In their eagerly‑awaited new flagship synth Roland have harked back to the sampling technology of their respected S‑series samplers, as well as adding some contemporary touches to equip it for the 21st century. Gordon Reid assesses past, present and future.
'Begats' are a bit like buses — they always come in bunches. Take the development of Roland's S&S synthesizers as an example. First there was the T110, but you won't have heard of that one, as it was never released. However, the T110 became the U110, and this begat the popular U220 and its keyboard counterpart, the U20. The U20 then begat the U70 — or it would have done if Roland's marketing gurus hadn't decided to cash in on the success of the D50 and call it the D70 instead. The D70 then begat the JD800, the first incarnation of Roland's 'four tones to a voice' synthesis system, which has since become the trademark of the company's sound engines. The JD800 in turn begat the JD990 module, and then the JV80, the first in Roland's most successful family of synths to date. TheJV80 begat a plethora of JV keyboards and, more importantly, the JV880 module, which in turn begat the JV1080, the JV2080, and the JV1010. Recently, the JV2080 begat the XV3080 and the XV88 keyboard version. There's been an awful lot of begetting going on.
If you trace the developments of Roland sound engines from the T110 onwards, you'll see that each has led logically and smoothly to the next, culminating in the XV3080. But hang on... why haven't I included the XV5080 in my story? Surely this is simply a sexier version of the XV3080? Well, yes it is, but it's somewhat more than that. Unlike its predecessors, the XV5080 has both a father and a mother, and it's the first Roland synthesizer in more than a decade to incorporate traditional sampling technology. It's not a sampler, but it owes its existence to both the '90s JV sound engines and the S‑series samplers of the late '80s. Intrigued? Then read on...
The XV5080 is Roland's new flagship. Surprisingly supplied only in 2U rackmount format (there's no keyboard version), it's based on the XV synthesis engine already discussed in our XV3080 review (see the July 2000 issue). This means that, like its little brother, it offers a substantial 128‑voice polyphony, GM Level 2 compatibility and advanced multi‑effects, and it can host Roland's new SRX expansion boards.
Unlike older Roland synths, the heart of the XV engine uses stereo PCM waveforms and, with well over 1000 multisamples, the XV5080 offers more than four times the ROM of the previous 'JV' models. This includes all the JV2080 waveforms, many waves from the respected JD990 module, and still more from Roland's extensive sample libraries. With huge editing potential and literally hundreds of Patches, Performances, and Rhythm Kits taking advantage of these waveforms, it's going to take you (and me) quite some time to fathom the depths of this synthesizer.
Hundreds? Well, there are seven preset groups of 128 Patches, a group of 256 GM Patches, 14 preset kits of 64 percussion instruments, nine GM kits of 64 instruments, and 64 preset Performances. Add to this the 128 slots for user Patches, 64 slots for user Performances, four slots for user Rhythm Sets, and the (up to) eight banks of Patches and Performances that appear when you install expansion cards... I'm sure you get the picture.
In addition to all of the above (which already exceeds the capabilities of the XV3080 by a considerable margin), the XV5080 also offers extensive sample replay options. You can install up to 128Mb RAM, and use this to load Roland S‑series samples, Akai S1000/3000 samples, WAVs and AIFFs. But don't for a moment think that the XV5080 is just an XV3080 with more memories and a bit of RAM thrown in for good measure. It's much more than that, as we shall see.
The synthesizer engine within the XV5080 is in most ways identical to that of the XV3080, but with a number of important additions, which we'll come to presently. Externally, there are a number of more obvious improvements, the best of which is the tasty 320 x 80‑pixel screen. This makes all the synth's facilities more accessible and more pleasant to work with, removing the user‑unfriendly barrier imposed by the XV3080's 2x40‑character display. Indeed, the front panel on the XV5080 is altogether better designed and laid out, although it lacks a numeric keypad — a real shortcoming when you have thousands of sounds and parameters from which to choose.
Other physical differences between the two machines become apparent on the XV5080's rear panel, which has two additional analogue outputs, plus digital S/PDIF and optical outputs. There are also two MIDI Ins, a SCSI port, a word clock input, and Roland's proprietary R‑Bus interface for direct connection to its V‑series digital mixers.
