VOGONS


First post, by tadeu

User metadata
Rank Newbie
Rank
Newbie

Good night,

I got two sound cards. The older one, installed in my 486 DX-2 is a Jazz Vision 16, which does FM synthesis through the ubiquitous YMF-262 chip. And, well, it's OPL-3, and works out of the box with everything (except Windows NT sometimes, but it's another story). The newer is an Audigy 2 ZS, originally installed in an old Dell P4HT.

MIDI music in the Jazz card is a point-and-shooter... You just tell wherever the board is, and AdLib+FM synthesis chip does the rest. The Audigy, however seems not to posses any FM synthesis equipment. It only works when I choose a SoundFont.

So, alright, Audigy does not have FM, end of history, right? Well, not quite. AWE32/64 (EMU8000/EMU101K) seems to provide a certain "compatibility layer" in which it does provide some kind of integrated synthesis. What I'm curious about is if my Audigy card contains such mechanism. It seems not to, but documentation is a bit scarce. How did this emulation kind worked before?

Thanks,
Att.,
-trp

Main rig: Asus P8H61-M LX2 R2.0, Xeon E3-1245 v2, 12 GB DDR III, GTX 760 - running Debian Testing
Retro rig: AcerAcros, 486DXII/66, 16 MB SIMM, Cirrus Logic GD5428, Jazz16 SB, U.S. Robotics 36.6K Modem, FreeDOS

Reply 1 of 7, by leileilol

User metadata
Rank l33t++
Rank
l33t++

AWE32/64's FM wasn't "part" of the Emu8000 chip involved. AWE32 had YMF262s integrated, some later AWE32s and SB32s had CQM instead, AWE64 also has CQM. Being ISA also helps. Creative's PCI sound cards are another story, starting off with the acquision of Ensoniq.

SBLive had Emu10k and its FM was done by a DOS TSR that only provided software-driven FM for DOS games (at least in the later drivers anyway. earlier drivers and the AudioPCI used ECW wavetable approximation synthesis and you can't use this 'fm synth' with Windows games/apps)

apsosig.png
long live PCem

Reply 2 of 7, by realnc

User metadata
Rank Oldbie
Rank
Oldbie

Yep. There was no emulation involved with SB cards up to and including the AWE32 (except with some later models of the AWE32.) They actually had a real OPL chip on board.

The OPL software emulation in later cards was horrible though. It sounds really bad at best, and like garbled garbage at worst.

Reply 3 of 7, by Jo22

User metadata
Rank l33t++
Rank
l33t++

EMU8000 was a creation by EMU systems originally, if memory serves.
Two voices channels of the EMU8K on AWE boards were used to inteface with the OPL3/CQM "chip".
This saved a dedicated DAC for FM and provided digital output of FM synthesis, I believe.
It also allowd for adding echo/surround to FM music (see Phil's AWE64 Gold videos).

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 4 of 7, by tadeu

User metadata
Rank Newbie
Rank
Newbie

Thanks for the replies, all of you. I think it pretty much answers my question.

So, this is it. On what comes to synthesis, Audigy just does on hardware what Fluidsynth does on CPU. A fancy hardware Fluidsynth, I'd call it. hahaha

I always got the impression that FM synthesis had kind of a personality which no SoundFount can 100% mimic. Same goes for those external modules like MT-32, SC-880 and so on (although Munt does a really good job emulating MT-32, at least for me, an average John, on what comes to music).

But guess it's just a question of loading up any good SF and things should sound pretty good. I was looking for something like this:
https://www.youtube.com/watch?v=-y3PttJRyAM&t=59s

AWE64 had a very crisp and realistic feel without losing that “punch” OPLx had.

Last edited by tadeu on 2018-11-12, 23:19. Edited 1 time in total.
Main rig: Asus P8H61-M LX2 R2.0, Xeon E3-1245 v2, 12 GB DDR III, GTX 760 - running Debian Testing
Retro rig: AcerAcros, 486DXII/66, 16 MB SIMM, Cirrus Logic GD5428, Jazz16 SB, U.S. Robotics 36.6K Modem, FreeDOS

Reply 5 of 7, by realnc

User metadata
Rank Oldbie
Rank
Oldbie
tadeu wrote:

So, this is it. On what comes to synthesis, Audigy just does on hardware what Fluidsynth does on CPU. A fancy hardware Fluidsynth, I'd call it. hahaha

Well, doing it on the CPU back then wasn't an option. It would have to be lower quality and you'd take a big performance hit while gaming. The AWE32 came out when many people were still gaming on 486 CPUs. By the time the Audigy came out CPUs were much faster, but still not fast enough as to allow high quality wavetable synthesis though software without hogging the CPU.

Reply 6 of 7, by tadeu

User metadata
Rank Newbie
Rank
Newbie

Well, to be fair, I feel that things like Fluidsynth still somewhat CPU hungry, specially when running on underpowered chips like my Pentium G2020. So an Audigy is welcomingly useful. I was just curious on wheter it had ou hadn't a “personality” like AWE had. I ended up loading a good soundfont in and pointing my DosBOX to the EMU10K. Tyrian soundtrack sounded gorgeous without eating my CPU for launch. Even being slightly old, it's a very good card for cheap builds, like mine! hahaha

Main rig: Asus P8H61-M LX2 R2.0, Xeon E3-1245 v2, 12 GB DDR III, GTX 760 - running Debian Testing
Retro rig: AcerAcros, 486DXII/66, 16 MB SIMM, Cirrus Logic GD5428, Jazz16 SB, U.S. Robotics 36.6K Modem, FreeDOS

Reply 7 of 7, by leileilol

User metadata
Rank l33t++
Rank
l33t++
tadeu wrote:

AWE64 had a very crisp and realistic feel without losing that “punch” OPLx had.

I don't know about that. Wavetable takes a lot out of Tyrian's sound, and oddly of all the wavetable renditions, I prefer how the AudioPCI renders it. Some of the 4mb/8mb ECW synths work out really well. Shame that the sound driver doesn't! It's unplayably crackly ingame.

apsosig.png
long live PCem