I am testing an AWE32 model CT3900. I run diagnose.exe, and it tells me that the DSP is 4.13. The AWE32 has a waveblaster DB attached to it, model CT1910. I have set up the soundcard+DB inside a slot 1 Abit AB-BX6 mobo, with Windows 98. However, I am testing the soundcard inside "pure" DOS 7.1. I run hexen -warp 02, and hear no "instant" stuck note. I leave the music running for 10 minutes, and constantly check for stuck notes, and there are none. I run doom (1.9 shareware), and perform the same test. After 5 minutes of no stuck notes, one note gets stuck, but this stuck note sound only lasts about 1 second. The music then carries on as normal.

There's something interesting about this CT3900 card. It's got a sticker on one of the small chips, just north east of the IDE jumpers, in the bottom left hand corner. It reads -

Creative
01482290 10
Rev A

The last part of the number section, 10, the 1 is not completely upright, and almost looks like a / character.

ux-3 wrote:

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.

Yes, precisely.

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

So, here are some of my results with Hexen...

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).

Personally, I think the problem with the "legitimate" hanging notes is a bit overblown. Depending on what games you play, you might never encounter a stuck/dropped note on an AWE32 at all.

From what you write, it is basically a speed/synchronizing problem. Perhaps the isa-bus speed has to do with it?

Possibly, but perhaps only as pertains to the "legitimate" hangs. I haven't had any success preventing the illegitimate, Hexen-type hangs through any combination of system BIOS or bus-speed settings. If you discover something different, please tell.

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

I'd have to check, but an All Notes Off MIDI message is probably being sent.

Thanks for your elaborate reply! Do you have a source for that table? Is there more info, where this came from?

ux-3 wrote:

Do you have a source for that table?

It's just something that I specifically created for the posted message, based on my own testing. Ideally, it might be nice to start a detailed spreadsheet comprising a list of every game for which hanging MIDI notes have been experienced, the soundcard used, DSP version, etc.

Is there more info, where this came from?

Like what? I suppose it depends on the question. 😀

retro games 100 wrote:

I run doom (1.9 shareware), and perform the same test. After 5 minutes of no stuck notes, one note gets stuck, but this stuck note sound only lasts about 1 second. The music then carries on as normal.

I guess I'd have to hear it, but a note that plays out for 1 second may not necessarily be stuck. 😀

In any event, this would likely be an example of a "legitimate" hanging note, where the Note Off message was "lost," but whereby a subsequent note of the same value correctly sounded and terminated, resolving the hang.

Cloudschatze wrote:

Like what? I suppose it depends on the question. 😀

I was assuming you posted an excerpt of a longer table.

Anyway, I am very glad to understand the cause and unfolding of the mysterious effect better now.

Are there any other manufacturers known with a stuck notes problem or is this an exclusive CL-feature?

ux-3 wrote:

Are there any other manufacturers known with a stuck notes problem or is this an exclusive CL-feature?

Well, Media Vision essentially scrapped an entire soundcard because of a hanging notes issue...

Skip McIlvaine - Media Vision/Aureal wrote:

Media Vision produced a sound card (initially intended to be the Pro Audio Studio XL)...that had one unique feature from all of the other Pro Audio Spectrum 16s: Wavetable DB connectors. Unfortunately, after a massive amount of testing, it was discovered that the Spectrum (MVD101) chipset could not properly support these headers without the side-effect of the occaisional hanging note. Media Vision continued to use these PCBs under the name Pro Audio Spectrum 16 (because they were in every sense of the specification a Pro Audio Spectrum 16) but did not mention the fact that they included a 26-Pin connector for Wavetable daughtercards.

Purportedly, Media Vision's later Jazz chipset has similar issues (though I've yet to experience them)...

John Neary - Media Vision wrote:

The Pro3D was the pinnacle of Media Vision...The only problem was that the Jazz chipset had a bug where the Midi buffer was too short and would drop data once in a while resulting in a missed note, or worse, a stuck note.

(emphasis mine)

Cloudschatze wrote:

http://www.symphoniae.com/misc/hexen_hang.jpg […]
Show full quote

Nice. It is what people need here. I saw 2 games were popular to test "hanging note" issue, - it would be good to see such table for "X-Wing CD" too. Just in case: to issue appear was needed "SB16/AWE + GM" mode in games, other modes such as "SB 2.0 + GM", and "GM only" may not work for this. So if you can, please give info for second test-game.

Add the CT2980 as having the hanging note bug with Hexen -warp 02

Add the CT2800 as having the hanging note bug with Hexen -warp 02

I read this thread with interest. The hanging note bug discussion is of course known to all and a frequent occurrence. What made this particular one interesting to me is ux-3's findings of having a SB16 with DSP 4.13 that does not have the bug.

