XQ333A0 is the sample ROM (2Mbytes) and the NEC chip is sample RAM (128KBytes) used for custom samples that you can toy around with bundled software. YMF290 is a chip that captures MIDI data and generates an interrupt that the TSR will then use to read back that data from the chip and emulate things through YMF278, it is a clever solution and will work with pmode and through Windows also. YSS205 is a poorman's effects processor + DAC for the YMF278 and the LHxxxx is PSRAM that holds the effect data. AD1848 is the DAC+ADC+mixer, WSS type stuff. OPTi translates SB data to WSS data in SB mode and provides CD interface and means to access the effects chip. AD1848 output is routed through the effects chip and get the currently loaded effect applied on top.

The software that sets up the effects chip doesn't do any sort of delays between the writes and it causes the process to fail on hardware much faster than P1 @ 133MHz unless you add a lot of wait states on ISA accesses or disable caches. One day I will hack the software to allow things to work with faster hardware. You will get strage noises or no sound at all on faster hardware as the effect coeffs are all mangled up as they get written to the chip.

Wow! Thanks for that! You certainly know your sound card stuff! 😀
So, this card would ideally suit say a 386/486 build.

Edit:
Any fans of this card out there? Keropi, kanecvr, others..?
End Edit.

I have the card and did a nearly full reverse engineering on it not so long ago. I am very fond of the MIDI side of it but only in Windows. The DOS TSR only uses a small subset of the samples available, probably due to memory requirements.
386 will probably not work so well for the MIDI emulation in DOS games, the manual suggests a fast 486 being minimum. The OPTi chip has terrible drivers and it seems it cannot do stereo for SBpro either, all games I tried came out mono. WSS side seems to be all ok though.

Oh yes, that's was my 2nd soundcard ever - bought it new back then for my 386sx20. Still have it complete/boxed. Back then it was a great board.
It is true the OPTI MAD16 driver/init program was crap. I also could not make the card work in anything faster than a 486, IIRC in pentium-land I was getting weird noised during boot.
Can't wait to see what Tiido releases now 😀

Re: the good sounding MIDI in Windows. Is this true if you run a DOS game inside a Windows "DOS Shell"?

DOS stuff only works via the TSR, if you don't load it you will not get any music in DOS games, Windows driver does no sort of virtualization for the DOS stuff. Windows programs/games work via the Windows driver and gets to use the fancy stuff that can also use FM in addition to samples.

Also gameport output is SB-MIDI only so you'll not be getting any sound from most games when using external MIDI modules via the gameport.

I have one of these for quite some time. Long time ago I used it together with a GUS ACE in my main build, Pentium 200 MMX, 64/96 MB SDRAM. It worked quite well and I do not remember any sound problems in games other than the normal issues of speed sensitive games, which manifested with all soundcards I tested.