Korg Poly800/EX800 Users group photo

Yahoo Groups archive

Korg Poly800/EX800 Users

Index last updated: 2026-03-31 13:55 UTC

Thread

Bugs, 303's and sequencers

Bugs, 303's and sequencers

2008-11-12 by Michael Hawkins

This email is much longer than it should be but anyway...

The Original Poly-800 Bug
The main reason that the nasty sequencer bug is fixed in the HAWK-800 is because the original Poly-800 had a shocking MIDI implementation and also a known bug. And the HAWK-800 code is based on the EX-800 code which doesn't have the sequencer bug. Then all of the HAWK-800 features have been added into that EX-800 code. That new code includes buffering of certain MIDI commands so that there is much less chance of MIDI data causing the HAWK-800 any problems. In the next minor firmware release, I will be including the stack overflow protection that I had previously talked about. The stack protection will cause the HAWK-800 to restart if the MIDI buffers overflow. This will put a complete stop to the Poly losing patches if it gets swamped by MIDI messages.
I do not think the Poly 800 MK2 suffered from the original MK1 bug at all. If patch 66 caused a crash of a MK2 Poly, I would think it was because of out of bounds data in the patch. The MK2 is more sensitive to out of bounds data in patches because of the effects parameters.

303 Accents and the Poly Sequencer
First of all, I am not that familiar with the 303 at all (I certainly know the sounds but never used or ever touched one myself). But I know the Poly sequencer is shockingly inadequate as it is now. The right way to improve it would be to replace it entirely with a completely rewritten software sequencer module. That is something I would LIKE to do but it is not in my plans to do so as yet. Having said that, there are a couple of short term improvements to be made.

i) I plan to implement an arpeggiator with features somewhat similar to a cut down version of the Kawai K5000 arp. Obviously how much "cut down" depends upon time and my inclination to do it which depends upon musician input and how much pizza and beer is provided. :-)
ii) Route the sequencer to VCF cutoff, resonance or EG1/2/3 attack (with the sequencer notes either playing or not).
iii) Provide multiple sequencer patterns that can then be sequenced in a user set order.

Now the big issue here, is that the ARP will take a lot of work and I have it as the highest priority of things to do. But the question should be directed to HAWK-800 owners as to what they would rather see first. ARP or better sequencer? Since I think using a Poly as a sequencer is just plain silly ( :-) ), that is why I chose to work on the ARP first. I do know that I would like to have points ii and iii though because they would make a really bad sequencer - well - a little less bad.

Of course, it drove me nuts to discover just how bad the sequencer was when I bought my first Poly back in 1984. And the same "drive me nuts" problem is that you cannot hold a note or notes over while dropping others. The ability to hold notes and not others is such a fundamental feature of a polyphonic sequencer. However, I respect the Korg engineers that built the Poly because they were working to build a budget synth and they simply ran out of space to fit features into the one 8K ROM. The annoying bit is that I would have to rewrite the entire sequencer in order to fix just that one thing so I don't see this being addressed until I one day get to rewrite the whole sequencer from scratch.

Anyway, the bottom line is ARP or sequencer features - which to do first?

Mike

From: Alx <the_voder@...>
To: korgpolyex@yahoogroups.com
Sent: Tuesday, November 11, 2008 4:54:23 AM
Subject: [korgpolyex] Re: Bug

--- In korgpolyex@yahoogro ups.com, Russ <russdaren@. ..> wrote:
>
> I found out earlier this year (and had it confirmed by others on this board) that sending too
much midi data can crash a poly 800. Like sending 1 minute's worth of note/pitchwheel in 1
second.It would wipe out your patches!
> possibly similar?
>

It's a well-known issue with both the Poly-800 and EX-800. Very annoying. It's a major
reason I was excited about the Hawk-800 mod, in fact. I was never able to use either my P-
800 or EX-800 under sequencer control, because it would periodically crash. Not only that,
when it did crash, it would overwrite parts of the memory with random data, so I'd end up
losing all my patches. Happily, after installing the mod, it seems perfectly stable, AND I can
now tweak parameters from my XStation controller while it plays, which is a bit of a
revelation!

a|x


Re: [korgpolyex] Bugs, 303's and sequencers

2008-11-12 by Alex Drinkwater


On 12 Nov 2008, at 14:00, Michael Hawkins wrote:

This email is much longer than it should be but anyway...

