thanks c.imp, when i have the scb-55, i try the test.
bye.

c.imp wrote:

Maybe you try just play this mid in any player?
Try this on your buggy CT cards.

Played the MIDI file on my worst card/daughterboard combination (CT2960 Vibra16C + Creative Waveblaster I 😳): No hanging notes to be heard at all.
In Doom these cards playing together make you turn off the music after a few seconds...

For MIDI playback I used the "mplay" program available here.

I don't think MIDI data alone will reveal the bug, the playback routines have more influence on this. After all, Hexen and Doom are based on the same engine, there seems to be something special about the way they use the MPU-401.

Btw, does anyone know what's the difference between selecting "General MIDI" "Wave Blaster" or "Sound Canvas" in the Doom setup dialog? They all appear to do exactly the same thing (and produce the same hanging notes).

🤣, that midi file is annoying!
I remember it sounds better in the game... 😁

5u3, hm... maybe some routines changed in midi file during conversation? Anyway try MEGAMID.

I think this different general midi variants in reference doom engine games setup is made just for people who don't know that is the same but know what they have. Just simpleing setup for those peoples.
Ofcourse results is the same.

keropi, For me this midi sounds in any mpu401 based player is the same absolutelly as in game via mpu401 device.

I played the midi file with several midi player and I didn't get any bug

I tried to play Hexen with Sfx disable (just midi) -> no bug

The bug seems to rely on simultaneous usage of PCM + MPU401 on the same sb16

This is the chief reason why there are, in my opinion, two better alternatives than any AWE32:

Sound Blaster 16 Multi-CD ASP (CT-1750) & AWE32 Upgrade Combo (CT-1920)
Pros: Improved compatibility w/v4.05 DSP, true OPL3 chip, possesses all chips & interfaces of original AWE32, smaller cards, no PnP or drives for basic operation.
Cons: Takes two ISA slots, requires custom cables to connect signals, some SB16s still have bug, worse SnR than later AWE32s.

Sound Blaster AWE64
Pros: No bugs, better SnR ratio, shorter size, more polyphony in Windows, better SPDIF output, equal or more RAM without upgrades.
Cons: Proprietary memory upgrades, no Waveblaster header, no PC speaker header

Great Hierophant wrote:

Sound Blaster AWE64
Pros: No bugs, better SnR ratio, shorter size, more polyphony in Windows, better SPDIF output, equal or more RAM without upgrades.
Cons: Proprietary memory upgrades, no Waveblaster header, no PC speaker header

Two small corrections: AWE64 cards do have a PC speaker header - at least all of mine (CT4390, CT4500, CT4520) have one.
And there is an issue with the emulated OPL3 - on AWE64 (and on some AWE32) cards FM sound has high frequency interference.

5u3 wrote:

And there is an issue with the emulated OPL3 - on AWE64 (and on some AWE32) cards FM sound has high frequency interference.

Is it interference? I know exactly what you are referring to, but thought that it was more of a high-frequency aliasing issue. If the AWE64 has adjustable filters, it might be an easy fix.

Can anyone judge how good the emulated OPL3 is on the AWE cards (none of which have a Yamaha OPL3 chip, only some Sound Blaster 32s do.)

Also, I may not be as enthusiastic about that Goldfinch AWE32 Upgrade board. The common ones require you to mix the board's output via the Sound Blaster 16's line inputs. This is noisier than the true AWE32. The Goldfinch uses a different midi emulation driver than the AWE32. See here for details as to why this might not be idea for games that natively utilize the AWE32:

http://america.hongfaith.com/_private/FAQ/AWE32FAQ.TXT

Cloudschatze wrote:

Is it interference? I know exactly what you are referring to, but thought that it was more of a high-frequency aliasing issue. If the AWE64 has adjustable filters, it might be an easy fix.

