http://razmo.ziphoid.com/E310U_GT.mp3
A small example of me messing with an Eminent 310 Unique sample on the V-Synth GT...
Very cool. And pretty much indistinguishable, as you said.
What exactly do the formant capabilities include? – Never been a Roland guy, so that's why I'm asking.
Well... in essence, all there is to it is that you can encode a sample in a way that when you play back the sample at any pitch, it will keep the formants of the sample steady thru the whole keyrange... you have to use a solo sound though, since it obviously recognize the pitch and does some FFT analysis, so if more than one solo sound is happening, it can sound a bit wrong... but it's perfect for sampling solo accoustic instruments.... flutes, guitars, bells etc... and of course voices.
When you have encoded your sound, you can change the formant character by shifting it up or down... with a voice that will make it either more female or male and into extremes... but also the time of the sound is kept... so any sound with rhythmic content will keep the BPM thru the whole sound... and then pitch can also be adjusted... these three building blovks are independant of each other, and that's where the fun part lies.
But this is not all... these blovks can of course be modulated by dosens of controllers... LFO's, Keynote, Wheels etc... Envelopes... basicaly there is no limits here. This is the basic oscillator settings for a sample.
You have two oscillators you can assign like this... totaly independant of each other, and these are then combined using different algorithms like simnple mix, FM, Ringmod and other types...
This then goes thru two COSM processors... these are per voice, and basicaly they are FX engines with stuff like distortions, filters, frequency shifters, decimators etc.
Then all that goes thru a VCA, and then finaly into an advanced FX section with three multi effects.
You can also configure these buildingblocks in different routings... and on top of that, you can add something called AP synthesis which is basicaly a kind of physical modelling behavior.
The possibilities are mind boggling to be honest... you also have VA synthesis that you can choose instead of a sample oscillator with dosens of waveforms.
But that's not all... this "block" is just ONE of two... you have an UPPER and LOWER block that are completely identiccal but separate fully... this allow for splits, stacks etc...
Strangely enough, even though this synth has so much power, it's extremely logicaly laid out, and very easy to program.
I can certainly say, that this machine is truly unique... there are nothing like it in hardware that I know of, and playing with formant really puts this in a totaly different league than any other sampler... just the way you can "play" thru a vocal sample with phrases in it is incredible... you can litteraly play it a bit like scratching, even playing the phrase backwards, stop it totaly for a freeze in the middle of a wovel.
Sometimes it does screw up the encoding of some samples.. where it "hickups" between different octaves in the sound playing back... but then you have two other encoding options that allow these sounds to work anyway, but just without formant control... good for loops and huge ensemble sounds etc.
When I look at the architecture of the V-Synth GT I see many resemblances to the LA synthesis that Roland put in their very famous older synth D-50... in fact I'd call the V-Synth a D-50 make II because it has the same elements basicaly... the samples and VA synthesis is also present in D-50... D-50 also had algorithms for mixing the two oscillators via FM, ringmod etc... and then the FX sectiion.... and the fact that the first version of the V-Synth actualy had a D-50 emulator on card/built in also hints that there are references to the D-50.