I have the same thing.

A SB16 (CT2230) with DSP 4.13, CSP chip and CT1745a interface chip and it does not have the bug.

And I also have a CT2910 again with DSP 4.13, same interface chip model number but a different date and NO CSP. This one DOES have the hanging note bug. The card also has a big square flat chip while the other has a small rectangular OPL chip. It also has IDE (which can not be disabled), the other has proprietary cdrom ports.

Although these are the only cards I personally tested these two do exemplify how much more complicated this issue is by both having DSP 4.13 and yet giving opposite results. There is a lot more to it than just the DSP and DSP 4.13 is on nobody's safe list.

I think it would be a step forward in understanding and closing this issue if ALL the differences of the cards are put side by side in a nice table. Or even a collection of high resolution scans or pics so comparisons can be done by anybody who wants to.

Cloudschatze, you reading this mate? What do you think? You seem to be the only one using a scientific method to understand this issue.

Can I just point out something minor please. I scrolled up to the top of this page, and noticed that my post has some strange Skype phone number "image" embedded in to it. I've never seen that before on Vogons.

Edit: This Skype phone number can be seen between the words Creative, and Rev A. I think what's happened is that the number I originally wrote in this message above has been "transformed" in to Skype advertising. I guess this transformation might only be visible on some people's machines, and not on all of them?

You have skype installed right?

It's a plugin I believe and everywhere it spots a phone number on a website, it puts itself there so you can click on it.

Either disable the plugin, look in the Skype options or uninstall Skype

Yes of course, that's the reason! Unfortunately, I wasn't able to "put two and two together", and come to that obvious conclusion. Oops. 😦

It's Spykeware 😀

So its really sound card problem? I know hexen not working properly in some cases try test sound on other games?

Here's another model to add to the Hexen hanging note = "Yes" list: CT2980.

I still wonder why people go for SB16s and not AWE64. Sure it might not have a real OPL chip, but any games from that era use MIDI anyway.

My "Sound Card comparison - Digital Speech - DOS Game Wing Commander II" has shown me that the AWE64 GOLD is one of the best, if not the best card to use for digital speech. It was the only card (from the ones I used) that don't produce the "cutoff" noise after someone stops talking.

Even DOSBox has these "cutoff" noises...

It doesn't have a wavetable header, but why not just get a external Roland Sound Canvas? You will avoid the cheap Sound Blaster mixer and the Sound Canvas has a line input at the rear, so you can just hook up your AWE64 GOLD through RCA leads.

The AWE64 GOLD is also less noisy and doesn't "POP" when you reboot the machine

AWE64 Gold is pretty nice. The AWE32, SB16 and SB32s aren't so nice, no. 😀 You don't get 16-bit 44.1 kHz out of a SB16/AWE32 regardless of what the spec sheet says. It sounds too shitty thanks to the awful analog circuitry. This also impacts a DB's output.

But having a real OPL3 around is great. It's just as impressive to me as the best wavetable because it's so unique and different. It's also better than having terrible wavetable IMO. I don't know what the cleanest ISA card with a real OPL3 onboard is.

I'm happy with the FM that my SB Pro 2.0 puts out. It's pretty quiet and clean actually. Put one of these next to a nice MIDI solution and you can run any DOS game ever made no problem. The lack of TSRs and PnP is a major plus too (software setup of any kind for that matter).

Yea the OPL3 argument is also quite interesting...

All the later games support MT-32 or General Midi anyway and the really old games that only have OPL music came out before the SB16 and sound most authentic on a old Sound Blaster because they didn't have the low and high pass filters...

So for me it's either a really old Sound Blaster for really old OPL only games, or get a really new one that has as clean as possible digital speech in combination with a Roland MT-32 or Sound Canvas.

Also the AWE64 GOLD has SPDIF, so you can get digital speech without any noise if you like! I haven't tested it, because my PC sound card does only optical and not coax.

I will do a FM music comparision video soon though. I always wondered "how different" the AWE64 FM sounds compared to the OPL chip.

It's funny that people insist on OPL3 for "authenticity" yet don't mind playing on hardware that didn't even exist when the game they are playing came out... 😉

I think it's simply because SB16s are sooo easy to find. They are everywhere on ebay for next to nothing...

Mau1wurf1977 wrote:

Also the AWE64 GOLD has SPDIF, so you can get digital speech without any noise if you like!

I thought the AWE32's SPDIF was only connected to midi. Is it different on the AWE64?