which case that output's voice allocation will be shared dynamically by all the Parts - but that does rather negate the purpose of having individual outputs in the first place.

What all this means is that you have to put some thought into output assignment. Although Roland's approach has the incidental effect of acting as an automatic voice reserve (one Part can't steal voices from another), overall it's a real nuisance when you just want to get on with the business of making music. It would've been far better to have dynamic voice assignment with the option to reserve voices for any Part, but presumably Roland have good technical reasons for not implementing this approach.

Verdict

THE U110 SHOULD be a big seller for Roland, combining as it does operational convenience and a respectable (and expandable) range of quality instrumental sounds with a flexible multitimbral organisation and the polyphony and outputs to make the most of that organisation - all within a 1U-high casing.

Editing the U110 is a non-tiring experience. With the notable exception of output assignment, Roland have made the U110 as straightforward as possible, while the logical parameter organisation and easy front-panel access ensure that you can zip around the instrument, no problem.

Although the U110 is essentially a preset instrument, the ability to combine sounds, detune them against one another, alter their attack and release times, and put them through digital chorus and tremolo does give considerable flexibility, as does the ability to EQ and effect sounds individually via the separate audio outs. The PCM cards should go some way towards alleviating any preset blues you might have - by giving you more presets to play with.

In conclusion, the U110 is essentially a utilitarian rather than an innovative produce, and as such it succeeds very well in the task it sets out to accomplish.

Prices U110 £599; PCM ROM cards £45 each All prices include VAT.

(Contact Details)