5u3 wrote:

elianda wrote:

My main use of gsplay was switching between MT32 and GM/GS mode. Now with DB50XG it gets really weird, gsplay MT32 says that the […]
Show full quote

My main use of gsplay was switching between MT32 and GM/GS mode.
Now with DB50XG it gets really weird,
gsplay MT32 says that the device does not support MT32 mode.
now if I try to send the *.mid that switches DB50XG to the MT32 bank gsplay detects some GS messages therein, skipping them since it detected the DB50XG as GM only.

Hmmm... I never used GSPLAY for anything other than switching between modes, but it works well enough for me (just tested it again with both SCB-55 and DB50XG).

Maybe it's caused by some setting in the config file?

I am not sure, you can give an option for the richt Patch Names and for the Reset Mode. Maybe the last defines the devices capabilities.
Whats your setting for the DB50XG ?

These are the settings I use for both DBs:

1MIDI_INTERFACE  MPU401
2MIDI_DEVICE  M
3MINIMAL_DEVICE  N
4RESET_MODE  GS
5PARTIAL_ALLOC_RESET  Y
6GS_SYSEX  Y
7NON_GS_SYSEX  Y
8ALLOW_RESET  Y

Another thought: Are you using this on a fast machine? The author mentions a 486/66 somewhere in the config file, maybe adjusting this or slowing down the CPU might help.

The DSP V4.04 does have the hanging note bug, doesn't it?

I always thought the initial Sb-16 release was DSP v4.05, which is said to not have the hanging note bug, but a bad SNR instead.

gerwin wrote:

I always thought the initial Sb-16 release was DSP v4.05, which is said to not have the hanging note bug, but a bad SNR instead.

I agree with that.

I've never seen a DSP revision older than 4.05. Every CT1740 I saw was equipped with it.
I always thought the CT1740 was the original model.
But maybe v4.04 exists, like that mysterious CT1730.

You can greatly improve sound quality.
Simply bypass the output amplifier (close pins 1+2 on OPSL,OPSR)
and use a pair of active speakers or an external amplifier with headphones.

Edit: There's something that makes me wonder..
People often complain about the bad SNR of the early SB16, but adore the SB Pro 2..
So do the original SB 1.x, SB 2.0 and SB Pro's have a higher SNR than the CT17xx ?

My new CT1740 just arrived. It came with the infamous CT1900 Waveblaster. The CT1740 does have the V404 DSP chip, as I was told. So this is really early stuff then?
Anyway, since I never had one of these and I got no manual with it, I shall give it a spin. Does it make sense to test for hanging notes with the Waveblaster?

ux-3 wrote:

Anyway, since I never had one of these and I got no manual with it, I shall give it a spin. Does it make sense to test for hanging notes with the Waveblaster?

Yes. The "hanging note" bug does not depend on the DB.

😊

How do you install this Waveblaster in Windows? Just run the hardware manager?

Edit: Just switch Midi from 220 to 330, I guess?

Next Edit:
With Hexen, I choose General Midi, not Waveblaster, right? Or doesn't it matter? And what do I look (hear) for then?

Well with Soundblaster you get MIDI from OPL2 and MPU401 interface.
The OPL3 plays through a driver conversion.
The MPU401 interface goes to the hardware which you plugged either on the external MIDI connector or the WaveBlaster Connector.
The driver or card does not know, what MPU401 card/device you have plugged in, it simply uses it for playback.

Usually you switch in Windows in the multimedia control the midi output from OPL to MPU401 (or FM-Synthesis to external).

As for the games, older games show different options for Waveblaster and General Midi but in most cases it is simply the identical driver that is chosen.
Someone knows a game that handles WaveBlaster different than any other connected Wavetable?

As for the CT1900 itself, for its age it doesnt sound too bad at all. It doesnt handle effects, but the samples are ok. So if you got a MIDI song that uses not much effects it still sounds good. There are much worse solutions. (like Orchid Soundwave 32).
To get an impression check this out:
http://www.voodooalert.de/board/index.php?pag … &threadID=17133

(Scroll down a bit for the WaveBlaster recordings).

Jo22 wrote:

I've never seen a DSP revision older than 4.05. Every CT1740 I saw was equipped with it.

FS: Sound Blaster 16 cards (first model)

OK, I cycled the board through all SB cards I have and ran Hexen -warp 02
on the CT1740 (ASP onboard) DSP 4.04 I found nothing unusual.
on the CT1770 (empty ASP socket), DSP 4.13 I got a high pitched humming noise right at the start of play that wouldn't go away,
on the CT3980 (ASP onboard) DSP 4.13 I found nothing unusual, but am not sure if the card truely uses the Waveblaster. No matter what device (GM Midi, Waveblaster, AWE32) I choose, I always need to use address 330. (I just cross checked by removing the Waveblaster - no more music in GM and WB). So the 3980 did not readily display the stuck note problem.

I wonder if the hanging notes bug inversely correlates with the presence of an ASP processor?

I wasn't aware of that topic.
Thanks, robertmo. 😀

ux-3 wrote:

I wonder if the hanging notes bug inversely correlates with the presence of an ASP processor?

Close pins on IFSD, APSD to disable the ASP (aka CSP).

Jo22 wrote:

Close pins on IFSD, APSD to disable the ASP (aka CSP).

The 1740 doesn't have pins. You have to live with the ASP or solder!

ux-3 wrote:

The 1740 doesn't have pins. You have to live with the ASP or solder!

I awaitet that response. 😉
Some of them have those pins, some haven't.