The Original Poly-800 Bug
The main reason that the nasty sequencer bug is fixed in the HAWK-800 is because the original Poly-800 had a shocking MIDI implementation and also a known bug. And the HAWK-800 code is based on the EX-800 code which doesn't have the sequencer bug. Then all of the HAWK-800 features have been added into that EX-800 code. That new code includes buffering of certain MIDI commands so that there is much less chance of MIDI data causing the HAWK-800 any problems. In the next minor firmware release, I will be including the stack overflow protection that I had previously talked about. The stack protection will cause the HAWK-800 to restart if the MIDI buffers overflow. This will put a complete stop to the Poly losing patches if it gets swamped by MIDI messages.
I do not think the Poly 800 MK2 suffered from the original MK1 bug at all. If patch 66 caused a crash of a MK2 Poly, I would think it was because of out of bounds data in the patch. The MK2 is more sensitive to out of bounds data in patches because of the effects parameters.

303 Accents and the Poly Sequencer
First of all, I am not that familiar with the 303 at all (I certainly know the sounds but never used or ever touched one myself). But I know the Poly sequencer is shockingly inadequate as it is now. The right way to improve it would be to replace it entirely with a completely rewritten software sequencer module. That is something I would LIKE to do but it is not in my plans to do so as yet. Having said that, there are a couple of short term improvements to be made.

OK, I have an idea:
If Mike can add NRPNs for all sound parameters, I can have a mess around and try and work out which parameters could be altered to make those accent sounds. The current MIDI controller setup is great for live sound tweaking, where you're only editing one parameter at a time, but is less easy to use when controlling parameters from a sequencer, so I'm still holding out for those NRPNs at the moment. Once that is in place, I can try some different controller tweaks, and see what works best, and post some examples here, and we can decide if it's worth pursuing some kind of accent feature, whether under local sequencer control, or as another play mode.

i) I plan to implement an arpeggiator with features somewhat similar to a cut down version of the Kawai K5000 arp. Obviously how much "cut down" depends upon time and my inclination to do it which depends upon musician input and how much pizza and beer is provided. :-)

I've never really been into arpeggiators, personally, but I know many people are. I can see it's a more obvious 'performance feature', mind you.

ii) Route the sequencer to VCF cutoff, resonance or EG1/2/3 attack (with the sequencer notes either playing or not).

That would be cool, but would it actually be possible to have the sequencer control ANY sound parameter? I actually think it would then make the sequencer a really inspiring tool for loop-making, rather than the rather sad afterthought it currently is.
If it only going to be able to control a selection of parameters, I'm not sure I'd list env attack as one of them. Off the top of my head, these are the parameters I'd want to be able to control:
Filter cutoff
Filter resonance
Filter env. decay
Filter env. attack (maybe)
One of more of the LFOs (rate, mod amount)
DCO1/2 Octave
DCO1/2 hormonics
Slide (I'd still want to be able to do this, whether it meant being able to overlap notes, or needed a special slide flag to be set for that step)

iii) Provide multiple sequencer patterns that can then be sequenced in a user set order.

I was thinking of this too. Maybe patterns could be switched from the keyboard.

Now the big issue here, is that the ARP will take a lot of work and I have it as the highest priority of things to do. But the question should be directed to HAWK-800 owners as to what they would rather see first. ARP or better sequencer? Since I think using a Poly as a sequencer is just plain silly ( :-) ), that is why I chose to work on the ARP first. I do know that I would like to have points ii and iii though because they would make a really bad sequencer - well - a little less bad.

