Did anyone ever get Paperboy GS to work on MAME or any of the KEGS-based emulators? The latter doesn't really get passed the title page, and in MAME, it crashes when certain sounds are triggered (like breaking glass). It feels like a sound or timing issue.

I've tried all the copies floating around the net, plus I downloaded the Woz-a-day version, cracked it myself and tried it in KEGS and GSPlus, and still no joy. Any thoughts?

I confirm that the sound timing is the issue in MAME. The game is waiting for the sound register to return a value in a specific range, which currently never occurs. I will see if I can work out why.

In article <c5a39a95-f2b6-429e-aeb7-38c53fb5d96fn@googlegroups.com>,
qkumba <peter.ferrie@gmail.com> wrote:
>I confirm that the sound timing is the issue in MAME. The game is
>waiting for the sound register to return a value in a specific range,
>which currently never occurs. I will see if I can work out why.

Ed Armstrong emailed me Feb 22 about KEGS not working, and it was an easy
fix (for KEGS).

Here's what I said:

I looked into running it on KEGS. At the title screen, Paperboy is
constantly reading the DOC last-sampled-played register, and is waiting
for certain values to occur. As an optimization 25 years ago, KEGS only
updates the DOC last-sample-played register at 60 times a second (to
basically make it work for DOCVU, and things like it). And KEGS code
had a comment saying this should be fixed.

So I fixed it, and then it seems to play. I'm quite bad at the game,
but I threw lots of papers, and even got 250 points eventually, and it
didn't hang.

I get the same results on the two .2MG files, and the .PO file. KEGS
does not work with the .woz file, I'll need to look into that.

If you want to patch KEGS.1.16 and see for yourself, modify sound.c at
lines 1776-1779, and change it from:

---
case 0x3: /* data register */
/* HACK: make this call sound_play sometimes */
g_doc_saved_val = rptr->last_samp_val;
break;
---

to

---
case 0x3: /* data register */
dsamps = dcycs * g_dsamps_per_dcyc;
sound_play(dsamps);
g_doc_saved_val = rptr->last_samp_val;
break;
---

Just add the "dsamps =" and sound_play() lines, and then recompile.

You can probably make a similar patch to GSPort.

On Sunday, February 26, 2023 at 1:57:30 PM UTC-5, Kent Dickey wrote:
> In article <c5a39a95-f2b6-429e...@googlegroups.com>,
> qkumba <peter....@gmail.com> wrote:
> >I confirm that the sound timing is the issue in MAME. The game is
> >waiting for the sound register to return a value in a specific range,
> >which currently never occurs. I will see if I can work out why.
> Ed Armstrong emailed me Feb 22 about KEGS not working, and it was an easy
> fix (for KEGS).
>
> Here's what I said:
>
> I looked into running it on KEGS. At the title screen, Paperboy is
> constantly reading the DOC last-sampled-played register, and is waiting
> for certain values to occur. As an optimization 25 years ago, KEGS only
> updates the DOC last-sample-played register at 60 times a second (to
> basically make it work for DOCVU, and things like it). And KEGS code
> had a comment saying this should be fixed.
>
> So I fixed it, and then it seems to play. I'm quite bad at the game,
> but I threw lots of papers, and even got 250 points eventually, and it
> didn't hang.
>
> I get the same results on the two .2MG files, and the .PO file. KEGS
> does not work with the .woz file, I'll need to look into that.
>
> If you want to patch KEGS.1.16 and see for yourself, modify sound.c at
> lines 1776-1779, and change it from:
>
> ---
> case 0x3: /* data register */
> /* HACK: make this call sound_play sometimes */
> g_doc_saved_val = rptr->last_samp_val;
> break;
> ---
>
> to
>
> ---
> case 0x3: /* data register */
> dsamps = dcycs * g_dsamps_per_dcyc;
> sound_play(dsamps);
> g_doc_saved_val = rptr->last_samp_val;
> break;
> ---
>
> Just add the "dsamps =" and sound_play() lines, and then recompile.
>
> You can probably make a similar patch to GSPort.

Yes, can confirm that this fixes the issue. It also helps with some other titles as well (more on this later).

Now we're just desperately trying to get a working exe on Windows and we'll be good!

MAME was dropping the fractional part of the phase from the accumulator, so some parts of the sample weren't playing.
I've proposed a fix which is waiting for approval or refinement.