Strange issues with the VxD driver... Windows randomly plays the windows start-up sound, or other previously played sounds like right-click menu sounds, there are random crackling sounds etc. Would love to go back to WDM, but there is just no sound with it.
I'm wondering if anyone had similar issues with Audigy 2 ZS?

Not sure if this is reason for the lack of audio with WDM, but when using the WDM driver, Windows gives an error about a missing DLL C:\Windows\System\drmstor.dll on start up.
No such error with VxD driver.

VXD is the driver of choice for this era of machines due to the higher performance and better support for sound in 16-bit applications and the guides focus on this driverset. It's possible that the shortcut install method used doesn't result in functioning WDM drivers, I only ever used the retail install CD for WDM and the shortcut/easy method for VXD. The issues you are having don't seem to be VXD specific either as people don't report crackling or repeated sounds as a normal experience.

My guess is whatever was messing with the graphics card, is now doing the same thing to the sound card.

Do you have Win98 or Win98SE?

Celeron 466 mostly shipped with Win98SE or Win2K.
The Audigy2 is supported in Win98SE-WinXP.

So if you trying Win98 (original) then you are out of step on both components.

MattRocks wrote on 2025-12-01, 11:00:

Do you have Win98 or Win98SE? […]
Show full quote

Do you have Win98 or Win98SE?

Celeron 466 mostly shipped with Win98SE or Win2K.
The Audigy2 is supported in Win98SE-WinXP.

So if you trying Win98 (original) then you are out of step on both components.

Celeron 466 and Audigy 2 VXD drivers do not work with Win98, only Win98SE? Seems odd to me. 😀

NeoG_ wrote on 2025-12-01, 07:24:

... people don't report crackling or repeated sounds as a normal experience ...

I have a thread almost dedicated to EMU10K1 crackling in Win98 like a saucepan - that is how I remember those Creative DSPs behaving on similar spec CPUs.

I don't recall problems on Linux or on *GHz CPUs so I suspect crackling stems from Big WAVs + Big Creative Labs drivers vs. Slow IDE HDDs + Slow CPUs. My view is that in that era people only bought Creative Labs cards because Creative Labs sponsored the games studios. Unless you specifically want EAX demos there is probably a sweeter sound card to pair.

There are utilities that measure latency.
For example, LatencyMon.
But not for Windows 98, unfortunately.
With their help, you can determine which driver or programme is abusing the bus and CPU time.
It could be anything.

shevalier wrote on 2025-12-01, 11:32:

There are utilities that measure latency. For example, LatencyMon. But not for Windows 98, unfortunately. With their help, you c […]
Show full quote

There are utilities that measure latency.
For example, LatencyMon.
But not for Windows 98, unfortunately.
With their help, you can determine which driver or programme is abusing the bus and CPU time.
It could be anything.

Win98SE is the right environment to surface how good (or bad) Creative Labs' EMU-based sound cards really are.

LatencyMon is for Win7, and EMU hardware acceleration (most of the driver stack) is disabled on Win7 -Microsoft removed hardware DirectSound and forced all mixing through a software audio engine that emerged with Windows Vista. So if you run LatencyMon on Win7, you’re not testing Creative’s stack at all — you’re testing Microsoft’s software mixer. And if Win7 is your environment for experiencing Creative Labs' EMU-based sound output then you have only ever heard the DAC, because everything else Creative Labs shipped was disabled or bypassed.

In the war between MacOS and Windows, maybe Microsoft considered Creative Labs a liability that needed silencing? I'm not joking - my recollection of period accurate Creative Labs noise is truly dire, and if Creative were as great as people seem to believe then why did Apple - the multimedia experts - never buy a Creative chip? And, Creative didn't compete in music studios either. If Creative had any credibility in audio fidelity then you'd expect at least some amateur musicians to sometimes choose Creative over M-Audio, Ensoniq, Digigram, Turtle Beach - they did not!

The differentiator is that numerous video games specifically targeted Sound Blaster boards - and that happened because Creative Labs provided free tools, middleware, and technical support to games studios. Here's the thread I started to explore our collective memory of these sound card issues, and I'm not winning the debate but I'm also not letting anyone rewrite my recollections 😉

marxveix wrote on 2025-12-01, 11:22:

MattRocks wrote on 2025-12-01, 11:00:

Do you have Win98 or Win98SE? […]
Show full quote

Do you have Win98 or Win98SE?

Celeron 466 mostly shipped with Win98SE or Win2K.
The Audigy2 is supported in Win98SE-WinXP.

So if you trying Win98 (original) then you are out of step on both components.

Celeron 466 and Audigy 2 VXD drivers do not work with Win98, only Win98SE? Seems odd to me. 😀

Err.. I'll point to someone more qualified to me..
Guide: Installing Windows 9x and DOS drivers on Audigy cards (version 3.1)

Thanks a lot for the responses guys. I realized I haven't given you enough information, sorry about the omission.

