search

Wohlstand / libOPNMIDI

A Software MIDI Synthesizer library with OPN2 (YM2612) emulator
GNU Lesser General Public License v3.0
96 stars 6 forks source link

add MAME OPNA #76

Closed jpcima closed 5 years ago

jpcima commented 5 years ago

This adds OPNA from MAME fm.cpp which supports build of different chips with common base. It's separate from fm2612, but it's notable it has also a support of YM2612 and several others.

Wohlstand commented 5 years ago

Okay, I have fixed a bug of CMake script and I have fixed C++98 compatibility of given code, so, it must be fine now. Anyway, what about rhythm ROM? Does MAME pre-icludes them, or need to pass them externally?

Wohlstand commented 5 years ago

So, I'll gonna to put this chipset into OPN2-BE, and then backport CMake-related changes from libADLMIDI.

Wohlstand commented 5 years ago

Just now I have checked out, PSG won't work on MAME, but works fine on NP2 (the implementation on a side branch where I am able to play PSG, but, shitty nosense on the side of my usage implementation, but it's able to produce sound)

jpcima commented 5 years ago

Anyway, what about rhythm ROM? Does MAME pre-icludes them, or need to pass them externally?

In MAME's original 2608intf.cpp the ROM is externally provided. I just made the bare-bones replacement needed for operation, and I/O callbacks are dummied, so rhythm functionality is expected not to work. (it's maybe the reason for PSG not working as well?)

ROM_LOAD16_WORD( "ym2608_adpcm_rom.bin", 0x0000, 0x2000, CRC(23c9e0d8) SHA1(50b6c3e288eaa12ad275d4f323267bb72b0445df) )

Wohlstand commented 5 years ago

About of rhythm rom: Let's make from that rom the .h file and load as memory block? @Papiezak have referred an OPNA emulator that contains ROM as C headers.

Wohlstand commented 5 years ago

About of PSG: I may guess, some code files are missing and PSG is implemented on another side of MAME...

jpcima commented 5 years ago

About of rhythm rom: Let's make from that rom the .h file and load as memory block?

I think pmdwin does it in this way. It's no problem to do the same. More importantly, I should study how MAME makes the interaction work in the 2608intf source.

About of PSG: I may guess, some code files are missing and PSG is implemented on another side of MAME...

From MAME code 2068intf:

class ym2608_device : public ay8910_device,
        public device_rom_interface

Probably I should add this emulation: https://fr.wikipedia.org/wiki/AY-3-8910

jpcima commented 5 years ago

In mame-2608 I committed some progress on the PSG part that is yet to complete. Arranged some parts which tie to much with the MAME core, so not to need pulling in dozens of sources. https://github.com/Wohlstand/libOPNMIDI/compare/mame-2608...jpcima:mame-2608

The PSG connects to ym2608's device by callback interface.

This annoying part is left : PSG has to resample its output. This device generates at its own sample rate that is 1/8 clock rate, for around 1 MHz.

In MAME it has an internal mixer which accepts multi-channel at different rates. For completion it's needed that I pull in more of MAME code. Resampling is located at src/emu/sound.cpp and relevant code is sound_stream::generate_resampled_data.