Technical questions

[Flora]

ok so paul if I understand it well you did have the 680x0 assembler and convert it to x86 assembler or was it at a higher level?

On the other hand how did you manage the sound?

[Flora]

I am asking that for my personal culture as well as to understand exactly how you managed this tour de force

[Paul Stevens]

I started with the Winston file A_192.ST It
looked like this: A couple hundred thousand numbers.

48e7
0088
287a
ffda
307c
fa00
4a6c
d12c
6628
4228
001b
117c
008b
0021
117c
0008

I wrote a 68000 disassembler and found the code
and data segements more-or-less by trial and error
(no mean feat in itself) and produced:

000b4c 48e7 0088 MOVEM.L A04,-(A7)
000b50 287a ffda MOVE.L $-36 (=0x000b2c),A4
000b54 307c fa00 MOVE.W #64000,A0
000b58 4a6c d12c TST.W -11988(A4)

Then I looked at each instruction, figured out what it
did and wrote the following by hand

void TAG000b4c(UI16 P1)
{
unsigned int A0;
A0 = 0xfffffa00;
if (d.Word11988 == 0)
etc.

Then I compiled and ran it.

[Flora]

so it is:
680x0 machine code -> 680x0 assembler -> C/C++
is that correct?

[Paul Stevens]

That is correct.

[Flora]

Man it must have been a nightmare.
that is a great job anyway.

For the sounds the backengeenering worked as well or do you have to include some tricks?

[Paul Stevens]

The sounds were relatively straight-forward. I will tell
you what I think I remember. Some details may be
wrong.

The Atari had some sort of log-scale digital-to-analog
converter for the sounds. CSB had 4-bit samples which
were fed to the D-to-A converter. About all I had to do
was build a lookup table to convert the samples on the
same log scale to 8-bit samples. Then I build a wavefile
header to put on front of the samples and sent it off
to Windows, telling Windows that it was an in-memory
copy of a .WAV file.

[Flora]

yeah that was smart and nice.

Did you get any money for that? :)