Less obviously, as they're tucked away behind the large panel on the upper surface, the XV5080 also boasts eight expansion‑board sockets (as opposed to the XV3080's six), thus increasing the synth's maximum ROM capacity considerably. There are also two slots for sample RAM, and these accept 16Mb, 32Mb and 64Mb SIMMs. Mind you, not just any old SIMM will do. These must be 60µS, with a board height not exceeding 36mm. Anything else may either (i) not work, or (ii) prevent you from refitting the cover. A local computer store quoted me £70, £160 and £230 (respectively) for these, so although you can probably obtain them cheaper elsewhere you should be aware that expanding the XV5080 into a sample‑playback machine is not a cheap prospect.
Before we move onto the sample section, it's worth noting the handful of significant differences between the synthesizer engines of the XV3080 and the XV5080. One of these positively leaps at you: thanks to its dual MIDI inputs, the XV5080 is 32‑part multitimbral. Clearly, Roland are making an effort to make the XV5080 an 'every synth you need' sort of product, and the ability to access 32 independent sounds makes a great stride in this direction. Or, at least, it would with a better effects structure. (We'll come to this later...)
The inclusion of Roland's COSM guitar effects processor is another big step forward. These physically modelled effects, first seen in the V‑Guitar, make the guitar simulations in the XV5080 far more authentic than those of any previous Roland synthesizer or sampler. If you're looking for a quick reason to justify upgrading from a JV1080 or JV2080, I would suggest, "because the XV5080 has COSM". If you're trying to create a 'band in a box', you need convincing guitars and basses, and this synth generates them.
Sticking with the effects for a moment longer, the XV5080 also boasts more effects slots — three MFX (multi‑effects) slots, compared to the 3080's one, just as the JV2080 had three EFX slots to the JV1080's one. Indeed, parallels between the XV5080/XV3080 and the JV2080/JV1080 are quite extensive, both of the bigger brothers featuring more effects, a better screen, more internal ROM, and more expansion slots.
Returning to the matter at hand, however, everything else about the XV5080's synthesis is identical to the XV3080's, so we can now move on to...
I know I'm repeating myself, but let's be clear about this: the XV5080 is not a sampler. Like the Roland SP700, it's a sample‑playback machine capable of loading and using Roland S770, S750 and S760 data. However, unlike previous RolandP, the XV will also load Akai data, WAVs and AIFFs.
Since there's only one way to test how well it does all of this, I located the SCSI port, hooked up a CD‑ROM drive that normally hangs off the back of an S770, and tested a variety of S‑series CD‑ROMs. These included Spectrasonics' Symphony Of Voices, Roland's own Keyboards of the '60s and '70s Volume 2, and two CDs of sounds and projects burned from a 500Mb HD using Toast software on a Macintosh computer. The operating system looked very S770‑ish, so I located the Patches I wanted and... everything loaded perfectly! All the samples appeared, the XV5080 allocated them to their Partials and then into appropriate Patches, and my S770's Performances appeared with all their parameters correctly defined. Much encouraged, I disconnected the hard drive from my S770, hooked this up to the XV5080, and tried again. As before, everything loaded as it would on the S770 itself.
Once loaded, the S770's Patches and Performances park themselves in the appropriate number of user memories rather than any form of buffer. Consequently, anyone wanting to load S770 data into an XV5080 had better be careful that they leave these memories free of any essential sounds or, at the very least, that they've backed up any important material. As you would expect, these Patches and Performances are retained in memory when you switch off the XV5080, but the sample data itself is lost. That seems a bit stupid to me. Since you're never going to want to load the 'wrong' samples into existing structures, why not put the whole shebang into sample RAM, and leave the internal memories alone?
Moving on, the data are in, and we can now ask what's missing from the XV5080, aside from the S‑series' sample inputs. The answer is: virtually everything to do with sampling and sample manipulation. None of the S‑series facilities, such as waveform compression and expansion, time‑stretching, resampling, wave filtering, smoothing and normalising, are implemented. Of course, this means that there's no need for a waveform display, so this also is missing, as is the S‑series' video monitor output. The only 'sample‑level' facilities to have survived the purge are the abilities to set start, loop, and end points, and to define loop mode — forward, fwd+release, one‑shot, fwd+one‑shot, alternating, reverse one‑shot, and reverse looping.
