badmojo wrote on 2023-10-28, 21:35:

Tiido wrote on 2023-10-27, 15:32:

EDIT: After some struggling, I have managed to make mixer locking and dual OPL2 simulation both fit into the design !

The mixer locking thing will be great, currently I have bat files to re-initiate my YMF719 based card after any game that messes around with the levels. This works fine but it's a bit messy.

Would this mixer locking with for other YMF-based sound cards or just yours?

It is hardware locking so just for Tiido card

MIDIto can lock a SB pro mixer too apparently - I haven't tried it though I must admit: MIDIto r12 (Midi and FM redirection, Mixer locking, Sound Blaster fixes)

Yeah, the locking is done via the CPLD on my card and it will not work for any other sound card, as with dual OPL2 emulation.

MIDIto is able to work with any card, as long as you can run the software itself (it needs 386 capabilities).

----

The card has had some more testing done, IDE works properly, WB stuff works properly. I'm continuing work on software so that rest of the features can be tested out too. Hardware should now be finished, all the CPLD stuff simulates properly and I expect no issues with it when software side finally catches up.

Some bad news, the dual OPL2 emulation is not going to work.

OPL3 has a bit that when set, enables all the extra features over OPL2, like extra channels and panning capability, and when cleared, the features become locked away.
The idea was that I enable OPL3 features, set panning on all channels to left or right, then disable OPL3 features, and the setting sticks...
And it does, but only as long as that register is not written to again, in which case both channels are forced on again, and unfortunately that same register holds parameters any OPL2 thing will touch too so it is guaranteed the panning gets reset.

While I can have OPL3 inside YMF704/721 panned to one side, through a dedicated volume register on OPL4 section that controls FM and PCM balancing which games will be unaware of, there's no such thing on OPL3 on the YMF71x side, and because external OPL3 is routed to same mixer level as internal OPL3 I cannot just pan the mixer and get the desired effect either.

Games would be able to detect presence of two OPL2 like chips however, but proper stereo will not be happening since one chip plays both channels always. Only way will be to mix the external OPL3 into CD input, then the FM+WB input can be panned from mixer and problem is solved. This however needs actual new board revision, and I am not going to be doing that and since it requires dedicated software+hardware support to set it all up and I won't spend time on a feature that cannot be properly implemented.

SB mixer locking is working good, games cannot change the levels that are locked. This is the most useful feature I think.

Software end has most of the hardware access related stuff done, it can properly initialize the whole card and has useful error messages etc. througout. Currently the init program is just about 8KB in size and I expect it to get to around 20KB once UI is added. It is several times faster and smaller than the previous program, primarly thanks to being an assembly program instead of a QB45 thing ~

can't say it's a huge loss - interesting to have for sure but the use cases are too few
mixer locking is the most useful of the two so it's nice this works

Is this somehow related to all other implementations of OPL2 being mono in other card manufacturers SB Pro 2 compatibility mode?

Mono is just how Adlib and OPL3 normally work over the normal FM access ports and vast majority of games work this way, totally ignoring any OPL3 or dual OPL2 specific possibility.

Dual OPL2 is a special configuration where ports that were used by CMS (SAA1099) on SB1.0 and 2.0 are used instead by two OPL2 chips on SBpro1 (and on SBpro2 these same ports are used by single OPL3 chip), and at the same time normal adlib (388+9) and SB FM (2x8+9) ports go to *both* chips for compatibility with existing software which expect a single chip and sound output in both L and R channels.

Since I have two OPL3 chips I could make their access appear as if there was dual OPL2 configuration, but because of hardware limitations I cannot actually get proper stereo output from this configuration.

Single OPL3 almost appears as two OPL2 chips, with main difference being lack of timer on second half that games use to detect dual OPL2 configuration. OPL3 also cannot physically support the Adlib compatibility mechanism where dual OPL2 card has both chips written to and because of the reason why I cannot have dual OPL2 like conf on my card. It is impossible to make single OPL3 look like dual OPL2 with actual stereo output either even if game is patched to ignore lack second timer and never accessing adlib or SBFM ports which are supposed to update both halves, you need two separate OPL2 or 3 chips for that stuff to actually work.

Dual OPL2 wasn't something I was interested in personally and on the plus side, less features means less testing and potential for bugs 👍

Tiido wrote on 2023-11-05, 18:54:

Mono is just how Adlib and OPL3 normally work over the normal FM access ports and vast majority of games work this way, totally […]
Show full quote

Mono is just how Adlib and OPL3 normally work over the normal FM access ports and vast majority of games work this way, totally ignoring any OPL3 or dual OPL2 specific possibility.

