Archive of the former Yahoo!Groups mailing list: Korg Poly800/EX800 Users

previous by date index next by date
previous in topic topic list next in topic

Subject: Re: [korgpolyex] Velocity sensitive

From: Michael Hawkins <korgpolyex800@...>
Date: 2009-01-10

the problem with the slow CPU is that the onset of notes tends to slow down as you add more features.

A good example is the four LFO's where once there was only one.

The original Poly only had to initialize one LFO. Now it has to initialize four.

The initialization of the LFO's (when they're not free running) has to occur at note down onset.

The code is small but when you multiply it by four from its original, it is going to have an impact.

Then, on top of that, there are numerous new features implemented in the Poly that also require initialization at onset of notes.

FM-800 EG follow and MG, Resonance follow and MG, to name just a few.

I've spent a great deal of time optimising the code for speed and size.

Then on top of that, the envelopes are software driven too. There is a large amount of mathematics used to create those ADBSSRR envelopes.

All of the math has to be able to be executed within a specific amount of time so that the interrupt service routine clock keeps accurate time. Otherwise, the envelopes would stretch or shrink according to the complexity of the programmed envelope.

I DON'T like the idea of using a heatsink on our Poly's because that implies that your batteries will go flat in no time at all.

One of the best features of the Poly is that you can walk the stage with it.

Whatever we do from this point forward must be able to be done without any additional hardware modification.

Mike

From: Alec Jahn <ocedtotehmax@...>
To: korgpolyex@yahoogroups.com
Sent: Saturday, January 10, 2009 5:03:47 AM
Subject: Re: [korgpolyex] Velocity sensitive

I'm only a novice at architecture (programmed gameboys in my day.... last year).  More cycles per second is great, but would that even solve the problem?  I can understand the simplicity of getting more done in the same ammount of time...  Is there anything in the circuit that relies on that particular/steady clockspeed (interrupts were mentioned) (I'm thinking of old games and the audio processor on (again...) the gameboy, a "four voice" analog synth basically - when the clockspeed was increased, IE playing a game on a Super Gameboy or some newer-than-the- old-grey- brick-models, the game would technically run slightly faster and the audio would be that "percent" higher in pitch... amongst some other minor glitches... this was just a ~2% increase or so, not 33% or higher as our possible suggestion/max) .  Could this be done through the HAWK mods, or would it have to be a separate addition?
Then I ask, (again, noobness showing), what controls the clock in our 800s?  Something as simple as the FSB on a normal PC?
 
Also, could you just swap out the chip itself for a better one - a simple "upgrade"?  I wouldn't think it'd be that easy, yet I'm unfamiliar with these 'vintage' architectures in comparison to stuff from the last decade.
 
I do like the thought of having a little heatsink in my synth if needed.  I think I'll fashion one anyway, a nice bright copper one.
 


 
On Sat, Jan 10, 2009 at 3:42 AM, Alexandre Souza <alexandre-listas@ e-secure. com.br> wrote:

> OVERCLOCK THAT SOB!

Or put a faster processor. There are faster 8085s around, the 800'one is
running at 6MHz, it can be run at 8 MHz and there are faster (but harder to
find) versions. What about a Z80 in its place?