OK — now we know what has been lost. But has anything been gained? Absolutely! You can now throw the enormous power of the XV5080 synthesis engine at your samples. To help you understand how you can do this, I should first explain the two sample playback modes that the XV5080 offers. The first of these is multi‑partial mode. If you ask the XV5080 to load data from any S‑series disk, it auto‑selects this mode, whereupon a number of XV facilities disappear, to be replaced by a menu structure that (notwithstanding the omissions) looks identical to that of the S770 itself. However, if you select 4‑part (XV) mode you can access all the synthesis capabilities of the XV's engine, and apply these to each of the samples independently. This, of course, means that you can use your samples in two distinct fashions. Firstly, you can treat samples and multisample sets as components within an S‑series Performance, just as you would on an S‑series sampler. Alternatively, you can treat any single sample (but not a multisample set) as a source PCM within the XV5080 synthesis engine.
At first sight this makes the XV5080 seem somewhat schizophrenic. However, if you've used the S‑series and JVs in the past you'll appreciate just how cleverly Roland have integrated the two domains of synthesis and sampling — not just by placing the two alongside each other in one box, but by allowing them to combine complementary capabilities in a simple and flexible way. One consequence of this is that you always have access to 128 voices — whether these are all PCM‑based, all sample‑based, or anywhere in between. It's a far cry from the S‑series' 24‑voice architecture. The only obvious deficiency in this area is the XV5080's inability to treat an S‑series multisample set as one of its own PCM multisamples within the '4‑part' mode.
Now for the acid test: do the same Samples, Patches, and Performances loaded into both the S770 and the XV5080 sound the same? I plugged the master outputs from both devices into adjacent stereo channels on my desk, flattened all the EQ, defeated all the aux sends, and balanced the gains so that both instruments achieved the same loudness. I then used the channel mutes to perform instantaneous A/B tests on a variety of samples, including harps, pianos, orchestral strings, Mellotrons and assorted percussion.
Do they sound the same? I think that the differences (if any) are extremely subtle. I tried to detect quantifiable changes in filter tracking, keyboard response, or even a consistent bias in warmth or brightness towards one or the other, but I was unable to do so. On some Patches I very slightly preferred the sound of the S770, while on others I felt that the XV5080 had a tiny edge. Even in my own studio I don't think that I could have identified one from the other with 100 percent success. It would be a far more critical user than I who would love one of these instruments and refuse to use the other.
Since Roland claim that the XV5080 will also read and understand Akai S1000 and S3000 samples, I had a snoop around my studio for an Akai CD‑ROM. Unfortunately, being a Roland sort of guy, I have only one Akai‑format CD. Placing this in the CD‑ROM drive and selecting the Load page caused the XV5080 to display the partitioned Akai disk structure rather than the Roland sampler structure. It then allowed me to select desired programs or samples and load them without further ado. Once loaded, the Akai samples and programs appeared in the same multisample structure as the Roland samples but (and I was very impressed by this) all the Akai's filtering and tuning information was retained and interpreted correctly.
Further encouraged, I then tried a CD of WAV and AIFF files. At first I thought that the XV5080 was going to trip over because, on selecting Load, I was presented with an 'Improper Disk' message. However, within a few moments the XV5080 recognised the CD and allowed me to load the WAVs and AIFFs; either individually, in user‑defined groups, or all together (up to the limit of the installed RAM, of course). My only gripe here is the lack of an on‑screen progress meter. It can take minutes to load larger clips and, while there's a 'barber's pole' graphic to show that something is happening, there's nothing more informative. Shame.
At this point I was a little concerned about how I would use these WAVs and AIFFs within the XV5080. Then I discovered the 'Create Patch' utility, which allows you to choose one or more samples and then automatically create a multi‑partial Patch from it or them. The XV places the resulting Patch in the 'Temp' buffer, thus allowing you to edit it further before saving it to a User memory. The utility does nothing that you couldn't do manually, but it does it almost instantaneously, and takes all the pain away — very nice.
Such facilities also make it simple to use the XV as a straightforward replay device for extended samples (rather than the snippets from which you create Patches). For example, many bands use samplers not as sound sources for playing 'live', but to replay complete musical sections derived from their — or other peoples' — recorded tracks. The XV5080 allows you to load several such sections, and use them much as you would use backing tracks replayed from tape or disc. You may or may not approve of such tactics but, either way, the Roland gives you the option.
