Korg Poly800/EX800 Users

Hi!
I'm interested, if in the new soft/hardware update known as Poly/EX
800 MKIII coould be a Unision Option. It would be very powerful, 8
DCOs, detuned, ufff...

Hi Kozmisch,

Unison mode would be great if all 8 DCO's could be detuned but the MSM
5232 tone generator chip only has two clock inputs. One for four DCO's
and another clock input for the other 4 DCO's.

So Unison doesn't get much although it might be of some use if each of
the four DCO's are an octave shift away from each other.

Mike.

--- In korgpolyex@yahoogroups.com, "kozmisch" <kozmisch@...> wrote:

>
> Hi!
> I'm interested, if in the new soft/hardware update known as Poly/EX
> 800 MKIII coould be a Unision Option. It would be very powerful, 8
> DCOs, detuned, ufff...
>

Hi Kozmisch,

I thought a little more about Unison and portamento modes.

When you look at the design and features on the Korg EX-8000/DW-8000
you can see how Korg's engineering evolved. In the Poly 800 they used
a tone generator chip that had 8 DCO's grouped as a pair of four with
a separate clock for each pair. This naturally allowed a master
oscillator to drive the clock input along with a simple detune circuit
that would be used to make the second set of four DCO's detune away
from the first four DCO's.

In the DW-8000, Korg moved to an EPROM based waveform generation
system along with 16 separate clocks for each individual oscillator.
So unison mode in the DW-8000 sounds extremely fat and deep because
every oscillator is detuned away from the master clock. But the
clocking and TG electronics in the DW-8000 is about ten times larger
than that found in the Poly 800.

I am still investigating the tone and envelope generation parts of the
EX 800 software. Until I have a clearer idea as to how all of that
software works, I am not going to be able to make a decision on how
portamento/mono/unison mode is going to be squeezed into the EX 800.
But as I understand more about the TG and EG's, I'll keep everyone up
to date as to what I find.

Mike.


--- In korgpolyex@yahoogroups.com, "korgpolyex800" <korgpolyex800@...>
wrote:

>
> Hi Kozmisch,
> 
> Unison mode would be great if all 8 DCO's could be detuned but the MSM
> 5232 tone generator chip only has two clock inputs. One for four DCO's
> and another clock input for the other 4 DCO's.
> 
> So Unison doesn't get much although it might be of some use if each of
> the four DCO's are an octave shift away from each other.
> 
> Mike.
> 
> --- In korgpolyex@yahoogroups.com, "kozmisch" <kozmisch@> wrote:
> >
> > Hi!
> > I'm interested, if in the new soft/hardware update known as Poly/EX
> > 800 MKIII coould be a Unision Option. It would be very powerful, 8
> > DCOs, detuned, ufff...
> >
>

Thank you, Mike, for the explanation. I wish you good luck in the
process.)

Jani.





--- In korgpolyex@yahoogroups.com, "korgpolyex800" <korgpolyex800@...>
wrote:

>
> Hi Kozmisch,
> 
> I thought a little more about Unison and portamento modes.
> 
> When you look at the design and features on the Korg EX-8000/DW-8000
> you can see how Korg's engineering evolved. In the Poly 800 they used
> a tone generator chip that had 8 DCO's grouped as a pair of four with
> a separate clock for each pair. This naturally allowed a master
> oscillator to drive the clock input along with a simple detune circuit
> that would be used to make the second set of four DCO's detune away
> from the first four DCO's.
> 
> In the DW-8000, Korg moved to an EPROM based waveform generation
> system along with 16 separate clocks for each individual oscillator.
> So unison mode in the DW-8000 sounds extremely fat and deep because
> every oscillator is detuned away from the master clock. But the
> clocking and TG electronics in the DW-8000 is about ten times larger
> than that found in the Poly 800.
> 
> I am still investigating the tone and envelope generation parts of the
> EX 800 software. Until I have a clearer idea as to how all of that
> software works, I am not going to be able to make a decision on how
> portamento/mono/unison mode is going to be squeezed into the EX 800.
> But as I understand more about the TG and EG's, I'll keep everyone up
> to date as to what I find.
> 
> Mike.
> 
> 
> --- In korgpolyex@yahoogroups.com, "korgpolyex800" <korgpolyex800@>
> wrote:
> >
> > Hi Kozmisch,
> > 
> > Unison mode would be great if all 8 DCO's could be detuned but the MSM
> > 5232 tone generator chip only has two clock inputs. One for four DCO's
> > and another clock input for the other 4 DCO's.
> > 
> > So Unison doesn't get much although it might be of some use if each of
> > the four DCO's are an octave shift away from each other.
> > 
> > Mike.
> > 
> > --- In korgpolyex@yahoogroups.com, "kozmisch" <kozmisch@> wrote:
> > >
> > > Hi!
> > > I'm interested, if in the new soft/hardware update known as Poly/EX
> > > 800 MKIII coould be a Unision Option. It would be very powerful, 8
> > > DCOs, detuned, ufff...
> > >
> >
>