Dual OPL2 is a special configuration where ports that were used by CMS (SAA1099) on SB1.0 and 2.0 are used instead by two OPL2 chips on SBpro1 (and on SBpro2 these same ports are used by single OPL3 chip), and at the same time normal adlib (388+9) and SB FM (2x8+9) ports go to *both* chips for compatibility with existing software which expect a single chip and sound output in both L and R channels.

Since I have two OPL3 chips I could make their access appear as if there was dual OPL2 configuration, but because of hardware limitations I cannot actually get proper stereo output from this configuration.

Single OPL3 almost appears as two OPL2 chips, with main difference being lack of timer on second half that games use to detect dual OPL2 configuration. OPL3 also cannot physically support the Adlib compatibility mechanism where dual OPL2 card has both chips written to and because of the reason why I cannot have dual OPL2 like conf on my card. It is impossible to make single OPL3 look like dual OPL2 with actual stereo output either even if game is patched to ignore lack second timer and never accessing adlib or SBFM ports which are supposed to update both halves, you need two separate OPL2 or 3 chips for that stuff to actually work.

Glad you are implementing this, i remember i asked about if this was possible years ago on one of our PM exchanges. Question:

Besides 220 will you also map second OPL3 at 38A/B? I think then combining your card with a Pro Audio Spectrum 16 (muting the OPL3 on a PAS16) it wouldn't be difficult to simulate the first dual OPL2 Pro Audio Spectrum. PCM wise the PAS 16 is completely compatible with the original PAS.

OPL3 specific ports are always at those addresess too, that is how it works on YMF71x

Too bad about the dual OPL problems. That was the feature I was most interested in. But I'm going to want the card anyway. Hardware mixer locking alone is a great feature.

Speaking of dual OPL. I do have a feature request. How hard would it be to disable Adlib ports for perfect compatibility with PAS or Goldlib? Sure, since YMF is PnP, the Adlib ports could be simply remapped elsewhere. But I'd still like some hardware way to completely disable Adlib ports to minimize issues.

YMF71x Adlib port cannot be disabled but it can be moved out the way, basically entire range between 210...FF8 is usable assuming nothing else gets in the way first (the setup program excludes all standard hardware locations for user convenience). YMF704/721 OPL3 can actually be disabled or reside in 380, 388, 390 and 398 addresses (controlled by the CPLD on the card that does the other funstuff like mixer locking).

Things have been pretty slow due to the Norwegian course I am attenting on most days of the week but I am slowly getting somewhere with the software.

Items can be now selected, although adjustments are not yet done. I'll be getting on them pretty soon though ~

Some complexities are involved in supporting different featuresets though. As before, the program will work with regular YMF71x cards and also all the different YMF71x revisions which have differences in how they work and there are also the extra features of my own cards that have to be handled as well. But things are definitely coming along ~

The new program is a lot smaller too, currently just over 15KB and when done I expect it to be just around 20KB, which is a lot less than the 120KB of the old version that is written in QB45.

Looks fantastic Tiido. Will the software run on an 8088?

This just goes to show that a lot of the chips that existed were much more capable then previously thought. It's often been the drivers that hold them back.

carlostex wrote on 2023-11-23, 08:30:

Looks fantastic Tiido. Will the software run on an 8088?

It will, I am not using any instructions 8088/8086 do not have, same as with original QB45 program, but it does require VGA card since it uses 80x50 for the UI part and I have no plans to support 80x25.

EDIT: Levels are now being shown, sliders come next and then I have to add actual adjustment capability.

Tiido wrote on 2023-11-23, 18:07:

carlostex wrote on 2023-11-23, 08:30:

Looks fantastic Tiido. Will the software run on an 8088?

It will, I am not using any instructions 8088/8086 do not have, same as with original QB45 program, but it does require VGA card since it uses 80x50 for the UI part and I have no plans to support 80x25.

EDIT: Levels are now being shown, sliders come next and then I have to add actual adjustment capability.

What's the status of the project?

The website is down; and there are no download links to be found..
Am I missing something? 😕

Website has been down since I moved to norway, getting it back up has been a very low priority... You can find the most important downloads peppered in this thread.

Statuswise, the project is not dead but I have not had time to work on the software side of things but the hardware is finished. I have very much focused on Norwegian language, I now don't have to pay for my courses either so I will be taking the maximum out of them. A2 level has started, and I strive for at least B2 since I will be staying here for rest of my life ~

Tiido wrote on 2024-03-28, 06:38:

Website has been down since I moved to norway, getting it back up has been a very low priority... You can find the most important downloads peppered in this thread.

Statuswise, the project is not dead but I have not had time to work on the software side of things but the hardware is finished. I have very much focused on Norwegian language, I now don't have to pay for my courses either so I will be taking the maximum out of them. A2 level has started, and I strive for at least B2 since I will be staying here for rest of my life ~

Glad to hear you are well. Thank you for all the work you have done so far, and will wait patiently for it to continue.

V