Finally, the time had come to save my new Patches. Oops... a problem! Try as I might, I could not get the XV5080 to recognise that there was free space on my S770's hard disk. Whether the source was originally Roland, Akai, WAV or AIFF, the result was always the same. Does this mean that the XV5080 will refuse to save to any S‑series disk? Yes, it does. The disk formats and data structures of the two instruments are completely different. You can save an XV's memories plus samples to an appropriately formatted disk and re‑load them into the same machine (or, of course, into another XV5080) but the S‑series is out of bounds. This is a shame, because it precludes you from building and configuring libraries on an XV, and then taking them elsewhere for use on the older machines.
Unfortunately, there are one or two other XV OS deficiencies to mention. For example, if you use a SmartMedia data card as your backup medium, you can only save the entire User area as a single entity; you cannot save individual Patches or Performances! This means that you could waste an entire 128Mb card to store just a handful of Patches.
The situation with external hard drives and Zips is a little better, but not much. If you use these, you can save individual Performances as 'folders' (which is more efficient) but you still can't save individual Patches. This makes it difficult to locate and load disparate Patches into new Performances unless you return to the original CD‑ROMs or S‑series drives. (I get the impression that even the person who wrote the manual was embarrassed about this, because it is very woolly regarding all these issues.)
Possibly the most infuriating news is that you can't quickly and easily load two Performances, mix and match the sample‑based Patches you want from within them, and then save the new, composite Performance to disk. To do this, you must load the first Performance, then manually delete the data in a second Performance slot, manually delete the required number of Patches for the second Performance and, finally, Append the second Performance, its Patches, and the Sample data into the user memory. Once you have done this you can choose which Patches appear in any given Performance, and normal service is resumed.
I can understand this shortcoming to some extent. After all, the XV5080 isn't psychic and, because it stores Performance and Patch data in non‑volatile memory, it needs to ask you what is expendable. Indeed, the above procedure might make some sort of sense, but for two facts. Firstly, it's not necessary to clear any Performances or Patches when you load the first Performance. Secondly, you don't have to jump through these hoops when you load and overwrite individual Patches. Finally, my (albeit volatile) S770 has no trouble loading multiple Performances and keeping track of all the appropriate Patches and Samples.
These faults are hard to forgive, bearing in mind that one of the best things about the later S‑series instruments was their intuitive and quick sample management and disk operating system. Despite its excellence in other areas, the XV5080 has taken a step backwards here, falling into traps similar to those that previously crippled the Kurzweil K2000 (now cured) and the Korg Triton (still waiting for surgery). Power‑users beware...
To err is human, so they say. To demand more, even when you've already been offered so much, well... that's human too. Presented with an instrument as powerful as the XV5080, I can still find areas I'd like to see improved. For example, the routing within Roland's effects section is fairly flexible, but it's not a patch (sorry!) on Korg's. Look at it like this: the 1995‑vintage Trinity is 16‑part multitimbral and offers eight independent effects busses. The new XV5080 is 32‑part multitimbral, and has three busses. Oops! Indeed, given the Roland's sophistication in other areas, it would be great if every part could have an independent effects buss. Going still further, I look forward to the day when all Patches retain all of their effects when imported into a Performance. If the Novation Supernovas and Novas can handle this, I don't see why a flagship Roland should not.
Secondly, I think that the XV5080 should sport a floppy drive. Not, you understand, that I am wedded to vintage technology, but because many people — myself included — have large sample libraries of DSDD 3.5‑inch disks. The whole S550 library came like this, and it would be useful to be able to load it directly into the XV. While we're talking about sampling, Mr Serious User of Woking wrote in saying "Dear SOS, Why oh why oh why oh why oh why doesn't the XV5080 have a sampling input? Surely it wouldn't have cost too much to add one?" Well, Mr User, all I can say is this: If this omission isn't a hint that there's a real Roland sampler in the works, somebody should take the company's President outside and give him a jolly good talking to.
