On Thu, 10 Jul 2014, Rob wrote:
On 07/10/2014 10:52 AM, Fede wrote:
I also think that producing OPL3-quality FM synthesis with the OPL3's
polyphony is trivial for today's CPUs, which is why you don't see
those technologies around much anymore.
Yes, it might be trivial, but as with every hardware piece there is a
sound to it, which I like and want to know what's possible with them
(specially since I can't afford a nice analog synth at the moment).
FM is actually digital synthesis, though some implementations included
analog filters.
Actually FM can be done in pure analog synthesis. It's nothing more or
less than one oscillator's output being used as a modulator to be sent
into the input of another oscillator on the same synth, and lots of
analog synths can do that. You can do FM patches on a Minimoog.
Yamaha likes to act like they invented it of course, and they've been
acting that way so long now that we tend to think now that it has
something to do with digital. :)
Actually where Yamaha came into is was this:
The part of a synthesizer where analog is most needed for the sound is
the filter. Making a digital oscillator is not that tough and doesn't
sound too bad, and neither is a good digital envelope generator hard to
make either, but in the 80's, it was proving really hard to make a
decent, musical-sounding digital resonant filter. By the mid-80's, most
keyboard manufacturers were making hybrids (Roland JX series, Ensoniq
ESQ-1, Korg DW series, Sequential Prophet-600, etc.), where everything
was digital *except* the filters, which were still analog. Mostly all
people wanted from digital if it came right down to it was stable
tuning, and since the filters don't affect your tuning, these hybrids
would have been enough to meet what people were expecting, and they
could have just kept on making them. Of course, the keyboard companies
kept going, and we got pure digital (with digital resonant filters)
later, but for what it's worth, the hybrids pretty much fulfilled
everything people were actually asking for, and we could have settled on
that and everybody would have been happy.
Anyway, Yamaha, unlike all the others, decided to try to get ahead of
the game by making a pure digital synth before anyone else, and they
tried to solve the filter problem...by NOT HAVING ANY FILTERS. Bahaha!
Their idea was that if you had enough oscillators cross-modulating
eachother, the resulting wave would have so many interesting harmonics
in it that you could forgo the usual resonant filter completely, and it
would still sound good enough for most people. The DX7 had six digital
oscillators per voice (though Yamaha calls them operators instead of
oscillators). Most people now agree, now that a popular taste for
analog has returned, that the DX series was pretty cold and sterile
sounding. However, they were dependable. You could count on a DX7 to
not need any warm-up after you turned it on, and a patch on a DX7 would
sound the same every time you used it. As for the sound, in pop music
it was rarely heard without some kind of effects processing, so usually
you never heard raw dry DX7 that much anyway. People actually bought
zillions of them.
The irony is, Yamaha was theoretically right. With enough oscillators
cross-modulating eachother, you could make filters optional. Where they
got ahead of themselves was thinking they could do that with only six
per voice. Dozens or hundreds, yeah maybe. Six, no.
What the OPL3 produces cannot be any different than what an effective
emulation of it can produce. It literally outputs digital data which
needs to be converted by an external DAC, which is normally right
there on the same board.
That's true -- even digital synths have character. An OPL3 chip is
probably not the hardest thing in the world to make an emulation of, but
there's always some quirk of the hardware that gets lost in the
translation when you try to make a software equivalent of almost
anything.
On the other hand, SID's synthesis is analog, something that really
confused me as a 15-year-old trying to make video game sound effects
but which I really appreciate today. I think the only reason we don't
see more "SID in a box" devices for the increasingly popular chiptune
culture is that chiptune composers seem to favor Nintendo devices like
the NES and Game Boy. But the SID is much harder to emulate accurately
than those, and an external box would really be desirable (if one
doesn't find a way to just network a C64 to do it, which may be
cheaper).
I used to have an Elektron SidStation in my studio (sold it), but it was
a device like you're talking about -- a little Midi-controlled silver
box with a real SID chip in it.
--
+ Brent A. Busby + "We've all heard that a million monkeys
+ Sr. UNIX Systems Admin + banging on a million typewriters will
+ University of Chicago + eventually reproduce the entire works of
+ James Franck Institute + Shakespeare. Now, thanks to the Internet,
+ Materials Research Ctr + we know this is not true." -Robert Wilensky
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@xxxxxxxxxxxxxxxxxxxx
http://lists.linuxaudio.org/listinfo/linux-audio-user