On Tue, 10 Jul 2007, korgpolyex800 wrote:

> In the Poly 800 they used a tone generator chip that had 8 DCO's grouped 
> as a pair of four with a separate clock for each pair. This naturally 
> allowed a master oscillator to drive the clock input along with a simple 
> detune circuit that would be used to make the second set of four DCO's 
> detune away from the first four DCO's.
======================

wouldn't it be possible/feasible to get a four (or eight) oscillator 
detune in a similar way to playing a chord: just tell each of the four (or 
eight) oscillators to play a slightly different note...? instead of the 
different notes being semitone intervals apart, they can be small 
fractions of a semitone apart, ideally with a parameter to adjust the 
detune.

i suppose that sound could also be done by inserting a chorus effect 
between the oscillators and the filter, but then we're in the realm of a 
hardware mod.

i'm just trying to imagine a poly-800 in mono/portamento mode, with my 
hardware mods, and some way to route velocity to the filter cutoff and/or 
resonance...


-- 
         ...atom

  ________________________
  http://atom.smasher.org/
  762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
  -------------------------------------------------

 	"I never gave them hell, I just tell the truth
 	 and they think it's hell."
 		-- President Harry Truman, 1956

i don t think playing a note with just any frequency is possible... there s probably a table of note-to-freq values somewhere hard-coded. if you remember, you

Great to hear from you again Atom,

The TG in the Poly 800 is the OKI MSM-5232. This chip was actually
used in many arcade games and that seems to be what it was designed
for. Regardless, the design has two master clock inputs. One clock
drives DCO1-4 and the other drives DCO5-8. In the Poly 800, chips
IC-18,19 and 20 provide the detune clock which uses the master clock
as input and detunes prior to feeding it to the second clock input on
the MSM-5232. The registers in the MSM-5232 itself only allow for the
selection of real notes, there is ability to detune the DCO's within
the MSM-5232 itself. That is why Korg moved to an sampled wavetable
design in the EX-8000.

So the only way to provide 8 fully detuned DCO's would be a large
hardware change (something I am considering for the future).

But you said "some way to route velocity to the filter cutoff and/or
resonance...".

I think that feature is very doable. In fact, it may be easier to
route velocity to the filter than changing the VCA EG's to track
velocity. So I have definitely noted that feature down for
implementation as soon as possible.

Thanks for thinking of it.

Mike.



--- In korgpolyex@yahoogroups.com, Atom Smasher <atom@...> wrote:
>
> On Tue, 10 Jul 2007, korgpolyex800 wrote:
> 
> > In the Poly 800 they used a tone generator chip that had 8 DCO's
grouped 
> > as a pair of four with a separate clock for each pair. This naturally 
> > allowed a master oscillator to drive the clock input along with a
simple 
> > detune circuit that would be used to make the second set of four
DCO's 
> > detune away from the first four DCO's.
> ======================
> 
> wouldn't it be possible/feasible to get a four (or eight) oscillator 
> detune in a similar way to playing a chord: just tell each of the
four (or 
> eight) oscillators to play a slightly different note...? instead of the 
> different notes being semitone intervals apart, they can be small 
> fractions of a semitone apart, ideally with a parameter to adjust the 
> detune.
> 
> i suppose that sound could also be done by inserting a chorus effect 
> between the oscillators and the filter, but then we're in the realm
of a 
> hardware mod.
> 
> i'm just trying to imagine a poly-800 in mono/portamento mode, with my 
> hardware mods, and some way to route velocity to the filter cutoff
and/or 

> resonance...
> 
> 
> -- 
>          ...atom
> 
>   ________________________
>   http://atom.smasher.org/
>   762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
>   -------------------------------------------------
> 
>  	"I never gave them hell, I just tell the truth
>  	 and they think it's hell."
>  		-- President Harry Truman, 1956
>

Be a PS3 game guru.
Get your game face on with the latest PS3 news and previews at Yahoo! Games.

