Closed GoogleCodeExporter closed 9 years ago
I do not have a Mega CD anymore to test it but i think people have reported it
happens with the "real thing" as well.
That's the first thing I noticed when trying the enhanced disc too but I
checked and there is no bug with PCM emulation, seems like it's a bug with the
game programming instead. Very likely, they did not tested with the stereo
cable connected ?
Original comment by ekeeke31@gmail.com
on 17 Oct 2012 at 5:10
here we go:
http://www.piersolar.com/board/viewtopic.php?f=30&t=1597
I would like confirmation from real PS + Mega CD owner though as it could be
working fine on different hardware revisions, which would be interesting to
know.
Original comment by ekeeke31@gmail.com
on 17 Oct 2012 at 5:18
And about "filename manipulations in load routines", if you look at the code,
you can see that 3 extension chars is only used when detecting SG-1000 (sg),
Master System (sms) & Game Gear (gg) ROM files (and eventually Genesis MDX
files which need to be treated differently). Anything else would be considered
as a Genesis ROM file, no matter of the extension, so there isn't any potential
problem that I know of ...unless a new "preservation group" invents new ROM
extensions for these systems as well.
Off course, it's up to the platform-specific "load_archive" function to make
sure the data copied to the extension buffer is a 3 characters string. Copying
everything from the . character to the end of the file would be a very bad idea
indeed but I assume good devs don't do that ^^
Both libretro and GX port does the following so they are safe:
memcpy(extension, &filename[strlen(filename) - 3], 3);
Original comment by ekeeke31@gmail.com
on 17 Oct 2012 at 5:31
thanks for quick reply and explantation, as always )
Original comment by 0vet...@gmail.com
on 17 Oct 2012 at 6:01
[deleted comment]
Actually, I had another look to this issue and it appears that there should be
*some* PCM sound getting out the right channel (according to the previous
thread link) but there is absolutely NONE with the emulator.
The reason is that the game does NOT initialize the PAN registers for the PCM
channels, at least not completely (for the title screen, channel 3 panning is
set to left & right, then channel 0 panning is set to the left only and only
channels 0 & 1 outputs are enabled, resulting in channel 0 being output to left
channel only and channel 1 output nowhere).
My hypothesis is that on reset state, PCM chip registers are all set to $ff,
which disables all channels outputs but also initializes the panning to left &
right by default for all channels.
It turns out that doing this when resetting the chip makes PCM sound outputting
correctly to left & right (but indeed a little more balanced to the left
because one or two channels are accidentally or deliberately panned to this
side only).
Should now be fixed in r724
Original comment by ekeeke31@gmail.com
on 17 Oct 2012 at 9:07
Original issue reported on code.google.com by
0vet...@gmail.com
on 17 Oct 2012 at 3:42