Paul Wiffen takes an in-depth look at Casio's flagship samplers

Casio are well known as a company for making technology available at previously unheard of prices, and when the FZ1 was first announced, a 16-bit sampler with a megabyte of memory for under 2 grand was a pretty radical innovation. At the time most 12-bit samplers with half a meg were costing at least that (most of them more) and even today there isn't a cheaper stand alone 16-bit machine on the market.

This was followed by the FZ10M rack mount version with double the memory as standard, making the FZ the most cost effective sampler on the market. Add to this the imminent arrival of the FZ20M with the same amount of memory, plus the very worthwhile SCSI interface for mass storage and the FZ family keeps abreast of the latest advances in sampling.

The FZ1 was also the first sampler to have a big enough display with the resolution to show useful sample waveform displays. Up until then, to see your sample data you had to either connect a computer monitor (if you had a Roland) or transfer the sample data to a computer software editing package. This worked OK with the Roland's (as long as you didn't mind the expense of an extra monitor), but as for transferring the sample data, the time taken to send it across and get it back (this was before the days of the faster than a speeding bullet SCSI interface) rather dulled one's enthusiasm for seeing the data, let alone performing any edits. With the FZ1, you made a sample and the pictorial representation was instantly available. In fact, it is not only sample data that can be shown. At many stages of the FZ1's operation you can press the Display button and the machine will convert the information shown numerically into a graphic display. This is particularly useful for enveloping filter and amplifier levels in addition to all the editing, looping, truncating and combining of waveforms which one would normally expect to use visual editing for.

Casio bill the FZ1 as a digital sampling synthesizer and whilst I normally object to samplers being described as synths, I'm prepared to make an exception for the FZ1, as it features many ways of creating sounds even if you never used any samples. It allows for subtractive (analogue-style) synthesis using standard waveforms and filter and amplifier envelope shaping, additive synthesis using sinewave harmonic addition, and waveform drawing as alternatives to sampling. But as the authenticity of sample sounds is the thing which will attract most people to the FZ1 let's look at that first.

Sampling is the first choice in the Source Select menu and once you have named the voice and decided where to put the original pitch on the keyboard, you get the usual options, sample rate and length, and input level. Sample rates available on the FZ1 are 36kHz, 18kHz and 9kHz. Now 36kHz is a good general purpose sample rate, but many samplers can sample at higher rates, in the middle or high 40s, which whilst not necessary for the majority of sounds, are useful for "difficult" sounds like cymbals, hi-hats or other things with a lot of high frequency content. Admittedly all the sounds I sampled in came back out in pretty good shape, but it is good to have a fall back position sample rate (in the form of a 44-50kHz option) in case you come across a really tricky sound.

The 18kHz sample rate was better than I expected it would be (on other samplers anything as low as this usually sounds completely grungy, if it's offered at all), but there was definite loss of top end frequencies. 9kHz was only really usable for those completely buggered up drum sounds on which Hip-Hop and House music seems to thrive, but I guess if that's your bag, then this is just the job!!!

Sample time is of course dependent on the sample rate being used. In this world, you don't get anything for nothing and you pay for the extra fidelity of the higher sampling rates with shorter sample time. However, the megabyte of memory in the FZ1 gives you a fairly generous time of 14.5 seconds and in the FZ10M, with its 2 meg as standard you get 29.1 seconds at this rate (as you do with the FZ1 if you add the MB-10 optional RAM board).

At 18kHz, all sample times are doubled and the 9kHz multiplies times by four, but I wouldn't use this last for anything but speech. The sample times on the FZ1 at the full 36kHz rate are enough to deal with most difficult instruments and if you are really stuck you can always get the memory expansion (unless you had the good sense to buy the FZ10M to start with). Comparison with other machines shows that in terms of memory, the Casios are perhaps the most generous, there is nothing else at £1299 which gives you anything like as much memory. The Ensoniq EPS (costing about the same) only has half the memory of the FZ1 and its rack version which has the same amount as the Casio rack costs £800 more. Akai's S950 is about the same price but also boasts only half the memory and the S1000 with the full 2 meg leaves little change out of 3 grand. Of course, these prices are not just reflecting memory size, there are other factors but it does give you a good idea about the Casio value for money.

Once you have set your sample's rate and time, you can look at the level coming in from your mic or line source and set high or low record level. This seems a bit primitive to me and you will probably need to be able to control the gain of your incoming signal at source to get the optimum level for sampling. However you can see the incoming signal in a VU-type display across the bottom of the Sampling screen on the LCD.

When you have made all the adjustments necessary, the actual sampling process can be started either manually or with an auto trigger which begins the sampling as soon as a pre-set Trigger Level is exceeded. You set this level against the VU-type display so that you can see the best point to set the trigger. Then when you arm the FZ by pressing the Yes key or by treading a footswitch as soon as the level exceeds the trigger threshold the sampling starts automatically.

Once you have made your sample, it can be shown immediately by pressing Display. Of course, you can change a sample's name or keyboard position once you have made it if either seem unappropriate. Now comes the digital cosmetics required to turn a raw sample into something that can be played musically. The first thing you may need to do is chop the front and/or rear of the sample. If you started sampling too early, when you play a sample back there will be a period of silence at the front of the sound, which will make your playing sound late. To get rid of it, you need to be able to set the point at which the sound starts. This is where the graphic display of the sample data really comes into its own. By zooming in with the display you can see very accurately when the really useful sample data starts and cut out the unwanted stuff at the front. Cutting the end of the sample is something you would normally leave until you've dealt with any looping operations.