Of course, it drove me nuts to discover just how bad the sequencer was when I bought my first Poly back in 1984. And the same "drive me nuts" problem is that you cannot hold a note or notes over while dropping others. The ability to hold notes and not others is such a fundamental feature of a polyphonic sequencer. However, I respect the Korg engineers that built the Poly because they were working to build a budget synth and they simply ran out of space to fit features into the one 8K ROM. The annoying bit is that I would have to rewrite the entire sequencer in order to fix just that one thing so I don't see this being addressed until I one day get to rewrite the whole sequencer from scratch.

Anyway, the bottom line is ARP or sequencer features - which to do first?

From my point of view:
well, it's dodging the question, but I'd like NRPNs before either. If I'm going to use it with a sequencer, then I can use my sequencer's arpeggiator, and if I'm going to be driving it from a sequencer, I'd like also to be able to change all the synth parameters with their own dedicated controllers. Don't get me wrong though; I think both the arpeggiator and the updated sequencer would be cool. I'd say, if I can have my slidey/accenty 303-style sequencer (or something that allows me to do something similar by altering certain parameters per-step, and having slides between certain steps), then I'd go for the sequencer as a priority. Otherwise, I'd plump for the arpeggiator, because you're obviously more excited about working on it, and I think more Hawk-800 users out there would want that, to be honest.


a|x




Mike

From: Alx <the_voder@...>
To: korgpolyex@yahoogroups.com
Sent: Tuesday, November 11, 2008 4:54:23 AM
Subject: [korgpolyex] Re: Bug

--- In korgpolyex@yahoogro ups.com, Russ <russdaren@. ..> wrote:
>
> I found out earlier this year (and had it confirmed by others on this board) that sending too
much midi data can crash a poly 800. Like sending 1 minute's worth of note/pitchwheel in 1
second.It would wipe out your patches!
> possibly similar?
>

It's a well-known issue with both the Poly-800 and EX-800. Very annoying. It's a major
reason I was excited about the Hawk-800 mod, in fact. I was never able to use either my P-
800 or EX-800 under sequencer control, because it would periodically crash. Not only that,
when it did crash, it would overwrite parts of the memory with random data, so I'd end up
losing all my patches. Happily, after installing the mod, it seems perfectly stable, AND I can
now tweak parameters from my XStation controller while it plays, which is a bit of a
revelation!

a|x



Re: [korgpolyex] Bugs, 303's and sequencers

2008-11-12 by Atom Smasher

On Wed, 12 Nov 2008, Michael Hawkins wrote:

> i) I plan to implement an arpeggiator with features somewhat similar to
> a cut down version of the Kawai K5000 arp. Obviously how much "cut down"
> depends upon time and my inclination to do it which depends upon
> musician input and how much pizza and beer is provided. :-)
================

http://optimolch.de/jens.groh/K5000/GregWaltzer/egw/equipment/k5000arp.htm

that sounds... ambitious. the more of a good job you do implementing it,
the more i'm afraid of how bad the UI will be ;)


> ii) Route the sequencer to VCF cutoff, resonance or EG1/2/3 attack (with
> the sequencer notes either playing or not).
==================

overkill. see below about the sequencer.

as long as velocity can be routed to amps, filters, etc (envelope
parameters would be nice, but IMHO not crucial) then a *reasonable* 303
emulation can be done with the hawk-800. *BUT* the hawk-800 will never
~really~ emulate a 303 because, among other things, the filters are too
different. so... i say do what's reasonable, but don't go overboard trying
to turn the hawk-800 into a 303 emulator.... make it do some of the 303
tricks (apparently it already does!) and support for a velocity/accent
(which i think it can do?), but then let people use it for the 2 DCO,
2/4-pole DCF, weird ENVs, 4(?!?!) LFOs, 21st century firmware/hardware
modified 80s geek synth that it is. the world does *not* need another 303
emulator.

to emulate a 303: all non-accent notes have a velocity 1. all accent notes
have a velocity 127. edit a patch so the velocity makes it a little
louder, opens up the filter a bit, and if it implemented, shorten the
filter-env attack (or something like that, it's been a while since i
studied the 303 schematics).

personally, i don't want a 303 emulator, as such. it would be nice to have
those features available so i could use it as a hawk-800 that's just that
much funkier, and be able to program ~other~ "accent" and/or slide style
tricks. ya know, not the kind of things where people hear it and say "that
sounds almost like 303", but they'd say "holy shit! that has a 303 kind
groove, but how the hell did he do that??"


