Re: [AH] sampling-based synthesis

From sam
Sent Tue, Mar 27th 2018, 20:50

Brent,

The samplers with analog filters came out during an era when sampling 
was brand new, and it was still easier to make analog filters (using 
dedicated analog filter ICs).  As technology marched forward, sampling 
became better, and at the same time it became easier/cheaper to use 
digital filters (for example, the EPS and then ASR-10 samplers).  The 
Mirage is a horrible beast to sample on, but boy, does it have 
character!  The ASR-10 is a breeze to sample on, but you don't get the 
rich filtering.

One correction to your post: the Soundprocess OS disk is $19.95 from 
Syntaur, the $90 product is a collection of 7 disks (21 banks) of 
Soundprocess sounds.  Soundprocess does add some cool features, and lets 
you create sounds more in a wavetable-synth way - but the voice 
architecture is still fairly limited.

Sam Mims

Syntaur



On 3/27/2018 3:35 PM, Brent Busby wrote:
> I've purchased an Ensoniq Mirage rack (not arrived yet), because I
> wanted to experiment with my own sampled waveforms and analog filters.
> Partially my choice of the Mirage was because of the amount of money I
> have to spend right now, not because it was necessarily the best choice
> for this.
>
> I am rather disappointed by its voice architecture though.  It offers
> two oscillator (sample playback) per voice, but there is only enough
> independant control over the pitch of the second oscillator to allow for
> chorus/detune type effects.  There is one LFO, which can only send
> triangle waves, and can only modulate the pitch of both oscillators.
> (Basically it's hard wired to be a modulation wheel effect.)
> Thankfully, the filter does have its own EG separate from the output
> EG.  It's 8-bit with playback rates as low as 8kHz, which should be
> charaterful, and the filters are CEM chips, but still, it's awfully
> hobbled and I would have liked more...like perhaps an SQ80 voice engine
> where you supply your own waves.
>
> It seems though that with sampler/synths you have to choose:  You can
> have a nice voice architecture, but digital filters, or you can have
> analog filters and a very crude voice engine.  The Korg DSS-1 is the
> only exception I've seen, and frankly I haven't been impressed by the
> actual sound of most of what I'm hearing in YouTube demos of it.  (This
> could very well just be the demos.  I've found you have to really
> actually have a piece of gear for at least six months to really find out
> anything, and I've never had a DSS-1.)  But there definitely seems to be
> a real vacuum out there when it comes to samplers with deep voice
> engines and real analog filters.
>
> I've heard about an alternative third-party OS for the Mirage called
> Sound Process that's supposed to hack in all sorts of things Ensoniq
> never thought of, including 4-oscillator/sample per voice (sounds
> promising), multitimbral/multitrack operation, etc...basically, some
> keen soul with a lot of machine language coding expertise on the Mirage
> has hacked it silly and created a whole new environment for it.  The OS
> disk is about $90 from Syntaur, so I think I'll want to check that out.
>
> If DSI or some other manufacturer today wanted to fill a void in the
> market that's really never been fully addressed though, not even in the
> previous analog era in the 80's, they should make something with a
> really good polyphonic voice engine with matrix type routing that has
> some good analog hardware filters, but have the oscillators be sample
> playback...preferably very low resolution sample playback, and you
> supply the samples.  To me, the problem with PPG/SQ80 style wavetable
> has always been that they already chose the waves for me before I had a
> chance to do anything.  I want to have lots of EG's and LFO's, and have
> analog filters, and supply my own waveforms...basically a Korg DSS-1
> with great tone that can interface well with modern computers over USB.
>