A rack-mount version of the EPS was something Ensoniq swore they would never produce, but they have finally bowed to public pressure. Paul Wiffen looks at the extra memory, extra outputs and SCSI applications, and concludes it has been worth the wait.
Ensoniq's stated philosophy behind the EPS (reviewed SOS Feb '88) was to produce a sampler that was geared towards performance (something that previous samplers had been lamentably poor at). The poly pressure keyboard, the Play While Load function, and the performance presets are all features which make that keyboard a joy to use live instead of the nightmare that most samplers can be in that rather exposed situation.
At first glance it would seem that most of these features would be lost in the rack-mount version. The poly pressure keyboard, of course, is no longer there, and although the EPSm can respond to polyphonic aftertouch, you can't take advantage of this unless you already have a keyboard which transmits it (which makes SQ80 owners the most likely customer for the EPSm, unless there are any Prophet T8 or Yamaha DX1 owners still out there). Play While Load is most easily managed when you have the buttons to control it right above the keyboard you are actually playing. The nice thing about the EPS's Patch Select buttons was that they were right next to the pitch and mod wheels at the performance end of the keyboard. Even if you have the rack-mount EPS within reach, it is unlikely that you will be able to use its Patch Select buttons so effectively (even though they perform exactly the same function on the rack). So you can understand why Ensoniq's original policy was that the EPS really would not translate very well to a rack-mount format.
However, the demand for a rack version has been so great that Ensoniq have knuckled down to making it as close as possible in use to the keyboard: Poly pressure can be received if you have the keyboard (or sequencer) to transmit it; Play While Load can be initiated from MIDI program changes, so you can call up new samples from a master keyboard; and the Patch Select can be controlled just as effectively (in fact more so, as both hands remain free for playing) from a dual footswitch accessory. Thus you can operate the rack version in a way that would make it just as effective as the keyboard.
All this supposes, however, that the EPSm is just an EPS in a rack. If that were the case, then it probably would be more deserving of a paragraph or two in the news pages than the space taken up by a Paul Wiffen review (notorious as they are in the industry for being a shade on the long side!). But the fact is that when you buy an EPSm, you get a lot more than you get with an EPS. This is just as well really, because the EPSm does cost substantially more than its keyboard counterpart. However, it only takes the briefest examination of the EPSm's spec to work out where this extra money is going.
The biggest expense in any sampler these days is the memory. Ever since the chip crisis of last year, contrary to expectation, the price of DRAM (dynamic random access memory) has not fallen in line with that of other electronic components. Therefore, any sampler with a generous amount of on-board memory is going to be more expensive. The EPSm comes with a full megaword of memory - four times as much as the keyboard version. So, whilst a natural comparison (memory-wise) for the EPS would be the Emax or S950, the only sampler on the market that comes with the same sort of memory size as the EPSm is the Akai S1000, which the EPSm substantially undercuts in price. Of course, you could buy a four times memory expander for the EPS keyboard, but it does add well over £500 to the price (simply because memory costs that much). So the bulk of the extra price of the EPSm can be accounted for by its increased memory. Indeed, it actually comes out looking like very good value in comparison with the S1000, which has exactly the same sample time available at the industry standard sample rate of 44.1 kHz.
