Simon Williams wrote:

>Of course my original scheme didn't allow for sequences longer than 255 
>bytes... so I had to add a little routine to move the counter up by one 
>page when necessary.
>
>Well I've gotta say this Assembly stuff is really a lot of fun... now
>that I've been at it for a while it actually seems easier than basic.
>
>Anyway, the sequencer works quite nicely (considering the obvious limit-
>ations of the A2). Once I'm done working out the kinks, I'll make it 
>avaialble to the millions who have been waiting so patiently.

Cool--can't wait!

>SW
>
>PS: Michael, if you're reading this, do you know what the delay factor
>is in calling DAC.522 from ML? Not sure if the timing needs to be that
>accurate, but it can't hurt... 

The entry overhead (before starting to produce timed samples) is
45 cycles, and upon returning, 28 cycles.  This is exclusive of the
6 cycles required to execute the JSR to DAC522.

In short, the overhead is less than 70 microseconds--probably negligible.

The situation is complicated a bit by the fact that the last sample is
temporarily replaced by a $00 sample, and it is not fully played before
returning. The net effect is that your program regains control (after the
JSR) 27 cycles earlier than the final sample would have ended.

So, if you are doing precise timing based on playing a certain number
of samples, the "calling overhead" is 45 - 27, or 18 cycles, again
excluding the JSR time.  ;-)

-michael

Check out 8-bit Apple sound that will amaze you on my
Home page:  http://members.aol.com/MJMahon/