Modulating the filter or osc from the noise generator is definitely a
hardware mod.

These sorts of mod's I am trying to address through digital
potentiometers that would be programmatically and or MIDI controlled.

So for example, the moog slayer mod would be able to be controlled via
one of the EG's or a MIDI controller.

I consider all of these to be potential stage 2 kits.

The stage one kit will upgrade the ROM/RAM and software with software
features only.

The stage two kit will provide a hardware mod board that integrates
all of the great hardware mods into one and provides MIDI and software
control of them.

To be honest, I'm not sure I'll live long enough to get the stage two
kit done.

Mike.

--- In korgpolyex@yahoogroups.com, Russ <russdaren@...> wrote:
>
> I dont think randomization is required so much as being able to use
noise as a modulation source for the filter or osc. not sure if this
is electrical or a programming issue. But I'd guess the former.
> 
>  
> ---------------------------------
> Be a PS3 game guru.
> Get your game face on with the latest PS3 news and previews at
Yahoo! Games.
>

theoretically, you could implement another MG (modulation generator) that would output randomized values... implementing a random from the scratch, on the

On Sat, 14 Jul 2007, jure zitnik wrote:

> theoretically, you could implement another MG (modulation generator) 
> that would output randomized values... implementing a random from the 
> scratch, on the other hand... :) it IS doable though. but i'm not sure 
> how much extra space it would take.
=====================

well, there's random, and there's RANDOM... i don't think any musical 
instrument needs a random modulation source that's cryptographically 
secure, just random enough. a 10-12 bit lookup table would provide 
"random" values with a short period, but probably "good enough (TM)" for a 
20 year old synth.

whether or not there's room or motivation for implementing it, i can't 
answer.


-- 
         ...atom

  ________________________
  http://atom.smasher.org/
  762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
  -------------------------------------------------

 	It's just a jump to the left
 		And then a step to the right.
 	Put your hands on your hips
 		And pull your knees in tight.
 	It's the pelvic thrust
 		That really gets you insa-a-a-a-ane
 	LET'S DO THE TIME WARP AGAIN!
 		-- Rocky Horror Picture Show

Yes, I've added that feature to my to do list.

We want to be able to create a randomising MG to compliment the
existing LFO and MG.

I think we definitely want to keep the existing MG and LFO but add a
second LFO that has square, sine, saw, reverse saw and triangle as
well as randomize. Plus an inverter.

We will want to be able to route them to VCF cutoff, VCF resonance and
DCO freq (vibrato) and DCO volume (for lovely tremolo). Any others?

How about that?

Mike.

--- In korgpolyex@yahoogroups.com, Atom Smasher <atom@...> wrote:
>
> On Sat, 14 Jul 2007, jure zitnik wrote:
> 
> > theoretically, you could implement another MG (modulation generator) 
> > that would output randomized values... implementing a random from the 
> > scratch, on the other hand... :) it IS doable though. but i'm not
sure 
> > how much extra space it would take.
> =====================
> 
> well, there's random, and there's RANDOM... i don't think any musical 
> instrument needs a random modulation source that's cryptographically 
> secure, just random enough. a 10-12 bit lookup table would provide 
> "random" values with a short period, but probably "good enough (TM)"
for a 

> 20 year old synth.
> 
> whether or not there's room or motivation for implementing it, i can't 
> answer.
> 
> 
> -- 
>          ...atom
> 
>   ________________________
>   http://atom.smasher.org/
>   762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
>   -------------------------------------------------
> 
>  	It's just a jump to the left
>  		And then a step to the right.
>  	Put your hands on your hips
>  		And pull your knees in tight.
>  	It's the pelvic thrust
>  		That really gets you insa-a-a-a-ane
>  	LET'S DO THE TIME WARP AGAIN!
>  		-- Rocky Horror Picture Show
>

On Sat, 14 Jul 2007, korgpolyex800 wrote:

> I think we definitely want to keep the existing MG and LFO but add a 
> second LFO that has square, sine, saw, reverse saw and triangle as well 
> as randomize. Plus an inverter.
>
> We will want to be able to route them to VCF cutoff, VCF resonance and 
> DCO freq (vibrato) and DCO volume (for lovely tremolo). Any others?
=============

how about making MG (LFO) 1 and 2 targets for MG (LFO) 2?

can detune be a target? or would it sound "zippery" because of the way 
that the detune is implemented?

if the new LFO 2 has an inverter, would it need a saw and reverse saw? or 
would the reverse saw really be handled by the inverter?

this may be overkill for a poly-800, but would the LFO 2 be free-running? 
trigger per note? re-trigger on each new note in a chord? my guess is that 
either of the first two options would be good; the third option would be 
not so good.