Looping is the way a sampled sound of finite length can be made to sustain indefinitely when played from a keyboard. Normally, most samplers offer one or two loops, but Casio in their generosity have allowed for up to eight on each sample. Whilst this is almost certainly excessive for 99% of all samples, it's better to have overkill available than be under-gunned.

Looping is one of those tedious tasks designed to give us a foretaste of purgatory and many people assume that when a sampler doesn't instantly give them a superb loop, it's the fault of the sampler. Unfortunately the truth of the matter is that there "ain't no easy road" to the perfect loop. As far as this is concerned the FZ does its level best to help you out. If the coarse and fine parameters don't find you a good natural loop, then you can "cheat" with Crossfade Looping. This is where the FZ alters the sound data to produce a neater match in the data.

Once you have a working loop/loops, it is time for you to think about truncating the sound (technical jargon for cutting off the unused ends) to use the memory of the FZ most efficiently - just because there is plenty of memory available is no excuse to squander it on data that will never be heard. Again the graphic display lets you see what you are doing.

Other parameters associated with the sample itself include envelope settings for both filter and amplifier and the effect that velocity has on them. Although many people don't use much shaping on samples, it's nice to have the facility if it's needed and velocity control of the filter especially can save you needing to crossfade samples to produce a realistic soft to loud transition in playback. The envelopes are 8 stage jobs, again more than equal to the demands usually made by samples. But when we come on to the other ways of creating sounds on the FZs then this flexibility really comes into its own.

Casio call the sample and it's parameters a Voice but Voices can also be created by selecting standard waveforms (as on analog synths), adding sinewaves together digitally or by drawing waveforms. These are all grouped together under the Wave Synthesis menu. The preset standard waveforms available are Sawtooth, Square, Pulse, Double Sine, Saw Pulse or Random. Obviously these waveforms need to be processed through the Casio's filter and amplifier envelopes (which we will examine shortly) in the same way as you would on an analogue synth, as raw synth waveforms tend to be a touch on the dull side when listened to straight! The same tends to be true of the majority of waveforms created by the Sine Synthesis method. This allows you to set a level for each of the first 48 harmonics in the natural series (many previous and often considerably more expensive systems have only allowed for the first 32 in the series) and then the FZ will compute these into a waveform. The number of harmonics available is quite generous (anything about the 48th harmonic can probably only be heard by the dogs in your neighbourhood) and this does give you a virtually unlimited palette of static waveforms. However some of them might take an age to set up (just to give each of 48 harmonics a level between 0 and 256 is no mean task) and so Casio have included other ways of arriving at complex waveforms. To "draw" them use the right cursor key to move across the x-axis and the data entry slider to give the y-axis position. This is similar to a technique first made available on the Korg DSS-1, but here you get to control how quickly or slowly you want to move across the x-axis. It definitely takes a bit of getting used to this "drawing" technique, but once you are used to it, it certainly provides a speedy way of coming up with an original waveform. It should appeal to the artistic temperament.

The scientific alternative the FZs offer is cutting a section out of an already sampled waveform and using this as a synthesizer waveform. Of course, in many cases it will sound nothing like the sampled source as it is the continuous variation in natural sound which makes it sound interesting and real. Some sounds will retain their timbre in a single cycle (most woodwind for example) and you could cut a section of a sample of one of these and make a fairly realistic patch which would be very economical on sample data. For the most part though, this Cut Sample approach will be a source of new waveforms for creating original synth sounds.

To create realistic representations of acoustic instruments it is often necessary to combine samples across the keyboard into a multi-sample. Once a voice has been created you can assign it to an Area. Each area has a Voice number which is assigned to it a highest and a lowest note (which obviously need to fall within the highest and lowest assignments within the Voice parameters), a volume level touch sensitivity maximum and minimum (to allow for velocity switching/crossfading) and most importantly (at least as far as this writer is concerned) an individual MIDI channel assignment (over and above the basic MIDI channel) and separate output assignment. These last two are the key to the all-important multi-timbrality which is the way to make any instrument capable of it really earning its living. In conjunction with a sequencer, you can have different instruments responding on different MIDI channels and coming out of separate outputs into your mixing desk/signal processing etc. The FZs allow for all 16 MIDI channels and there are eight separate outputs to mix to.

So far everything we have looked at has been common to the FZ-1, FZ-10M and FZ-20M (except that the two racks have double the keyboards memory). However, the SCSI capability is unique to the FZ-20M. All the FZs have a common interface (which looks like RS-232 connector but isn't) which allows System Exclusive Dumps between FZs, but SCSI allows much more exciting communication between the FZ-20M and SCSI storage devices like hard disks and the removable cartridge drives about which I waxed lyrical in last months EPS Update. These allow the storage of many floppy disks' worth of sounds in a format which loads many times faster than floppy disk. As the FZ-20M has only recently been announced, there hasn't been time for companies like Blank Software and Digidesign to state whether their Macintosh sample editing packages (Alchemy and Universal Sound Designer respectively) which already work via MIDI with the FZs will be updated to communicate with the FZ-20M via the much speedier SCSI, but as this seems to be the way things are going and the FZ-20M will be the cheapest SCSI sampler on the market, it seems highly likely. It would be the natural continuation of the way this family of samplers has grown over the past 18 months keeping pace with all the latest industry standards, but setting new and highly desirable trends in pricing.

Products: FZ-1, FZ-10M, FZ-20M
Supplier: Casio Electronics, (Contact Details)
Price: FZ-1 SRP £1299, FZ-10M SRP £999, FZ-20M £1699