What else? How about an external input to take advantage of the SRV and COSM algorithms? (Hmm... I think I know the answer to that one. The XV's buss structure isn't flexible enough — though, in my view, that's no excuse.) And what about adding a numeric keypad for entering Patch and parameter values? The sheer number of possibilities in the XV5080 would seem to make this a necessity, but it isn't there. Consequently (and this is also true of the XV3080) you'll soon be sick of the rotary encoder. Finally, it strikes me as very strange that the XV3080 displays waveform names in the edit palette, whereas the XV5080 only offers a number. This may sound trivial, but when you're scrolling through 1000+ waves to find the sawtooth it's a right royal pain in the posterior.
If you're in the market for a new rack synth, there are obviously two members of the XV family to choose from. I've explained what I would do, and why, in the 'Vive Les Differences' box, but to be brief, unless you really do have no way to raise the extra cash, I think you should do yourself a favour and ignore the XV3080. Provided that Roland sort out the sample management and disk operating systems, the 5080 is the way to go.
You might think that (as usual) I've found a lot to criticise about this instrument and, of course, you would be right. So let's put one thing into clear perspective: despite some significant faults, I think that the Roland XV5080 is a fantastic synthesizer. It sounds wonderful — clear and precise, or warm and thick, as you choose — and if you have one, you'll never run out of new sounds. Furthermore, if you are able to control it from a poly‑aftertouch keyboard, you could soon be playing with more expression and/or imitative realism than ever before.
The XV3080/XV88 and XV5080 share many capabilities, but big brother adds a number of important facilities not found on the lesser module or its keyboard sibling. The following table outlines these:
|<div align="left"> Parts:</div>||16||32|
|<div align="left"> Maximum Waveform Memory:</div>||256Mb*||386Mb*|
|<div align="left"> Effects Algorithms:</div>|| SRV‑3030|
| SRV3030, V‑Studio|
|<div align="left"> Expansion Board Slots:</div>||6||8|
|<div align="left"> Sample Playback</div>||No||Yes|
|<div align="left"> Sample Memory</div>||n/a||128Mb|
|<div align="left"> Sample Formats:</div>||n/a||Roland S‑series, Akai S1000, WAV, AIFF|
|<div align="left"> Display:</div>||2‑line||320x80 pixel|
|<div align="left"> Analogue Outputs:</div>||6||8|
|<div align="left"> Digital Outputs:</div>||n/a|| 2 Stereo (Coaxial |
|<div align="left"> MIDI Connectors:</div>||In, Thru, Out||In1, In2, Thru, Out|
|<div align="left"> SCSI Port</div>||no||Yes|
|<div align="left"> R‑BUS 8‑Channel Digital Output:</div>||No||Yes|
* If converted into 16‑bit linear format
Given the differences between the XV5080 and the XV3080, should you be thinking of finding the cash for big brother? I'm going to turn this question on its head. Given these differences, why would you seriously consider buying the XV3080? Its shortcomings may not seem of overriding significance, but in my view they render it significantly less attractive. Take the XV5080's larger screen as an example. Why struggle with the XV3080's 2‑line display when a full graphic LCD is available? Indeed, why deny yourself the myriad possibilities of the XV5080's sample playback capabilities, even if you don't intend to use them immediately? Why suffer the lack of the authentic guitars afforded by COSM? Why sacrifice the additional outputs (always very welcome) and the possibility to hook the thing directly to one of Roland's V‑wotsits?
It's also a real shame that you can't upgrade any part of the XV3080 to XV5080 specification. There's an extra price to be paid for the XV5080 — £500 — but I think it's too small a difference to justify the dissimilarities between the two machines. As Derek Johnson and Debbie Poyser pointed out in their XV3080 review in SOS July 2000, the Proteus 2000 is in many ways comparable to the 3080, and in some ways superior, but costs just £749.
Roland came late to the sample game. When, in 1986, it finally joined in, its offerings were meagre, to say the least. While Akai and Emu were streaking ahead with the S900 (1986) and the Emax (1987), establishing a dominance in professional studios that has never since been broken, Roland brought out the S10 and S50 keyboard samplers, plus the pitiful S110 and S220 modules. Of these, only the S50 used 3.5‑inch diskettes, and the other three were duly consigned to the 2.8‑inch Quickdisk dustbin of history.