-- 
         ...atom

  ________________________
  http://atom.smasher.org/
  762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
  -------------------------------------------------

 	"The sum of intelligence on the planet is a constant.
 	 The population is increasing."
 		-- Cole's Axiom

Hi Atom,

The existing and future LFO are implemented as software generated
cycles. I haven't quite worked out the exact math used to generate the
waveform (is the sine calculated or is it based on a lookup table?)
but since we already know how to set the MG via a MIDI controller it
would seem possible to modulate its rate using a second LFO.

You're right, saw with an inverter is reverse saw. Since we'll have to
implement triangle we might as well make the decision when we get to
it as to whether to use an inverter or not. Perhaps the inverter will
be a behind the scenes thing anyway. Also, square wave needs to be
pulse width ratio adjustable.

LFO 2 will be both free running and note triggered just like EG3 can
be re/triggered except that we'll add free running as an option. Also,
I am hoping to be able to change the EG3 triggering so that it can be
retriggered by LFO2. So then the LFO could be triggered by notes but
EG3 can recycle according to the LFO rate. That would give us the
ability to cycle through the EG3 envelope over and over.

Oh and I'm hoping to make the original LFO free running optionable too.

Detune would definitely be zippery because it is only 2 bits in depth.
A better way would be to implement a second master clock. But that
would take a hardware mod.

Mike.

--- In korgpolyex@yahoogroups.com, Atom Smasher <atom@...> wrote:
>
> On Sat, 14 Jul 2007, korgpolyex800 wrote:
> 
> > I think we definitely want to keep the existing MG and LFO but add a 
> > second LFO that has square, sine, saw, reverse saw and triangle as
well 
> > as randomize. Plus an inverter.
>
> > We will want to be able to route them to VCF cutoff, VCF resonance
and 
> > DCO freq (vibrato) and DCO volume (for lovely tremolo). Any others?
> =============
> 
> how about making MG (LFO) 1 and 2 targets for MG (LFO) 2?
> 
> can detune be a target? or would it sound "zippery" because of the way 
> that the detune is implemented?
> 
> if the new LFO 2 has an inverter, would it need a saw and reverse
saw? or 
> would the reverse saw really be handled by the inverter?
> 
> this may be overkill for a poly-800, but would the LFO 2 be
free-running? 
> trigger per note? re-trigger on each new note in a chord? my guess
is that 
> either of the first two options would be good; the third option
would be 

> not so good.
> 
> 
> -- 
>          ...atom
> 
>   ________________________
>   http://atom.smasher.org/
>   762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
>   -------------------------------------------------
> 
>  	"The sum of intelligence on the planet is a constant.
>  	 The population is increasing."
>  		-- Cole's Axiom
>

On Tue, 17 Jul 2007, korgpolyex800 wrote:

> Also, square wave needs to be pulse width ratio adjustable.
===============

very good stuff, especially that!


-- 
         ...atom

  ________________________
  http://atom.smasher.org/
  762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
  -------------------------------------------------

 	"I tremble for my country when I reflect that God is just."
 		-- Thomas Jefferson

i think you should make sine function with a lookup table. cheers, jure

Hey Jure,

That's good advice given that the CPU is only 8 bits and has no
floating point arithmetic logic unit.

LOL

--- In korgpolyex@yahoogroups.com, "jure zitnik" <kokoon@...> wrote:

>
> i think you should make sine function with a lookup table.
> 
> cheers,
> jure
> 
> On 7/17/07, Atom Smasher <atom@...> wrote:
> >
> >   On Tue, 17 Jul 2007, korgpolyex800 wrote:
> >
> > > Also, square wave needs to be pulse width ratio adjustable.
> > ===============
> >
> > very good stuff, especially that!
> >
> > --
> > ...atom
> >
> > ________________________
> > http://atom.smasher.org/
> > 762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
> > -------------------------------------------------
> >
> > "I tremble for my country when I reflect that God is just."
> > -- Thomas Jefferson
> >
> >  
> >
>

haha yes it s pretty obvious for this case, but i think many (probably most) math libraries out there also use some kind of interpolated lookup table. although

Well, your idea got me thinking so I quickly dropped the lookup table
values into excel and created histograms of the table values and I see
a linear graph, what appears to be a logarithmic or perhaps an
exponential but no sine wave anywhere to be seen. So I wonder if the
MG is a sine in the first place. Maybe it's just a kludge.

