>I can understand why this box has achieved cult >status. The reason the FS1R never sold well is >because most consumers are illiterate >technically and so they buy dumbed-down PCM >based romplers. The net result is they sound >just like everyone else

No the reason it never sold well is because it is impossible to program from the front panel, the presets are a joke as is the manual, operating system and most of all consumer support. The formant synthesis is so badly implemented that only a good guess would achieve the desired sound. The FSEQ's are mere gimmick and the machine doesn't allow you to create your own via software that would convert a waveform into an FSEQ (Yamaha created such a piece of software but refuse to relase it to the general public). The FM implementation is bog standard and based upon lookup tables. The sound quality is good, the FX are good, the filters are ok. Overall though, it's easy to see why Yamaha made such a loss from this machine: they haven't got a clue about the needs of real-world musicians and programmers and how to implement their ideas effectively and realistically so as to be actually usable by mere mortals.

Lets look at its limitations first - the hardware interface is downright nasty to use - in keeping with the entire FM family. The FS1R has only 4 parts multi-t and limited polyphany. There is no arpegiator or real sequencer onboard. Add in the fact that Yamaha dropped it after only 18months on the catalog - so future support will be limited. Why then are some claiming that the FS1R is one of the best synths of the last 10 years?

In short, its a sound designer's synth. It has one of the most interesting and fully featured sound engines since the golden age of the modular. Its a facinating synth to program because every sound you create is an adventure into uncharted acoustic terratory. In an age where "me too" pseudo-analogs and PCM ROMplers account for 99% of all the electronic instruments made, the FS1R offers the musician the oppertunity to create voices that are really unique and personal.

If you can find a FS1R in a music shop, don't flip through the presets. Select just one e-piano sounds and really work with it. Notice its quality, its responsiveness, its musicallity. The FM sound engine allows the creation of sounds which truely emulate the differing timbral qualities of acoustic instruments in response to the players touch. Sure the e-piano doesn't sound as like a rhodes as a sample based synth but after a few seconds this becomes irrelevant since the e-piano behaves like a rhodes, its sound shifting and mutation in response to keyboard technique. This is a huge inprovement for player and listener compared to a couple of velocity switched samples.

I think this is the key to the FS1R. The best way to think of the FS1R engine is that it is a type of physical modelling synth.

Everyone has their idea of FM synths. The do bells and basses and tinkly 80 e-piano's right? Wrong. The FS1R can create virtually any sound - from really good analog impressions noted by the other reviewers, to Waldorf wavetable like sounds, to acoustic instruments.

With a reasonable computer editor like the free SoundDiver version that can be downloaded from the net you can rapidly get inside the most powerful FM and FS engine ever commerically available.

Inside you have 88 patch combinations linking 8 modular oscillators (algorithms and ops as Yamaha termed them). Each oscillator can be switch to generate sine, two tpyes of square, two saw and formant wave forms. The combination can be fed through a near analog quality filter. The levels of each oscillator can respond to virtually any midi signal. Add to this another 8 noise sources and a wonderful effects section and you are already far far away from the type of programming and thinking that one has when confronted by a "classical" osc-filter-amp combination in an VA synth. The programmer has so many resources at his or her disposal that it makes more sense to think of the FS1R as a modular-saurus or a powerful academic physical modelling rig.

An example: create a santur like instrument ... dual strings - so let two oscillators chourus through detuning. The strings are metal so introduce some non integral harmonics using FM of the first two operators. Lets set that up so that the amount of aharmonic tones increases as the string get thicker at the bass end.

The instrument is struck with hammers so lets use some the 8 non-voiced oscillators to create the percussive noise of hammering the string. Lets introduce more noise components to model the resonance of the impact through the body of the instrument.

Lets model the instrument's wooden resonator with a resonant band pass like filter setting ....

Get the picture? Because they have a lot of resources available, the sound designer is thinking about building a phyical/acoustic model of the instrument they are creating rather, as one tends to do with the "classic" model, thinking about how to best utilise the few basic units available from the VA or ROMpler engine. Sure, its a long way from "lets twist this knob and see if it squeals", but the sound quality is the reward.

The FS1R is not for everyone but if you are intrigued by FM or want some very unique voices or like complex sound design engines then you simply have to get one! The second hand price is certain to rise.