Suggestion for ymfm variant/emulator

[tobybear]

The ymfm library contains a variety of enulated FM chips, including the OPL2/OPL3, and is under a BSD-3 license: https://github.com/aaronsgiles/ymfm
Performance and compatibility is good afaik, however the code needs a C++14 compliant compiler.

There is however a single-header port of the OPL part of ymfm back to pure C here: https://github.com/mattiasgustavsson/libs/blob/main/opl.h
This header also includes MIDI playback routines from the DOSMidi project that we would not need, but maybe the OPL2/3 part could be an easy to integrate alternative backend for ADLMIDI?

[Wohlstand]

Hello
I applied it recently (at the CMake I did the check for C++14 support, and if target compiler doesn't suppot that, then YMFM will be disabled), and even more I implemented support of OPL2 (with 9 voices and mono), and, according to benchmark, YMFM aren't so super effective like DosBox emulators, but faster than Nuked.

[Wohlstand]

This header also includes MIDI playback routines from the DOSMidi project that we would not need, but maybe the OPL2/3 part could be an easy to integrate alternative backend for ADLMIDI?

Speaking about MIDI driver in general, I thinking to completely redo it. My current MIDI driver which is derivative from ADLMIDI, is flexible and functional, and of course I should preserve all the features I and @jpcima did, but get rid of all Bisqwit's origin and that should allow me to switch license to LGPL without waiting when Bisqwit will consume my pull-requests that he requested to grant LGPL. I thinking to turn the core part into pure-C to allow building hardware drivers targeting to legacy systems, and I will keep the C++ part that works with emulators and implements the audio streaming.