[sdiy] Solaris at NAMM

Thu Jan 21 07:19:56 CET 2010

On Thu, Jan 21, 2010 at 00:44, Richard Wentk <richard at wentk.com> wrote:
> Solaris is probably a bit cheaper though. :)
>
> But yes - synthesis has in fact stalled.

Yes, in fact Solaris is only an extension of things done already in the 60s:

http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1050166

which later resulted in SPICE, which then resulted in PSPICE, and we
all know that VAs steal their algorithms from that.

Funny that we had the tools to make virtual analogs before real
analogs existed. *

> You *have* to do clever algorithmic things to make really special noises, and clever algorithmic things turn out to be desperately hard to do well. (Or in some cases, at all.)
>
> So we have products like Solaris, which are basically simulations of other stuff that are easy to understand, and easy to work with, and make pleasant sounds, but don't push the envelope at all.
>
> For comparison see the Hal Alles synth from the 1970s. It eventually turned into the Synergy and Crumar GDS, but the bare version was an open bank of oscillators, multipliers and minimal filters that you could patch together ad lib.
>
> Since you basically needed at least a degree and more likely a PhD to understand how to use it, it was a one off. Likewise the Di Giugno 4X machines at Ircam.
>
> What's frustrating is that there are no equivalent R&D efforts happening today. There's a lot of hobbyist interest, especially in alternative interfaces, and some mainstream academic effort in that direction too. But - so far as I know - there are no ground breaking Alles, Moorer or Di-Giugno-style synthesis projects happening.

You're just not looking hard enough. Trawl through the kvr audio
forums, they've got a lot of esoteric junk going on there.

Or check out the Max/MSP community sometimes. They have a head office
somewhere in Williamsburg.

> There's also little or no musical AI, which is something that could easily have been pushed further than it has been - not in a playalonga way, but to create context- and performance-dependent synthesis.

Play-alongs are the domain of DSP (as in, dsp the degree you get at
university, not dsp the kind of silicon chip) and more exactly pattern
recognition.

> Here's the Alles synth and Laurie Spiegel from 1977:
>
> http://www.youtube.com/watch?v=fWKDsfARXMc

Neat. I bet nowadays Laurie would be using MSP while wearing
thick-frame glasses and a black turtleneck.

I dig the tune, recognize it from somewhere. Real nice! Thanks for a
nice reminder.

* This makes me wonder: if the analogs are now considered 'retro' and
the 'original thing'... can we say virtual analogs have actually been
first? Can we go and state that real analogs were the 'new' thing, and
virtual analogs are the 'retro'? And where does this put the DX7?..

D.

>
>
> On 20 Jan 2010, at 19:50, cheater cheater wrote:
>
>> I like my 816 but it is not a supercomputer, in fact due to a 'trick'
>> people at Stanford came up with the whole algorithm could be executed
>> with simple additions and the output had to go through a single LUT.
>>
>> Nope, the THX Deep Note was not done on a Synclavier. It was also not
>> a CS-80. And it is not an organ and not a recording. It is an
>> algorithmically produced sound created by Dr James 'Andy' Moorer with
>> the use of a mainframe computer on which a program ran (what could be
>> called a DSP program, if the term DSP had existed then), all this at
>> the Lucasfilm computer division. I know little about what the hardware
>> was actually, but here's what the algorithm was doing according to
>> Moorer:
>>
>> " I could get about 30 oscillators running in real-time on the device.
>> (...) The oscillators were not simple - they had 1-pole smoothers on
>> both amplitude and frequency."
>>
>> "20,000 lines of code [written in C] produce about 250,000 lines of
>> statements of the form "set frequency of oscillator X to Y Hertz" "
>>
>> So, considering that this is about 13 seconds of sound and 'Andy's
>> recollection of the numbers is correct, that is an update rate between
>> 640 Hz and 19.2 kHz depending on how those "statements" looked. That's
>> quite a good amount of performance and I would take a bet that not
>> much worse than the Solaris. Bear in mind that the calculation in the
>> solaris does not happen at 96kHz, it's just the output rate. The
>> algorithms probably get updated at a lower rate. Maybe 2 kHz, that's
>> the standard right now - ever wondered why VAs have LFOs that only go
>> up to 2kHz? That's why.
>>
>> Considering that the Solaris still isn't out and the THX sound was
>> made on something that has existed 30 years ago, I think the
>> comparison is quite fair.
>>
>> D.
>>
>> On Wed, Jan 20, 2010 at 20:23, Graham Atkins <gatkins at blueyonder.co.uk> wrote:
>>> Yes Cheater,
>>>
>>> You're absolutely right, that was done on the Synclavier which does
>>> have a lot of power. You could also include the Yamaha TX816 I suppose
>>> which could also make some impressive noises.
>>>
>>> Graham
>>>
>>> On 20 Jan 2010, at 16:54, cheater cheater wrote:
>>>
>>>> http://www.youtube.com/watch?v=Sfg9DVwOd9w
>>>>
>>>> On Wed, Jan 20, 2010 at 17:40, Tom Wiltshire <tom at electricdruid.net>
>>>> wrote:
>>>>>
>>>>> On 20 Jan 2010, at 14:44, Richard Wentk wrote:
>>>>>
>>>>>> It's basically a 1980s instrument that's very late - like a PPG Realizer
>>>>>> that actually works (or will work) - not a 2010s instrument that's ahead
>>>>>> of
>>>>>> the game.
>>>>>
>>>>> That's a bit unfair, I reckon. Show me a single 80's synth that has
>>>>> anything
>>>>> like the processing power in Solaris. It's got 6 SHARC processors, and
>>>>> does
>>>>> all it's audio calculation in 32-bit floating point at 96KHz. That's
>>>>> definitely a 21st century specification.
>>>>>
>>>>> Whether it does anything genuinely new with all that power is obviously a
>>>>> separate question.
>>>>>
>>>>> T.
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Synth-diy mailing list
>>>>> Synth-diy at dropmix.xs4all.nl
>>>>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>>>>>
>>>> _______________________________________________
>>>> Synth-diy mailing list
>>>> Synth-diy at dropmix.xs4all.nl
>>>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>>>
>>>
>>>
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
>