[sdiy] MIDI too fast?

[ASSI]

On Tuesday 15 June 2010, grant at musictechnologiesgroup.com wrote:
> I was thinking even a queue would have helped in the JX3P case. 
> I'm a big fan of the queue, even if it takes slightly more
> code/time to execute rather than try and do everything in the ISR
> as was all the rave back then (nothing like calling a huge
> subroutine from an ISR, to mess things up). It's possible to give
> enough priority to emptying the queue that it doesn't become
> audible. Certainly an order of magnitude better than having to
> pace MIDI events by milliseconds.

The Poly-800 overwrites system memory and patch storage if you swamp 
it with MIDI bytes...

-------Michael Hawkins <korgpolyex800 at yahoo.com>
Believe it or not, there is no MIDI buffer in the MK1 or MK2 at all.

Incoming bytes are handled as they arrive. That works ONLY if the rate 
of input is slower than the Poly can handle the MIDI interrupt.

Many new pieces of MIDI equipment can definitely send MIDI faster than 
the Poly can handle it.

So if there is no buffer then how does it handle the messages?

Well, each time a MIDI byte arrives the Poly gets an interrupt. The 
hardware interrupt causes the CPU to save its 8 registers on the 
stack.

Thus, the stack grows and grows and grows until it reaches right down 
into the top of the patch storage.


The MK2 is susceptible to the same problem but since there is an empty 
memory space above the patch storage and below the stack, a MK2 simply 
crashes instead of losing patches.
-------


Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Factory and User Sound Singles for Waldorf rackAttack:
http://Synth.Stromeko.net/Downloads.html#WaldorfSounds