In 1988, Roland released the S550 (a 2U rackmount version of the S50 with twice the sample memory of its predecessor) and its little brother, the S330. But it was not until the launch of the S770 in 1990 that the company truly entered the professional sampling arena. By this time, Akai had streaked ahead with the S1000 (1988) and the S1100 (1990), so the S770 was never going to challenge Akai's dominance. Nevertheless, Roland followed it up with the S750 (1991), the SP700 Sample Playback module (1993) and, in 1994, the S760. Excellent instruments, all of these, they offered users two benefits not found elsewhere: intuitive operating systems and unique on‑screen editing capabilities. On the other hand, they were constrained by their 24‑note polyphony and limited sample RAM. (The S760 and SP700 were the biggest, with just 32Mb RAM). Consequently, Roland's samplers were overshadowed by the likes of the Akai S3000 (1993) and S3000XL (1995), Emu's E4 (1995) and, more recently, the Akai S5000 and S6000.
But maybe 2000 will be the first year in which Roland takes the lead in the world of sampling. With the VP9000's unique Variphrase technology, and the replay flexibility of the XV5080, the company just needs a true sampler to replace the ageing S‑series...
There's no doubt about it — the sound quality of the XV5080 is exceptional, especially for imitative and orchestral sounds. Unfortunately, it's very difficult to translate these qualities into text. If I try to describe what makes the XV5080 special, the best words I can find are "clarity" and "openness".
Part of this impression comes from the stereo PCMs. These produce a more open soundfield than you'll ever achieve by passing mono samples through a stereo reverb, no matter how good that reverb may be. But, notwithstanding that, the XV5080 still has an indefinable "who took the cotton wool away from the speakers?" quality. Furthermore, minimal aliasing suggests some careful engineering in the digital‑to‑analogue conversion. OK... there's some aliasing, but it doesn't grab you by the throat, unlike some other contemporary synthesizers that I'm too polite to name.
Sounds that rely upon clear, accurate high‑frequency content benefit greatly. These include bells, mallets, struck strings and, for that matter, any other sounds naturally produced by hitting lumps of metal and wood with other lumps of metal and wood. I also like the harps very much, but perhaps the greatest beneficiaries of the XV's qualities are the acoustic guitars. I'd go so far as to say that I have never before heard such playable emulations, whether produced by a synthesizer or a sampler. I'm particularly enamoured of the guitar Patches that offer slides, plucked notes, and harmonics, with the nature of the note determined by velocity. Given a bit of practice, you can make these sound really real. (Great sentiment, lousy grammar.)
Then there are the basses, which are greatly enhanced by the XV's clarity — no muffled "whoompfs" here. The transients are all present and correct, and very nice they are, too. That's not to say that the XV is incapable of thick, smothering warmth. Try some of the strings and pads, and they'll show off this side of the instrument's character, too. And the horns? Ooh... the horns! Unmissable.
If there's just one deficiency (and there always is), it's in typical analogue synth sounds and effects. These simply don't have the complexity, the depth, the je ne sais quoi of the real McCoy. But hey... why would you choose to buy a sophisticated samploid synth if you wanted to create another bunch of unexceptional Minimoog and SH101 emulations? That would not be clever.
The bottom line? I really want Roland to sort out the problems I've reported elsewhere in this review. The XV5080 sounds superb and, because of this, I want no justifiable reasons why I shouldn't buy one.
- Powerful, high‑quality synthesis.
- Sample‑replay capabilities.
- Generous display.
- Decent multitimbrality.
- Eight expansion slots.
- Good‑quality effects, including COSM‑modelled guitar effects.
- Eight analogue outputs.
- Digital I/O
- Sample management, 'Load' and 'Save' algorithms need a serious overhaul.
- Supplied without any sample RAM.
- Sample Patches overwrite existing XV Patches rather than loading to sample RAM.
- Limited effects busses by modern standards.
- No external input to take advantage of the filters, envelopes and effects.
- No numeric keypad.
- No floppy drive.
- Manual very poor in places.
The XV5080 is Roland's most powerful and best‑sounding digital synthesizer to date. Add some of the numerous JV expansion cards or the newer SRX cards and — with the exception of true analogue sounds — it can cover almost all of your sound‑generating requirements. If a future OS update sorts out the faults in the sample management and disk operating systems, it could become hard to resist.