> iii) Provide multiple sequencer patterns that can then be sequenced in a
> user set order.
===================

no comment. see below.


> Now the big issue here, is that the ARP will take a lot of work and I
> have it as the highest priority of things to do. But the question should
> be directed to HAWK-800 owners as to what they would rather see first.
> ARP or better sequencer? Since I think using a Poly as a sequencer is
> just plain silly ( :-) ), that is why I chose to work on the ARP first.
> I do know that I would like to have points ii and iii though because
> they would make a really bad sequencer - well - a little less bad.
======================

i'd rather see a good arp, and don't care about a sequencer. but... maybe
an arp can be a mini-sequencer too?

good hardware or software sequencers are easy to find and cheap. the same
cannot be said for good arpeggiators. my logic, then, is that nearly
anyone with the resources to get a hawk-800 should have the resources to
find a suitable external sequencer. the same does not apply to a good
arpeggiator. so (IMHO) the hawk-800 should skip the sequencer and focus on
the arpeggiator.

maybe a good feature that would blur the line between a mini-sequencer and
an arpeggiator would be this: enter a series of notes that can be played
back entered. program arpeggiator functions. then (per "sequence") assign
a value that determines what percentage of the notes are arpeggiated.
so... select a "arp" value of zero and the notes play back as entered;
select an "arp" value of 99 and all of the notes are arpeggiated; select
an "arp" value of 50 and each note has a 50% chance of either playing as
entered or being arpeggiated. i'm not sure if the hardware can handle that
(we've already discussed the random generation issues of the CPU) but it
might solve the "arpeggiator or sequencer" question.


> Anyway, the bottom line is ARP or sequencer features - which to do
> first?
=================

arp.


--
...atom

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

"Non-violence leads to the highest ethics, which is the
goal of all evolution. Until we stop harming all other
living beings, we are still savages."
-- Thomas A. Edison

Re: [korgpolyex] Bugs, 303's and sequencers

2008-11-13 by Alex Drinkwater

Ok, if the arp is anything like as sophisticated as the K5000 one then I'd happily go for that as a priority over the sequencer. As long as I can still have my slide notes and accents. The arp on the Novation Nova might also be a good, simpler model. It's somewhere between an arp and a sequencer too, in that it allows different gate-lengths, slides and accents to be preprogrammed, while still acting like an arpeggiator.

I still want my NRPNs though,and velocity mappable to various things.

a|x


--- On Wed, 12/11/08, Atom Smasher <atom@...> wrote:

