Probably not much incentive to make one, since there are a lot of inexpensive sound cards that offer SB Pro compatibility.

Shponglefan wrote on 2025-12-21, 01:42:

Probably not much incentive to make one, since there are a lot of inexpensive sound cards that offer SB Pro compatibility.

I considered that, but the problem with that theory is that any card that can emulate the stereo SB Pro, can emulate the old mono Sound Blaster.

As long as new old stock ISA sound chips were still available from reputable suppliers, it was easier to just use an integrated SB Pro 2 compatible chip for sound card projects like the ESS 1869F , CMI8330 or YMF714. Now, those stocks are drying up and other options need to be considered.

The SB Pro 2 and SB16 DSP firmware has been dumped (it's just an 8051 derivative) and OPL2 and OPL3 chips are still widely available, but what still seems to be missing is the Mixer IC. What the mixer does on the SB Pro 2 isn't that complicated but since it's an analog/digital hybrid chip it's not easy to replace with programmable logic.

The RP2040 on the PicoMEM and PicoGUS doesn't seem to be quite powerful enough to run the Nuked OPL3 core though an updated version of the PicoGUS with the RP2350 could probably manage it. @polpo is busy with an optical drive emulator project at the moment though. Will anyone else step up to accomplish the project?

As someone pointed out, there's an abundance of SB Pro 2 compatibles. In my opinion, as soon as its done on FPGA, no one will ever look back.

carlostex wrote on 2025-12-21, 16:57:

As someone pointed out, there's an abundance of SB Pro 2 compatibles. In my opinion, as soon as its done on FPGA, no one will ever look back.

I was under the impression that the supplies of FPGAs usable for retro computing projects had dried up a couple years ago, which is why Raspberry Pi based stuff is so common now.

Moogle! wrote on 2025-12-21, 17:31:

I was under the impression that the supplies of FPGAs usable for retro computing projects had dried up a couple years ago, which is why Raspberry Pi based stuff is so common now.

FPGA's are expensive per unit, and with the advent and expansion of AI I don't see the price coming down. It is also harder work, probably, to implement gate level accurate logic, it requires serious reverse engineering on original hardware so that it can then replicated on FPGA via a hardware description language like VHDL or Verilog.

The advantages on the other way are obvious, if you make a mistake you fix and reprogram. There's no need to have several revisions of fabbed silicon which happened too frequently during circuit development. I would say the biggest advantage is that you can improve on the original design, but that might put off purism and go beyond if the original goal is 100% gate level recreation.

The chips are available but they do cost sometimes twice as much more than they did compared to pre-covid times. I think the problem comes down to the fact that there just aren't enough hardware people at the problem. Most things I see done nowdays are mainly by software-first people, who are able to eventually come up with a working hardware side although often with some shortcomings. You can buy FPGAs capable of doing any retro sound card for under 10usd price, you absolutely do not need tho many tens to hundreds and even thousands of usd priced ones.

FPGA gateware is also not necessarily harder than wiriting software on a MCU, it definitely is different but if you never ever did something similar it may take a while to wrap one's head around. PLD is able to hit exact timings at same speeds as original hardware (i.e 1...25MHz clocks you see on all the old things), while an MCU often needs to run at performance levels multiple times greater than the host hardware is at. Improvements are possible in both ways, but in either case one has to know quite well what the original hardware is doing and then spend time getting it going and even more time refining it to perfection.

IMO there is very little nostalgic demand for genuine sound blaster PRO/2 boards, and that is normally what motivates people to create projects around something. My feeling is that most people's experience with the soundblaster PRO/2 standard was via a compatible card as it became a psuedo-standard for DOS sound amongst many brands. The SB1/2 and SB16 on the other hand were far more prolific.

Hang on, are we talking SBPro (CT1330A) or SBPro2 (CT1600 etc)?

Fully agree the added value of a CT1600 clone is tiny as pretty much anything with a YMF262 will (unless their analog filters are very different) sound exactly the same. But the original SBPro with dual YM3812 is a different kettle of fish, at least for the tiny handful of games that were written for its version of stereo.

Tbh, I'd be more interested in a PAS replica as there's more stuff out there for it, but I've seen cards with smaller unique software libraries than SBPro 1.0 be reproduced to play those few titles with authentic audio.

dionb wrote on 2025-12-22, 23:40:

Hang on, are we talking SBPro (CT1330A) or SBPro2 (CT1600 etc)?

I was referring to the CT1600, with stereo PCM, though when I made the thread, I had been under the impression all the Pros had stereo PCM.

Sorry for the confusion.

the PCM on the 1330A and 1600 is indeed the same, it's the FM that's different

thye both have "stereo" but the 1330A accomplishes it with two OPL2 chips where the 1600 used an OPL3

I don't really follow. The only unique thing about SB Pros is the CT1330 and its two OPL's. The CT1600 has been duplicated, endlessly. SB Pro PCM and OPL3 FM is the combo every clone went for, with somewhat variable quality.

*cough* SoundBlaster Pro MCV

(although a SoundPiper clone using the ES1869 would be fine too)