LunarG wrote:

Well, not the chip on it's own, but the whole soundcard, along with its drivers, function as one. It is the same way with many devices... The synth engine it self is designed as a native midi device, but together with the rest of the pieces it forms a midi device. I'm now talking about such things as musical instruments etc. They rely on many separate components to make them a full "midi device". So in the same way, you could argue that while the EMU8K (or EMU10K) chip alone isn't a midi device, the whole sound card makes up a midi device. I mean, outside of simple daughterboards, find me a midi synthesizer that doesn't run some sort of software to actually function.

Very true. Probably all MIDI synths are built up of a microcontroller to interpret the MIDI data, and then translate it to raw writes to the proprietary registers of whatever sound generator the synth uses.
I suppose PC is a special environment, where you do not necessarily have to rely on MIDI support in the hardware layer. Instead of using a dedicated microcontroller, you could simply move the MIDI interpreter to the host CPU, to save cost.

And getting back to the original discussion: most FM solutions for PC are low-budget, and as such do not have a MIDI implementation. So they rely on the host CPU to translate the MIDI data. The only exception I know is the IBM Music Feature Card, which is essentially a Yamaha FB-01 and a proprietary MIDI interface on an ISA card. The FB-01 part uses a Zilog Z80 to process the MIDI and drive the YM2164 FM chip.

The Roland LAPC-I is the same concept: it is basically an MPU-401 and MT-32 integrated on an ISA card, and both the MPU-401 and MT-32 have their own microcontrollers (the MPU-401 interface has a very advanced 'intelligent' mode, which is a complete 8-track sequencer, hence requiring a microcontroller of its own).

Same for the Roland SCC-1, which is like the LAPC-I, except the MT-32 is replaced with a Sound Canvas.

The GUS however, has no microcontroller onboard, and no MPU-401 compatibility in hardware. So it relies on the host CPU to emulate an MPU-401, and translate the MIDI data to register writes for the GF1 synth chip. This is very similar to how OPL2 and OPL3 chips are used.

I actually wonder if the CA20K2 (EMU 20k) on my X-Fi Titanium HD takes care of interpreting midi data (seeing as it had a built in RISC core, and capable of 10.000mips, it certainly should be capable enough) or if it has to rely solely on a driver level implementation. Not that it matters of course, as today, "midi" has become much more wibbly, wobbly, timey, wimey thanks to the use of DAWS. Sure, "midi" is still a part of it, but there's a reason they just launched Midi 2.0 😁

dr_st wrote:

LunarG wrote:

SirNickity wrote:

The AWE music synthesizer (EMU 10K) is also not a MIDI device. It takes software to translate MIDI data into EMU 10K commands. Same goes for the Gravis Ultrasound products.

You mean the EMU8000 of course. 😀

He sure does. But the EMU10K (Live/Audigy) is not a MIDI device either. 😁

😢 *sigh* Hang on, I'll go get my official Audio Nerd card...

Altought the MT32 uses MIDI I wouldn’t call it that because unlike GM, it’s proprietary to Roland. Games of the time tended to call it Roland.

lvader wrote:

Altought the MT32 uses MIDI I wouldn’t call it that because unlike GM, it’s proprietary to Roland. Games of the time tended to call it Roland.

You have to separate two things though:
MIDI and General MIDI.

MIDI itself is a standardized interface, it's not proprietary.
The MT-32 uses the standardized MIDI interface, so it surely is a MIDI device.
However, the MT-32 (like the Yamaha FB-01 for example) predates the GM standard, and as such is not compatible with it.

Yes, it's like an old web browser that can speak TCP/IP/HTTP just fine but doesn't know how to render the latest HTML/Javascript standards.