Re: Moving binary files to /RAM

["Laine Houghton" <laine@intergate.com>]

"Simon Williams" <email@DELETE_THIS.luddite.no-ip.com> wrote in message 
news:1gr5zon.d2hs7hoiep0yN%email@DELETE_THIS.luddite.no-ip.com...
> On a tangent, but still related:
>
> Just so you know why this matters, I'll explain what I'm up to: I'm
> trying to put together another drum sequencing program using DAC522 a la
> TIMELORD. I thought that by using /RAM to store the samples I could make
> use of more/bigger samples thus extending the programs capabilities
> however:
>
> I have 16 samples named "A"..."P", each 3072 bytes copied to
> /RAM/SAMPLES/
>
> I'm using this code to play them in order:
>
> 160 FOR F = 65 TO 80
> 170 PRINT CHR$(4)"BLOAD /RAM/SAMPLES/" CHR$(F) ",A$2000"
> 180 CALL 35840 : REM DAC522
> 190 NEXT F
>
> What happens, oddly enough, is that the first 12 samples play
> back-to-back as one would hope, but then there is a noticeable delay
> between the last four. I've tried this on a IIe, IIgs and IIc+ with the
> same result...
>
> Any ideas...?

 I did a test saving  blocks of 1's 2's 3's......... 3072 bytes just as the 
program would do. The 12's representing the 12'th file ends right at $BFFF
The 13'th to 16'th files would be contained in the bank 1 bank 2 and beyond 
memory.

I'm guessing that the bank switching overhead introduces an unavoidable 
delay that a program such as yours is sensative to.

A good incentive to get down to the assy work. With a 1 meg card you could 
move 12 files worth of samples to successive 64k banks.

If I did my napkin math correctly that would be 180 files and it should work 
as smooth as the 12 alone.

Laine...