I first installed the drivers from https://www.vogonsdrivers.com/getfile.php?fil … 130&menustate=0 by running the installer, no shortcut/easy methods involved. It installed DirectX 9 and WDM driver. After reboots, I switched to VxD driver using Creative's Driver Utility Program in the start menu.

I also did a similar installation with the modified drivers from https://www.vogonsdrivers.com/getfile.php?fil … 169&menustate=0 and in another attempt followed the corresponding guide MattRocks also linked to Guide: Installing Windows 9x and DOS drivers on Audigy cards (version 3.1)

If NeoG_ is right that this all comes down to something being wrong with the PCI slots #2-#5, I'm not sure if there's anything I can do about it. To test it, maybe I can get an AGP card and install the Audigy in PCI slot #1.

I do have Windows 98SE installed.

Coming back to what actually is happening with VxD, this doesn't sound like a CPU bottleneck. When Windows starts, I heard the opening music from the Plus! theme, and it plays perfectly fine. I right click on the desktop, and I heard the sound associated with it perfectly again. I right-click again, and no sound this time, or in any following attempts. However in about 30secs to 1 min, the sounds play randomly and repeatedly. And I also hear static-like noise from time to time.
When I play something in Winamp, it plays flawlessly. However, I still get additional static-like noise and rarely crackling sounds from time to time, and Windows sounds still play randomly.

With WDM, there is no sound at all.

I do have a quick way of restoring Windows to a known good-state, by the way (because Windows reinstallation through C:\WIN98 folder [after deleting everything other than COMMAND.COM. IO.SYS, MDDOS.SYS and WIN98 in the C: drive] is pretty slow on this machine). I made a backup before installing any drivers on my modern Linux computer, and when I need to restore to that, I just plug the SSD to it through a USB enclosure, and use rsync as

1rsync -av --delete /backup/path    /windows/c/drive/path 

By the way, I attached an atto disk benchmark for the SSD connected via StarTech IDE2SAT2 + 40-wire IDE cable to the UDMA33 motherboard. I was hoping it could at least saturate 33MB/s but for some reason, it doesn't.
One issue for sure is the first partition is not properly aligned to 2048; to get the BIOS to to Windows, the first partition needs to start at sector 63

1Disk /dev/sde: 111.79 GiB, 120034123776 bytes, 234441648 sectors
2Disk model:                 
3Units: sectors of 1 * 512 = 512 bytes
4Sector size (logical/physical): 512 bytes / 4096 bytes
5I/O size (minimum/optimal): 4096 bytes / 4096 bytes
6Disklabel type: dos
7Disk identifier: 0x00000000
8
9Device     Boot    Start       End   Sectors  Size Id Type
10/dev/sde1  *          63  65529134  65529072 31.2G  c W95 FAT32 (LBA)
11/dev/sde2       65529856 208136191 142606336   68G  c W95 FAT32 (LBA)

but despite that, it is still very fast on the modern PC, so that is probably not the bottleneck.

hydraphone wrote on 2025-12-01, 18:01:

One issue for sure is the first partition is not properly aligned to 2048;

Partition alignment is more related to wear than speed.
As far as I remember, the difference in throughput is 5%.

Try to use aida64 version 5.
Its got disk bench utility, those have cpu utilization monitoring

shevalier, thanks for the suggestion, I will try that. Disk speed is a luxury at this point though 😀 Will probably use the PCI SATA card eventually.

Going back to sound, fresh Windows 98SE install, Audigy in PCI slot #4, following Guide: Installing Windows 9x and DOS drivers on Audigy cards (version 3.1) again, computer freezes at step 3.3, there is the following tiny window on screen

New Hardware Found
PCI OHCI Compliant IEEE 1394 Host Controller

Reboot results in the same freeze. Tried using safe mode to disable the firewire in the device manager, but it is not listed in the device tree yet.

Different IRQs for different PCI slots - do you find the same result with every PCI slot? It might also be worth disabling on board stuff you don't use (i.e. parallel prt) to free up IRQs.

Found a post here that might help: https://msfn.org/board/topic/84886-the-comple … ws-98se/page/2/
None of the fixes I read seem to match your exact symptoms, but there may be other fixes in this: http://www.mdgx.com/spx/NUSB23E.EXE

I haven't used it myself.

hydraphone wrote on 2025-12-01, 17:53:

If NeoG_ is right that this all comes down to something being wrong with the PCI slots #2-#5, I'm not sure if there's anything I can do about it. To test it, maybe I can get an AGP card and install the Audigy in PCI slot #1.

The reason why I think this is because of the repeated sounds. The crackling and noise could be anything but the repeated sounds indicate the card is roaming over a memory area because the computer has lost control of it. The fact that it randomly stops producing sound on demand also points to this.