> From: Atom Smasher <atom@...>
> Subject: Re: [korgpolyex] Bugs, 303's and sequencers
> To: korgpolyex@yahoogroups.com
> Date: Wednesday, 12 November, 2008, 11:58 PM
> On Wed, 12 Nov 2008, Michael Hawkins wrote:
>
> > i) I plan to implement an arpeggiator with features
> somewhat similar to
> > a cut down version of the Kawai K5000 arp. Obviously
> how much "cut down"
> > depends upon time and my inclination to do it which
> depends upon
> > musician input and how much pizza and beer is
> provided. :-)
> ================
>
> http://optimolch.de/jens.groh/K5000/GregWaltzer/egw/equipment/k5000arp.htm
>
> that sounds... ambitious. the more of a good job you do
> implementing it,
> the more i'm afraid of how bad the UI will be ;)
>
>
> > ii) Route the sequencer to VCF cutoff, resonance or
> EG1/2/3 attack (with
> > the sequencer notes either playing or not).
> ==================
>
> overkill. see below about the sequencer.
>
> as long as velocity can be routed to amps, filters, etc
> (envelope
> parameters would be nice, but IMHO not crucial) then a
> *reasonable* 303
> emulation can be done with the hawk-800. *BUT* the hawk-800
> will never
> ~really~ emulate a 303 because, among other things, the
> filters are too
> different. so... i say do what's reasonable, but
> don't go overboard trying
> to turn the hawk-800 into a 303 emulator.... make it do
> some of the 303
> tricks (apparently it already does!) and support for a
> velocity/accent
> (which i think it can do?), but then let people use it for
> the 2 DCO,
> 2/4-pole DCF, weird ENVs, 4(?!?!) LFOs, 21st century
> firmware/hardware
> modified 80s geek synth that it is. the world does *not*
> need another 303
> emulator.
>
> to emulate a 303: all non-accent notes have a velocity 1.
> all accent notes
> have a velocity 127. edit a patch so the velocity makes it
> a little
> louder, opens up the filter a bit, and if it implemented,
> shorten the
> filter-env attack (or something like that, it's been a
> while since i
> studied the 303 schematics).
>
> personally, i don't want a 303 emulator, as such. it
> would be nice to have
> those features available so i could use it as a hawk-800
> that's just that
> much funkier, and be able to program ~other~
> "accent" and/or slide style
> tricks. ya know, not the kind of things where people hear
> it and say "that
> sounds almost like 303", but they'd say "holy
> shit! that has a 303 kind
> groove, but how the hell did he do that??"
>
>
> > iii) Provide multiple sequencer patterns that can then
> be sequenced in a
> > user set order.
> ===================
>
> no comment. see below.
>
>
> > Now the big issue here, is that the ARP will take a
> lot of work and I
> > have it as the highest priority of things to do. But
> the question should
> > be directed to HAWK-800 owners as to what they would
> rather see first.
> > ARP or better sequencer? Since I think using a Poly as
> a sequencer is
> > just plain silly ( :-) ), that is why I chose to work
> on the ARP first.
> > I do know that I would like to have points ii and iii
> though because
> > they would make a really bad sequencer - well - a
> little less bad.
> ======================
>
> i'd rather see a good arp, and don't care about a
> sequencer. but... maybe
> an arp can be a mini-sequencer too?
>
> good hardware or software sequencers are easy to find and
> cheap. the same
> cannot be said for good arpeggiators. my logic, then, is
> that nearly
> anyone with the resources to get a hawk-800 should have the
> resources to
> find a suitable external sequencer. the same does not apply
> to a good
> arpeggiator. so (IMHO) the hawk-800 should skip the
> sequencer and focus on
> the arpeggiator.
>
> maybe a good feature that would blur the line between a
> mini-sequencer and
> an arpeggiator would be this: enter a series of notes that
> can be played
> back entered. program arpeggiator functions. then (per
> "sequence") assign
> a value that determines what percentage of the notes are
> arpeggiated.
> so... select a "arp" value of zero and the notes
> play back as entered;
> select an "arp" value of 99 and all of the notes
> are arpeggiated; select
> an "arp" value of 50 and each note has a 50%
> chance of either playing as
> entered or being arpeggiated. i'm not sure if the
> hardware can handle that
> (we've already discussed the random generation issues
> of the CPU) but it
> might solve the "arpeggiator or sequencer"
> question.
>
>
> > Anyway, the bottom line is ARP or sequencer features -
> which to do
> > first?
> =================
>
> arp.
>
>
> --
> ...atom
>
> ________________________
> http://atom.smasher.org/
> 762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
> -------------------------------------------------
>
> "Non-violence leads to the highest ethics, which is
> the
> goal of all evolution. Until we stop harming all other
> living beings, we are still savages."
> -- Thomas A. Edison
>
>
>
> ------------------------------------
>
> Yahoo! Groups Links
>
>
>