« Last post by drpgleeson on July 28, 2016, 11:49:51 pm »
Re "missing" resynthesis: my recollection (I was an early Synclavier adapter back in the day) is that while NED promised resynthesis, there never was a program doing actual resynthesis, which requires two parts: an analysis method of the input waveform and a reproduction method to create the synthesized output waveform. The analysis, to be useful, would have to be more than static -- it would require analyzing many samples of the actual instrument. I won't even get into blown or bowed instruments, which are relatively trickier. But even a resynthesized piano would probably need at least 50 loudness levels (and some classical pianists claim that they have 100 loudness levels at their command). Each of these levels would have a different overtone profile varying over time. You'd have to capture that in some way -- either laboriously analyzing the original waveform many times--every millisecond would be, I'd guess, pretty much the minimum -- or else figuring out some clever way of building an algorhithm that produced an analogous dynamic overtone change based, for instance, on the overtone decays subsequent to the vibration of a taut metal string.
Anyway, what I'm saying is that actual resynthesis has never been successfully managed. I've owned a couple of instruments that purported to produce resampling and they were only able to approximate a "toy" version of the original -- toy flute, etc.
But it's probably more interesting and a whole lot easier to resynthesize a waveform statically -- in other words, reproduce it's overtone series at a single moment of time -- then figure out "interesting" rather than "accurate" ways of dynamically changing it: FM, AM, periodic or aperiodic manipulation of individual overtones or groups of overtones, , overtone frequency shifting, etc. For instance, you could resynthesize an mf piano, then arrange a way of shifting the overtone series from the usual one for most melodic instruments to the overtone series for a struck rod. If this were available as a dynamic modulation using a mod wheel, listening to the transformation could be musically and conceptually pleasing.
Having said this, I doubt that Arturia's going to spend a lot of money developing a resynthesis capability (although I'd love to be wrong); for one thing, not many musicians really want to dive into the complexities of a resynthesis program. Or any synthesis program, for that matter. One of the guys at Sequential Circuits told me that almost all the original Prophet 5s that came back for service still had all 40 of the original factory patches, without the slightest alteration.