OK, my next card is here. Again AWE32, but this time it is no plug&pray device. Again, it has the DSP 4.13, and again, I could not find any hanging note problems with the two daughter boards. Perhaps I miss them?

I now have 3 SB cards with DSP 4.13. The AWE32 ones seem to play fine on both DBs, the SB16SCSI starts hexen warp 02 with a loud enduring tone that won't go away.

Nice about this card is the ability to change IRQ without reboot. Can be done by batch it seems. As far as I have tested, I can disable L1 cache in Dos without crashing.

So I am wondering if I should use Yamaha DB with my SB16 with DSP V4.04 or my AWE32 with DSP V4.13?

I have made a new observation that may be of interest to some of you:

I used the (fascinating) game Buzz Aldrin's race into space, cd-version. This older gem can only be configured for irq7 (if not, please tell me)

I configured the game to play sound effects through SB and MT32 through DB50XG.
This worked fine on my AWE32 with V4.13. It also works fine on my SB16 CT-1770 with V4.13. However, it failed on my SB CT-1740 with V4.04.

Failing in the sense that the PC freezes within a second upon hitting enter in DOS, when the red/blue color sceme appears. I swapped the two SB16 directly, identically jumpered. One works, one doesn't!

I experimented a little further: The DB50XG works, when sound effects are disabled, and sound effects work, when no music (or plain FM synthesis) is selected.

Apparently, the crash occurs only when both voice and mpu-401 are active. Sounds like another version of the hanging notes bug? Anyway, this is DSP V4.04.

Any comments?

ux-3 wrote:

Apparently, the crash occurs only when both voice and mpu-401 are active. Sounds like another version of the hanging notes bug?

No, it's a different issue, related to the SB16's use of a shared IRQ. The "SBMPU401" utility should stop the crashes.

Many thanks! You are right - this util does indeed do the trick.

Would it make sense to keep this permanently in autoexec.bat?

Curiously, the SB16 has both irq5 and irq7 reserved in bios. Don't see where the conflict would sit.

I am still puzzled by that hanging notes issue. I have tried two AWE32s in Hexen warp 02 and either I am just too stupid to notice or it doesn't happen. Others report that it does happen quickly. If it doesn't happen here, then perhaps there is another factor that has do with it?

ux-3 wrote:

Curiously, the SB16 has both irq5 and irq7 reserved in bios. Don't see where the conflict would sit.

It's not a system conflict, rather, it's a problem with the game's MIDI interrupt handler. Per NewRisingSun: "On the SB16, the single Interrupt Request Line is shared both among DSP and MPU interrupt requests. If it doesn't work without SBMPU401, it's because the game is not prepared for IRQ line sharing (which wasn't necessary before the SB16 came out). All SBMPU401 does is disable the MPU IRQ completely."

I am still puzzled by that hanging notes issue. I have tried two AWE32s in Hexen warp 02 and either I am just too stupid to notice or it doesn't happen. Others report that it does happen quickly. If it doesn't happen here, then perhaps there is another factor that has do with it?

As far as the "hanging notes issue" is concerned, there seem to be two types of hanging notes - legitimate, and illegitimate. Hexen and Doom (and others I'm probably not aware of) fall into the latter category.

Take a look at a "Hexen -warp 02" hang:

See those A7 notes? They shouldn't exist. Basically, the card (in this case, a CT2940) appears to be re-reading the velocity byte from the last legitimate note (E2) multiple times, from which a bunch of spurious notes are generated. Since Note Off messages for these bastard notes don't exist, they simply hang. As of yet, I'm not sure why this problem occurs.

For an example of a legitimate hanging note, try an AWE32 with X-Wing CD. Though mostly inconsistent, the occasional hanging/dropped notes do occur. In this case, I believe the issue to be both related to how the game handles MIDI, and either an undersized FIFO buffer on the AWE32, or the fact that its DSP may not be fast enough to handle everything being thrown at it (MIDI + PCM).

Furthermore, both the DSP version and the bus interface chip appear to be factors with both types of hanging. You didn't experience the Hexen hang with your AWE32s because the cards with the CT1747 interface chip don't exhibit this problem, nor do the SB16's with the CT1746 interface chip and 4.04/4.05 DSPs.

Whew. Confused yet? 🤣

I wonder if the hanging notes bug inversely correlates with the presence of an ASP processor?

This is one of the things I keep meaning to look into. After looking at the circuit traces, I think I came to the conclusion that the CSP would need to be physically removed in order to prevent the DSP from (theoretically) wasting processing cycles on it. For the most part though, the hangs are so few and far between (with the AWE32), that I haven't thought it necessary. Besides, why deprive yourself of the chance to play the one known game that actually supports QSound via the ASP/CSP?

Thanks for your detailed reply!

Is that Hexen blog from the very start of the warp run? Because my "hanging" note appears very early (near instantly) when the ingame starts.

And it appears on only 1 of 4 cards: A V4.13 SB16 without the CSP/ASP processor. The working ones have the CSP.

I have read through quite a few threads, trying to figure out why some AWEs with V4.13 hit the sack while mine don't (or I fail to notice). From what you write, it is basically a speed/synchronizing problem. Perhaps the isa-bus speed has to do with it? My troubled SB16 V4.13 requires a bios tweak in order to fire up in windows. Otherwise, the win fanfare dies after the first sound (regardless of DB presence). Could be a coincidence of course.

Unfortunately, I don't own x-wing CD, so I can't test that.

A quick glance at the SB16s revealed: The working one has CT1746A, the hanging one has CT1746B.

A different question: Why does the hanging Hexen note disappear, when I lower the ingame volume to zero and back to normal?