Extra memory can be used in various ways. If it is entirely given over to reproducing one instrument, then much higher sound quality is certain to result. This is because you can have more multi-samples and/or each sample can sound for longer before it needs to be looped. Experiments I conducted with piano, string and guitar sounds soon convinced me that this was just as true in the case of the EPSm as with any other sampler. Alternatively, you can continue to use the typical size of sound that Ensoniq have been creating with the much smaller basic memory of the EPS in mind, and you suddenly find that you can hold quite a few of them in memory at the same time - to the point where you run out of instrument locations before you run out of sample memory. Indeed, on several occasions I found myself wishing that there were more than the permitted eight instruments available. I soon discovered how to use the patches available within presets to get around this (by copying a whole instrument into each of the layers available within an instrument location and then saving the whole thing as a multi-instrument), but it cuts down on the flexibility and speed of use, which are the trademarks of the EPS/EPSm. I would have preferred to see 16 instruments available. Of course, a second bank of eight instruments accessed through some software switch may still be possible in a future software update, and this would give you the added advantage of being able to sequence 16 different sounds from the internal sequencer's tracks, or respond to all 16 channels of incoming MIDI in Ensoniq's excellent Multi mode. Judging by Ensoniq's well-deserved reputation for responding to user wishes (to which the very existence of the EPSm bears witness), if enough people ask for it this might happen.
Another major advantage of the increased memory size I noticed was that I didn't need to 'Select Instrument To Delete' (a message which unexpanded EPS owners will be only too familiar with!) anything like as often when I was loading a new sound to use. This is one of the things which makes remote disk loading via MIDI program changes difficult to use (you can't select the instrument to delete via MIDI program change even if you know that this message is liable to appear), so it's good to see that this is less of a problem on the rack version, where you're more likely to want to use this facility. Of course, if you're careful you can always cause a MIDI load to occur over the top of a larger instrument you no longer need, by placing the EPSm in Multi mode and sending the Program Change on the MIDI channel corresponding to that instrument (which removes the problem altogether). In fact, I found that I was using the rack in Multi mode so much that I saved the global parameters with this selected, so that it booted up in this mode every time.
Memory size isn't the only area where the EPSm justifies its price tag; any sampler which falls into the professional category needs to offer separate audio outputs. Whilst there is an optional expander for the dual output EPS to give eight individual outs on a separate box, the EPSm comes with these fitted as standard. Nor are they your standard monophonic outputs, which are really only any good for drum sounds and bass lines; these are fully polyphonic outputs and the EPSm can dynamically allocate its 20 voices between them without any manual configuring. Apart from the EPSm, only the Akai S1000 and Roland S550 samplers have this facility. Even expensive systems like the Fairlight Series III and the Emulator III are still stuck with monophonic individual outputs. Score a major point for Ensoniq.
For me, the real revelation was the innocent looking 25-pin connector on the back of the EPSm - the SCSI (pronounced 'scuzzy') port. True, this is nothing new, in that there is already a SCSI interface for the EPS, but until now they have been like gold dust. Making features like SCSI an option is often not a good idea, as the options are rarely available when a machine is first launched, and when they are there never seem to be enough to go around. There are many unhappy S1000 owners out there right now, as Akai's promises of a SCSI update for their machine have still to be fulfilled.
This is the first time I have ever had the opportunity to use an instrument with SCSI (I had no hard disk when I got the Emulator III for review, and at that point in time it wasn't compatible with Blank Software's Alchemy sample editing program). The problem is that normally when an instrument is brand new (ie. when a reviewer gets to play with it), there is no software available to interface through the SCSI option. However, because the EPSm behaves exactly the same way as the EPS, and because the SCSI is standard, Alchemy can be used with it from the outset. Undertaking sample transfers via SCSI instead of via MIDI is a real revelation; samples are sent in less time than it takes to play them back, instead of hundreds of times more. This means that you can contemplate doing a quick edit to a sound on the computer (provided you keep it hooked up) rather than wondering to yourself if its really worth that wait while the sample data crawls across from one to the other.
