Re: Carte Blanche

[sicklittlemonkey <Nick.Westgate@gmail.com>]

On Apr 4, 7:47 am, "Michael J. Mahon" <mjma...@aol.com> wrote:
> Nick, to capture the bitstream it is only necessary to continue
> to shift read data into the shift register continuously, then
> sample it (by loading it and storing it into memory) often enough
> that no bits are unobserved.  The Apple's speed is sufficient to
> do this without a problem.

Yes, but I was trying to do this at better than 4-CPU-cycle resolution
to reliably capture time-sensitive parts of the stream (including sync
nibbles of course).

The best results I had were at the highest resoltion possible: 2
sequencer clocks = 1 CPU clock. The read pulise line is shifted into
the latch at this rate, which means a full latch every 8 CPU clocks.
So load, store, load, store ... ;-)

That's what I meant about bandwidth and memory. There's not enough
main memory to store an entire track's this high resolution data, and
the CPU is too busy to do anything else. Even on the GS. Running this
data through an emulation of the normal sequencer looks good though.

Cheers,
Nick.