[sdiy] DIY Polysynths
music.maker at gte.net
Tue Jul 28 22:43:53 CEST 2009
paul.maddox.mail-list at synth.net wrote:
>> It seems to me that the pinnacle of ambition (and, optionally, success)
>> in the Synth DIY world is to build a polysynth.
>for me, it's almost the holy grail of SDIY.
>I lvoe creating new modules and new sounds, but building 16 of the
>same thing seems tedious (one advantage of code).
>> There are only a few
>> people in the world who seem to have done this. I've managed to find
>> these below. Can anyone add anyone else to this list?
>> Roman Sowa: http://www.sowa.synth.net/poly/index.html
>> Juergen Haible: http://www.jhaible.de/tonline_stuff/hj_jh4.html
>I might be tempted to add <don's flame proof suit> my Phoenix polysynth ;-
>Though not 100% finished, it's missing some modulation routing, it's a
>perfectly useable 6 voice polysynth, though it's DSP based, it's still
>DIY and it's still synth.
>> There are also a couple in progress:
>> Paul Maddox's Defender: http://www.vacoloco.net/synths/defender/
>> George Hearn's XL8:
>I'd actually push Georges to "almost complete" having played with it at SDIY.
>> Special Mention: Scott Gravenhorst's FPGA projects.
>Awww, they at least need to go in the "in progress", I seem to recall
>he did release a poly version of one.
Actually, more than one finished and/or released:
Released: GateManPoly 8 voices linear arithmetic subtractive
Released: GateManPoly_FM/8 8 voices 8 operator FM organized as 4 x 2-OP per voice
Finished: GateManPoly_FM/16 16 voices 8 operator FM organized as 4 x 2-OP per voice
Released: PolyDaWG/8 8 voices physical model, Karplus-Strong string
Finished: Xarp-56 56 voices physical model, Karplus-Strong string
In progress is a 16 voice 8 operator FM synth which allows far more than 2-op generator
combinations. eg. 2 x 3-op with 1 x 2-op, 2 x 4-op. such combinations can also include
sinewave oscillators. It will also operate as an 8 oscillator sine additive synth. Each
oscillator will have it's own ADSR and level. This synth is planned to have at least one
state variable filter (with ADSR) per voice for final timbre shaping.
There are also other experiments that I've not said anything about here. One I may return to
is a synth I call the phLUTe which is a physical model of a flute. While the steady tone is
very flute-like and it can be overblown (a good thing), it has problems with a transient that
occurs when the pitch is changed. The transient makes a kind of "clack" sound that is not
very flute-like. Others I won't mention because they are too neophyte at this stage.
I'm finding that there is great DSP power inside modern FPGAs. One can get nice development
boards that are capable of polysynth design for around US$50. I have 2 Avnet Spartan-3A 400
boards that I've ported the GateManPoly (subtractive) and the PolyDaWG/8 synths over to.
Both support 8 voices.
All of these synths use a free 8 bit embedded microcontroller (PicoBlaze) for MIDI and Sysex
updating of synth patch parameters and in some cases it also drives the LCD display.
Creating these synthesizers took far less time than I imagined. Granted they are digital,
but IMO, an FPGA allows the designer to use wide data words so that very high resolution
computations can be accomlished. While that isn't the same as analog, in a sense it begins
to approach it.
Some of the most productive time I have spent was to learn Verilog and read at sites like
-- Scott Gravenhorst
-- FPGA MIDI Synthesizer Information: home1.gte.net/res0658s/FPGA_synth/
-- FatMan: home1.gte.net/res0658s/fatman/
-- NonFatMan: home1.gte.net/res0658s/electronics/
-- When the going gets tough, the tough use the command line.
More information about the Synth-diy