I also used the EPSm with a 45 megabyte 'removable media' SCSI drive that I borrowed from Argent's in London, and found it to be the most wonderful thing imaginable. Of course, there is nothing to stop you from using a standard Macintosh hard disk with SCSI (I had already seen the EPS working with a Nexys hard disk and found the performance disappointing), but these removable drives are a much better idea. The data access time is as fast as a hard disk and the storage size as big, but when you have filled it up you only have to buy another cartridge at £150, instead of another hard disk at £600-700. Plus, you can use the drive with several different devices (eg. an EPS/EPSm, a Mac, the new Atari ST, which is supposed to have SCSI, etc) by changing the cartridge, rather than having to use a different hard disk for each device. This is possible because each cartridge can be used in a different format [further details from Argent's (Contact Details)],
In conjunction with an EPS/EPSm, the load times are incredibly quick. A full floppy disk's worth of data - which normally takes 55 seconds to load - loads in less than three seconds, and the complete megaword of memory in the EPSm loaded in under eight seconds! Typical EPS files load so fast that you can hardly see the load message in the display before it disappears. This is the first time I've seen a sampler able to switch between lots of different sounds just like a synth.
But my favourite thing of all is being able to use MIDI Program Changes to load sounds from SCSI. This means that you can use an external sequencer to load new sounds in the middle of a piece of music. As an example, this powerful feature would allow you to sample vocals in 10-second segments, save these to a SCSI storage device, and then use a sequencer to load them just before they are needed, and to send a MIDI Note-On to trigger them on cue. I even managed to have the EPSm playing the musical backing as well, but it occasionally missed some Note-Ons as the Program Change was received. In truth, it's probably a bit much to expect one instrument to do everything including vocals, but with care you can get more out of the EPSm than any other sampler I've ever tried. It could certainly handle the backing vocals for an entire gig, for example, without breaking into a sweat.
Other changes from the keyboard version include a headphone jack on the front panel - how I wish my S1000 had such a sensible location - I get sick of having to climb around the back of my rack to plug them in. This is great for when you don't want to bother connecting the EPSm to an amp, but unfortunately does not cut the sound to the main outputs, so you can't use it to check everything is working fine before you play it to 30,000 people - unless you can kill the sound externally on a mixing desk fader.
It goes without saying that we have not even touched on the excellent sequencing and sample editing features of the EPSm, as these were covered in depth in our EPS review [SOS Feb '88]. What we have covered here is all the additional things which come as standard on the rack, and which you can buy as accessories for the existing EPS keyboard. But you shouldn't underestimate the up-to-date nature of the EPS operating system, which includes digital signal processing features normally only found on sample editing software for computers - such as Sample Rate Conversion and instantly undo-able Crossfade Looping Macros, which normally take ages to set up. Although the EPSm's sequencer doesn't really have the flexibility of the top hardware sequencers that are currently available, it is nice to have sequencing built-in. I certainly preferred it to anything currently available on a computer, with the possible exception of C-Lab's excellent Notator. I particularly like the ability to choose between the old or new version of anything you do. I am forever thinking that I could play a part I have just recorded a little bit better next time, and then finding out that I couldn't when it's normally too late (like when I've recorded over the previous performance). With the EPSm's sequencer, I found I could always get the old version back. My only real criticism is that each track is tied to an instrument - you can't record a second track using the same sound without copying that instrument to another instrument location (and these are already a bit thin on the ground considering the amount of memory available). But this is only a small niggle in an otherwise excellent onboard facility.
Overall, the EPSm is a serious professional piece of gear, with the amount of memory, separate polyphonic outputs and SCSI interfacing facilities which top-end users and recording studios need. I am pleased that Ensoniq have chosen to release the EPSm as a fully-expanded version of the EPS. This means that the full potential of the design is much more likely to be seen in music stores, as the EPSm comes straight out of the box ready for the most advanced applications. But what is equally good is that an EPS owner seeing the system working at full stretch in a store, with high speed SCSI loading of one megaword sample files being played through 10 fully polyphonic outputs, need not feel ripped off that he bought an EPS a year ago, but pleased that he can now update the keyboard he has used for a year to exactly the same spec, at around the same sort of total cost that the EPSm now retails for. If only all manufacturers followed this creed!
£2199 inc VAT.
Ensoniq GB, (Contact Details).
Review by Paul Wiffen
Previous article in this issue:
Next article in this issue: