That being said, I may as well make an addendum of my review to Arturia's Synclavier V. Part one of the review can be found here.
As someone obsessed with the sound of old-school samplers, I struggled for the past few years to have the perfect sound for sample playback in software while conserving both disk space and CPU, which is practically impossible to do trivially. It even got to the point that, as some of you know, I even went as far as trying to code my own software synthesizer.
Well, the good news is that that may no longer be necessary: the new version 2 of the Synclavier V VST by Arturia incorporates both sample playback and sample resynthesis, both being reminiscent of later Synclavier systems.
Will try to update this post with my thoughts about the resynthesis function, which is revered among many other Synclavier fans, later on, but for now, I'm focused on the sampling support.
The Synclavier was known to be a rival to the Fairlight CMI due partially to the fact that later Synclaviers had support for sampling. In fact, in many ways, the Synclavier overpowered even the Series III in specs: up to 768 MB of RAM (even the Akai Z8, the last traditional hardware sampler from the early 2000s, could only support up to 512 MB), 9 GB of external hard disk space, and up to a 100 kHz sample rate. The only benefit the Fairlight had over the Synclavier, to me at least, was the Fairlight's inclusion of an antialiasing filter to muffle the ringing frequencies that come from transposing samples too low: the Synclavier simply did not have the capacity to do this, let alone per voice!
All this in mind, I was pretty wary to try out the updated V VST's sampling, especially given I already did try out the Fairlight CMI V VST from the same company. Thankfully, I was pleasantly surprised: whether Cameron Jones ended up using my infinite linear oversampling interpolation code or not, this VST actually sounds faithful to the variable clock samplers of the early to mid 80s. I tried transposing a sample (which, oddly enough, was from the Prosonus Library, which was originally intended for the Synclavier) chromatically, and long behold, the only artifical frequencies I clearly heard were the imaging frequencies when I transposed by more than an octave, which, for variable clock samplers, is supposed to be the case. Too often VST developers, including the CMI V team, take the easy way out and just use a mere piecewise interpolation for sample playback, without any oversampling.
This update even includes an antialiasing filter like the CMI had, though one minor gripe is that the cutoff value is not editable like for the CMI at this time. The only other disappointment I can think of is that we're only allowed a single sample per partial, meaning we can only have up to 12 samples in a program, a major contrast to the powerhouse that was the hardware (a convincing piano sound from the factory library had a whopping 84 samples, mapped by both key and velocity!). Hopefully both of these qualms of mine will be addressed later, especially since the multisampling was a big selling point for the original hardware (in the early 80s, the only other sampler capable of multisampling was later OS versions of the Emulator I).
UPDATE: tried out the resynthesis functions as well. Certainly impressive to my ears: while I have no experience with the analysis functions of the original hardware, it sounds fairly useable thanks to its timbre frames' support for cross fading, a contrast to the wavetable-based analysis of the CMI and the CMI V. It also is easier to analyse samples to a certain pitch than for the CMI V, which, in my experience, seems to hardwire the root pitch to a given sample rate.
This is definitely great progress to recreating the sampler-based sound of the 80s, and this update to the VST is most definitely recommended to people who want lo-fi sampling but can't afford a CMI or an actual Synclavier.
Well done, Cameron and Arturia! The VST can be bought directly from Arturia here: www.arturia.com/products/analog-classics/synclavier-v/overview