Yeah, I was picking the wrong word for it, it's definitly not interference from other chips or parts of the system, it's more like artifacts generated by the emulation itself. First I thought it was a side effect of the EMU8K chorus & reverb effects, but it also occurs with the effects turned off. The Vibra16 chipsets don't have this issue though.
"Aliasing Issue" describes it much better 😊

Overall the OPL sound generated by the "new" integrated Vibra chips is generally correct, but sounds a little tinnier and "softer", and with AWE64 cards there is a high-frequency "ringing" sound in the background -- one who has heard an AdLib long ago might not really notice it at first, but it's really obvious in direct comparison.

Found out something funny about the AWE64 OPL emulation issue:
In Win9x DOS boxes the aliasing noise is gone. Also, if you quit Win9x and enter DOS mode, the emulation sounds great, even if you reset the synth with AWEUTIL.

Cloudschatze wrote:

If the AWE64 has adjustable filters, it might be an easy fix.

Do you mean the OPL emulation output can be filtered while passing the EMU8K synth?
Looks like Creative had another reason to "reserve" two channels on the EMU8K for the OPL, besides saving a third DAC 🙄

EDIT: Executing the DIAGNOSE.EXE tool and going through the detection routines also gets rid of the aliasing noise.

i have try SB16 ct2960 dsp 4.17 witch roland scb-55.
with hexen demo (-warp 02) i have the hanging note
with the midi files, it is ok.

Update on the "OPL3 emulation high frequency aliasing issue": It seems to be caused by my mixer program. I thought I tested against this possibility, but obviously I made an error.

By the way: I have a CT3980 AWE32 PnP and I am experiencing the hanging notes bug with a DB-50XG. So I guess, that revision isn't bugfree after all.

I just tried the following test:

CT2760 + DB50XG

Hexen -warp 02

As far as my ears could tell, I did not have any hanging note(s). Inside Hexen setup, I set the music card to General Midi with address of 300 (not 330) because I have set my CT2760 to use 300 address.

Another thing to note: I have no idea at all if this is significant, but I removed a jumper on the AWE32, I forget exactly what it is called (and I can't check it out because the DB50XG is right over the top of this section of the AWE32!), but it's something very similar to MFBEN, something like that - I removed this jumper. (I'm just testing things out, I like to try stuff to see what happens.)

I also have a CT3990 which I may test with Hexen later today, and I also have a CT3900 which unfortunately prevents both my testing PCs from POSTing which appears to indicate that this card (the CT3900) may be damaged in some way.

Update; I just tried another test:

CT3990 + DB50XG

Hexen -warp 02

(Set music card to be General Midi, and this time set address = 330 because this is what this particular card accepts.)

Result = horrible hanging note.

I can update my earlier test now and say that I definitely did not get this horrible sounding hanging note with the CT2760 + DB50XG.

I wonder if I can set the CT3990 to use address 300 (and not 330), just to see if that helps at all? I'll take a look and see if I can do this...

I'd just like to add something to my post directly above regarding:

CT3990 + DB50XG
Hexen -warp 02

If I run this inside Windows 98, I do not get the hanging note. (If I reboot my PC to DOS, I do get the hanging note.)

Have you tried this with and with no in-game sfx sound playing in the same time?

I just did the following additional tests using:

CT3990 + DB50XG
Hexen -warp 02

Test #1 = Windows 98, both midi and sfx = works OK (no hanging note)
Test #2 = Windows 98, midi and no sfx = works OK (no hanging note)
Test #3 = DOS, both midi and sfx = does not work (you hear hanging note)
Test #4 = DOS, midi and no sfx = works OK (no hanging note)

If you need me to retry these tests, but using a CT2760 + DB50XG, I will do so tomorrow morning. Thank you.

Hmm, no hanging bug in Win98 with 3990 rev, yeah? Very strange.
Maybe you just forgot to disable MPU-401 emulation in Windows environment (where MFBEN jumper is totally ignored) and has not paid attention for difference in synthesis? Look in AWE Control Panel (if you did not have it), select Device... tab and uncheck Allow MPU401 Emulation on this device. And then test it again in Win98 DOS box with SFX sound.