Mike.

--- In korgpolyex@yahoogroups.com, "jure zitnik" <kokoon@...> wrote:
>
> haha yes it's pretty obvious for this case, but i think many
(probably most)
> math libraries out there also use some kind of interpolated lookup
table.
> although there are approximation functions, they are quite time
expensive.
> 
> cheers,
> jure
> 
> On 7/18/07, korgpolyex800 <korgpolyex800@...> wrote:
> >
> >   Hey Jure,
> >
> > That's good advice given that the CPU is only 8 bits and has no
> > floating point arithmetic logic unit.
> >
> > LOL
> >
> > --- In korgpolyex@yahoogroups.com <korgpolyex%40yahoogroups.com>,
"jure

> > zitnik" <kokoon@> wrote:
> > >
> > > i think you should make sine function with a lookup table.
> > >
> > > cheers,
> > > jure
> > >
> > > On 7/17/07, Atom Smasher <atom@> wrote:
> > > >
> > > > On Tue, 17 Jul 2007, korgpolyex800 wrote:
> > > >
> > > > > Also, square wave needs to be pulse width ratio adjustable.
> > > > ===============
> > > >
> > > > very good stuff, especially that!
> > > >
> > > > --
> > > > ...atom
> > > >
> > > > ________________________
> > > > http://atom.smasher.org/
> > > > 762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
> > > > -------------------------------------------------
> > > >
> > > > "I tremble for my country when I reflect that God is just."
> > > > -- Thomas Jefferson
> > > >
> > > >
> > > >
> > >
> >
> >  
> >
>

A graph of log(x) looks a little like a quarter of a sine wave... is it
possible the table is read forward then backwards, then again through an
inverter?

Tim

On Thu, 19 Jul 2007, korgpolyex800 wrote:

> Well, your idea got me thinking so I quickly dropped the lookup table
> values into excel and created histograms of the table values and I see
> a linear graph, what appears to be a logarithmic or perhaps an
> exponential but no sine wave anywhere to be seen. So I wonder if the
> MG is a sine in the first place. Maybe it's just a kludge.
>
> Mike.
>
> --- In korgpolyex@yahoogroups.com, "jure zitnik" <kokoon@...> wrote:
> >
> > haha yes it's pretty obvious for this case, but i think many
> (probably most)
> > math libraries out there also use some kind of interpolated lookup
> table.
> > although there are approximation functions, they are quite time
> expensive.
> >
> > cheers,
> > jure
> >
> > On 7/18/07, korgpolyex800 <korgpolyex800@...> wrote:
> > >
> > >   Hey Jure,
> > >
> > > That's good advice given that the CPU is only 8 bits and has no
> > > floating point arithmetic logic unit.
> > >
> > > LOL
> > >
> > > --- In korgpolyex@yahoogroups.com <korgpolyex%40yahoogroups.com>,
> "jure
> > > zitnik" <kokoon@> wrote:
> > > >
> > > > i think you should make sine function with a lookup table.
> > > >
> > > > cheers,
> > > > jure
> > > >
> > > > On 7/17/07, Atom Smasher <atom@> wrote:
> > > > >
> > > > > On Tue, 17 Jul 2007, korgpolyex800 wrote:
> > > > >
> > > > > > Also, square wave needs to be pulse width ratio adjustable.
> > > > > ===============
> > > > >
> > > > > very good stuff, especially that!
> > > > >
> > > > > --
> > > > > ...atom
> > > > >
> > > > > ________________________
> > > > > http://atom.smasher.org/
> > > > > 762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
> > > > > -------------------------------------------------
> > > > >
> > > > > "I tremble for my country when I reflect that God is just."
> > > > > -- Thomas Jefferson
> > > > >
> > > > >
> > > > >
> > > >
> > >
> > >
> > >
> >
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>

> > So I wonder if the MG is a sine in the first place. Maybe it's 
just a kludge.

- I am not sure, but here (http://www.geocities.com/diffused_light/
stuff/poly800ii.html) and somewhere elsewhere (i can not remember) i 
red, "LFO" has triangle wave ...

if i recall correctly the original poly-800 MG has triangluar waveform.. as far as cheating with log(x) goes... i think it would be better to use a lookup