I tried checking for clipping at all stages, then I realized it was mixing with a 32bit int, so that wouldn't have made a difference - but I ran the new code anyways, and sure enough, it made no difference. I've seen this before, I think its coming from the psg/ssg part of the ay8910 - in some games (like the games from the new(to FBA)-UPL driver iq_132 had recently wrote), the ssg sound channel was about 5x louder than the fm parts (Atomic Robo Kid, Mutant Night), somehow this had been fixed in MAME when they re-wrote the entire 8910 code - at this point it would be impossible to graft in a fix, so I took the ez route and simply lowered the psg(ssg) channels to match the fm ones. I can't see doing it for Metal Slug 5 though, the crackle of the gun actually sounds pretty good/non-irritating to me, even with headphones. Thanks for reporting it, though best regards,- dink
Since I added support for the independent sound routes you can reduce the volume of each channel individually.I can't test at the minute but I always assumed the gun sound was ADPCM, so not related to the AY part anyway?
not sure if it applies to Metal Slug 5 because I'm not hearing it, but some of the later NeoGeo games were clearly developed using an inaccurate emulator (probably NeoRageX) as part of the development.As a result Nightmare in the Dark has a really annoying crackle when you jump, it didn't happen in very old versions of MAME, and it didn't happen in NeoRageX, but it did happen once we started emulating actual limits of the chip properly, and it does happen on real hardware (so we emulate it) Not respecting the chip limits actually breaks some other games in worse ways!
64 bits version is stamped 0.2.97.32 ^^
The explosion is the only sample. The other sound should be emulated and working.....