NandemoFM v1.10 released

Nice!

warau: there's no need. It's just for fun!

Great follow-up to NandemoSCC!

Aw, didn't work with Hype or Moon Patrol

But it did work with Ye Ar Kung Fu, and it was amazing

Edit: Even recorded it on video!
https://www.youtube.com/watch?v=8ClK5xyQ8wQ

Judging by how much better it sounds with both PSG and OPLL playing together, I'd suggest creating a tool to add PSG sound to FM-only games.

Or even better, a tool that added SCC sound to FM music. This way there could be no chance of PSG conflict, and it would dramatically improve FM sound.

Very inventive stuff, I love it

I have success with Out-Run (MegaROM Cartridge) with NandemoFM v1.2 converter on PC.
To convert correctly I use this configuration:
Auto Start ticked
Original ROM Type= Megarom (ASCII8 Mapper)
BANK1 Control=Fixed to 4000H
PSG Noise=Force On
OPLL Data=Copy from PSG
BIOS=Custom(DRAM)
The others adjust I mantain the original
The instruments can vary and you can try any combination of instruments.
I also have runned Mopiranger directly from openMSX using a FS-A1GT config with FMBoot.COM with success.
Good luck to everyone!

I would love to understand how this works under the hood - I imagine implementing this into emulators - so all games could work with it (even games that writes directly to ports).

Knowing how this works, games that uses ports could also be adapted ^_^

PS:I do know this intercept BIOS calls - what I'm speaking of is how data is converted from PSG to FM...

Yeah, that part I understand - what I keep thinking is how to convert the values written in the PSG port to the ones in the FM port.

Maybe a basic (and inevitable) crude and global analysis of the ADSR (if any) used by the PSG software 'instruments' can be interpreted closely enough to decide, "this has a slow attack, so it must be a string section", or "quick attack and decay, then it fades down into a prolonged sustain, so probably they meant a piano there". If any variation is wanted, then you don't need a crystal ball to predict that the algorhythm to program such a thing will be extremely tricky. HW PSG presets would be easier, but they are generally not used in 'Japanese' game music.
Also, what people seem to forget here, is that the SCC volume is strictly linear, while PSG volumes are 'logarithmic'. Hence, there is no 100% accurate 'conversion table' to be made; it'd be all on approach, and 'close enough'.
With no disrespect meant, I am surprised by how fast people are satisfied by results of just having SCC or FM mimicking the PSG channels. Really good results are only reached by re-arranging the whole thing. There is no 'magical' algorhythm or code that will automagically make stuff sound good by just traversing the tone frequency data to another chip, but people already seem to be happy with just hearing the tones on their favorite chip, so who am I to criticize?

Algorithm to convert from PSG to OPLL is this.(Not accurate)
In fact, all of this I am writing in Z80-assembler.

// Relationship between PSG and frequency
//  frq = 111860.78125 / PSGparam
//
// Relationship between OPLL and frequency
//  OPLLparam = ( frq * (2^18) / (50*1000) ) / (2^(b-1))  ; b=octave
//
// Relationship between OPLL and PSG
//  OPLLparam = ( 586472.6528 / PSGparam ) >> (b-1)

for (ch=0; ch<=2; ch++ )
{
	// Calc OPLL data
	OPLLparamNow = 586472.6528 / ( GetPSGport(ch*2) + GetPSGport(ch*2+1)*256 );
	PSGvolNow = GetPSGport(8+ch);
	
	// Calc octave
	fNum = OPLLparamNow;
	for ( b=1; fNum>=512; b++ )
	{
		fNum = fNum >> 1;
	}
	
	// KeyOn by note change
	if ( ABS(OPLLparamNow - OPLLparamPrev[ch]) > NOTECHG )
	{
		// NOTECHG is a constant for determining whether to play another note.
		OPLLKeyOn(ch,fNum,b);
	}
	
	// KeyOn by volume up
	if ( PSGvolNow == 0 ) OPLLKeyOff(ch);
	if ( PSGvolNow > PSGvolPrev[ch] ) OPLLKeyOn(ch,fNum,b);

	// Save data
	OPLLparamPrev[ch] = OPLLparamNow;
	PSGvolPrev[ch] = PSGvolNow;
}

@Tiny Yarou

Very interesting, thanks for sharing!
I expected it to be a bit more complex, but maybe you left out some details here.
But then again, the best solutions are usually deceptively simple and to the point

@Tiny Yarou
Thanks for explaining! You made my day