MattRocks, unfortunately, yes, same results with PCI slots #2-#5.
The link you have seems to be dead, and it doesn't seem to be archived on web.archive.org

New attempt:

Fresh install -> chipset drivers -> renamed C:\Windows\INF\1394.inf -> did full installation from https://www.vogonsdrivers.com/getfile.php?fil … 130&menustate=0

For the first time, the Audigy Driver installation completed successfully, and it installed Windows Media Player as well, so there was no issue with drmstor.dll. After reboots, no sound again with the WDM driver. Switch to VxD again using Driver Utility Program from Start menu, a couple reboots, same weird sound issues.

NeoG_, that gives me another idea: if I can find a way to run Linux on this computer , I would be able to see whether the computer keeps losing control of the Audigy or not, it should show up in dmesg. I suppose I can install Arch Linux 32 (or maybe some old ISO if nouveau driver dropped FX 5200 and Audigy 2 ZS is also dropped from the kernel, need to check this), via 86Box to an .img HDD and dd it to SSD again.

Perhaps the Audigy card itself is problematic? I bought it as used. I suppose I can check this by buying another PCI sound card.

hydraphone wrote on 2025-12-01, 19:58:

shevalier, thanks for the suggestion, I will try that. Disk speed is a luxury at this point though 😀 Will probably use the PCI SATA card eventually.

Testing the drive in AIDA64 is not to determine the read speed, but to check the CPU load.
If the speed is low and the load is high, most likely the motherboard has not enabled UDMA and is operating in PIO mode.

As for using a PCI video card with Audigy, well, that's not a very good idea.
It's not a port, it's a bus.
Its bandwidth is shared among all connected devices.
And Audigy is very sensitive to this.
In addition, nVidia drivers from those years did strange things with priorities and interfered with chipset settings.

I would install 512MB of RAM (the maximum for LX, although it depends on the motherboard), install Windows XP, and see how it all works on a newer operating system.

By the way, did you install the driver for the motherboard chipset?
Windows XP recognises the 440 chipset well, but I'm not so sure about Windows 98.

hydraphone wrote on 2025-12-01, 22:41:

I suppose I can check this by buying another PCI sound card.

6ALX
I440X Main Board 23
SB-LINK is used to attach any “PC/PCI” standard sound card like Creative
AWE64D or Yamaha XG…for compatibility under DOS mode

If you find a Yamaha with a soldered SB-Link header, it will greatly expand the application options for this motherboard.

shevalier wrote on 2025-12-02, 07:02:

6ALX I440X Main Board 23 SB-LINK is used to attach any “PC/PCI” standard sound card like Creative AWE64D or Yamaha XG…for compat […]
Show full quote

hydraphone wrote on 2025-12-01, 22:41:

I suppose I can check this by buying another PCI sound card.

6ALX
I440X Main Board 23
SB-LINK is used to attach any “PC/PCI” standard sound card like Creative
AWE64D or Yamaha XG…for compatibility under DOS mode

If you find a Yamaha with a soldered SB-Link header, it will greatly expand the application options for this motherboard.

SB-LINK reduces the gap between PCI and ISA cards in pure DOS only - it wouldn't do anything to fix Windows audio.

I don't think it's PCI bus issues - audio should be tiny amount of data when busy, and idle most of the time. I think it's Audigy driver issues. Having the Firewire port lockup when there is nothing attached to the firewire port is very telling. Something is going wrong during initialisation.

MattRocks wrote on 2025-12-02, 07:14:

shevalier wrote on 2025-12-02, 07:02:

6ALX I440X Main Board 23 SB-LINK is used to attach any “PC/PCI” standard sound card like Creative AWE64D or Yamaha XG…for compat […]
Show full quote

hydraphone wrote on 2025-12-01, 22:41:

I suppose I can check this by buying another PCI sound card.

6ALX
I440X Main Board 23
SB-LINK is used to attach any “PC/PCI” standard sound card like Creative
AWE64D or Yamaha XG…for compatibility under DOS mode

If you find a Yamaha with a soldered SB-Link header, it will greatly expand the application options for this motherboard.

SB-LINK reduces the gap between PCI and ISA cards in pure DOS only - it wouldn't do anything to fix Windows audio.

it will greatly expand the application options for this motherboard.

The 440LX fundamentally does not support processors with SSE instructions.
When an application using these instructions is launched for the first time, these motherboards freeze.

Therefore, only Celeron Mendocino is supported.
And that is essentially DOS+Windows 98, by and large.
And nothing more

I don't think it's PCI bus issues - audio should be tiny amount of data when busy, and idle most of the time. I think it's Audigy driver issues. Having the Firewire port lockup when there is nothing attached to the firewire port is very telling. Something is going wrong during initialisation.

Perhaps, but graphics cards and sound cards are also very sensitive to delays as such.
All this needs to be checked only by replacement and experimentation.