I tried using my SC55 under DOS as MIDI with DOOM and DOOM2, and hanging notes were really obvious. Caused me to put a separate MPU401 interface in my system. Try some DOS games.

Ozzuneoj wrote:

If anyone has experienced these issues with a PAS16 in the past, I'd just like to know in what
games and under what circumstances they experienced issues.

Hmm.. I was a long time user of the PAS16, but had no MIDI synth back in the day, sorry. 🙁
Anyway, I vaguely believe to remember that the PAS16 uses the AdLib port as its control port (388h/389h).
In a dual soundcard test with a SB16 this caused a conflict with the SB16's OPL3 chip during init of MVSOUND.SYS:
A horrible noise came out of the SB16. Running SB16's DIAGNOSE.EXE afterwards fixed the issue, though,
and I was able to use both cards (even with the Thunderboard enabled, but on an oddball adress).
This makes me believe that the PAS16 has a special logic for its on-board OPL3,
also because the PAS16 had been designed to work with multiple sisters of its own kind in a single PC.
That being said, I didn't do much tests with the MPU-401 interfaces at the time.
I had a DB50XG on the SB16 installed, though, I believe. If memory serves, it worked fine under Descent.

Just wanted to bump this thread. I've had some interest in this again but haven't had time to really dig into it. If anyone has recordings or hands-on experience with this problem I'd like to hear about it.

Wow... it has been FIVE YEARS since my list post in this thread and over SIX since I started it. Kind of surprised there haven't been any more posts from people wandering in here after messing with a PAS16.

Anyone have any recordings or detailed experiences with running a Media Vision card and experiencing hanging notes in MIDI? If so, please specify the model of card in as much detail as possible (preferably a picture of it), the synth being used, the game\application with hanging notes and your impression of just how bad the problems were. Bonus points if you can compare it to one of the more well known types of hanging notes experienced on Sound Blaster cards (type 1, type 2, etc.).

The lack of feedback can probably be partially explained by the fact that few folks seem to use the PAS16 at all today, and fewer still have used it for external MIDI functionality.

For my part, the hanging note behavior was never consistent enough for it to happen in a 100% repeatable manner, but it would occur without fail during the course of either general gameplay, or even just playing MIDI files using various, DOS-based players. This is a fundamental difference compared to the MIDI issues of the SB16 MPU architecture, which only manifest during concurrent digital-audio playback.

The problem has been described as being a flaw/deficiency of the MVD101D chip itself, affecting all PAS16 variants with MPU-401 UART functionality. It's also been described as a problem without a known solution, though I suspect that just altering MIDI drivers/routines to provide for a longer or even interminable busy/wait MPU timeout might suffice (at the potential expense of gameplay slowdown).

You'd mentioned not experiencing any trouble in Windows; I'd suggest doing any testing from DOS instead. Windows' multimedia layer and even the soundcard drivers themselves will often leverage additional software MIDI buffering, which would mask the problem.

Finally, the USENET archives contain several hundred messages concerning PAS16 MIDI troubles. Here are just a few quotes:

PAS16 and Stuck Notes
"I finally got through to the Technical Support people at Media Vision
and they explained to me that if you try to send more than about 6
channels of data simultaneously through the external midi port, the
buffers will get overloaded and you will experience stuck notes on the
synthesizer."

PAS 16 "dropping" MIDI control messages?
"After many hours on the phone with Media Vision, I talked
with one of the "experts" in MIDI there. He said that it was a known
problem that the PAS-16 would occasionally drop MIDI messages when
dealing with a large amount of data (in the same way that a 16450
UART will overflow its 1-character buffer if an interrupt is missed,
occasionally the PAS-16 will drop MIDI control messages)."

PAS-16 is half-baked?
"Supporting only MPU-401 uart mode only is ok, but it occasionally
drops a byte here or there. Their reason? They really didn't expect
we would USE it... All it takes is one dropped byte to create one
stuck note and ruin the entire sequence."

Pro Audio Spectrum-MIDI?
"Well, not to tell the truth here or anything, I used to work at MV,
have LOTS of experience with the PAS16. The MIDI FIFO (buffer) on the
thing is fundamentally busted and typically results in lots of hung
notes."

Stuck Midi Notes/PAS-16
"I've also had similar problems with 'stuck notes' using the PAS-16's
external MIDI so you aren't alone. I've talked to MediaVision tech
support twice--they're aware of the problem but don't have a solution."

PAS 16 and MIDI
"I tried using PAS16 thru a Midimate(their interface). Got a lot of
hung notes (probably some missed notes as well, not as easy to hear.)
Mediavision's "midi specialist" admitted he used an MPU-401 for the
same reason."

PAS16 MidiMate Internals
"Unfortunately, the PAS 16 MIDI port simply doesn't work right. It
generates random notes that hang. After calling at least a dozen
times, I finally got someone to duplicate the problem at MV. Their
engineering department is currently investigating the problem."

Problems with PAS16 Midi connection
"The PAS-16's MIDI port is notorious for dropping MIDI events."
"This is a known problem with the chip that does the MPU401 emulation on
the PAS16 board. It will occasionally lose bytes in the midi stream
causing wrong notes, missed note ons, missed note offs, wrong velocity,
etc."

Any WaveTable daughter board??
"Media Vision produced a sound card (initially intended to be the Pro
Audio Studio XL) called the Pro Audio Spectrum 16 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."

Cloudschatze wrote on 2025-03-24, 04:57:

The lack of feedback can probably be partially explained by the fact that few folks seem to use the PAS16 at all today, and fewe […]
Show full quote

The lack of feedback can probably be partially explained by the fact that few folks seem to use the PAS16 at all today, and fewer still have used it for external MIDI functionality.

For my part, the hanging note behavior was never consistent enough for it to happen in a 100% repeatable manner, but it would occur without fail during the course of either general gameplay, or even just playing MIDI files using various, DOS-based players. This is a fundamental difference compared to the MIDI issues with the SB16 MPU architecture, which only manifest during concurrent digital-audio playback.

The problem has been described as being a flaw/deficiency of the MVD101D chip itself, affecting all PAS16 variants with MPU-401 UART functionality. It's also been described as a problem without a known solution, though I suspect that just altering MIDI drivers/routines to provide for a longer or even interminable busy/wait MPU timeout might suffice (at the potential expense of gameplay slowdown).

You'd mentioned not experiencing any trouble in Windows; I'd suggest doing any testing from DOS instead. Windows' multimedia layer and even the soundcard drivers themselves will often leverage additional software MIDI buffering, which would mask the problem.

Finally, the USENET archives contain several hundred messages concerning PAS16 MIDI troubles. Here are just a few quotes:

PAS16 and Stuck Notes
"I finally got through to the Technical Support people at Media Vision
and they explained to me that if you try to send more than about 6
channels of data simultaneously through the external midi port, the
buffers will get overloaded and you will experience stuck notes on the
synthesizer."

PAS 16 "dropping" MIDI control messages?
"After many hours on the phone with Media Vision, I talked
with one of the "experts" in MIDI there. He said that it was a known
problem that the PAS-16 would occasionally drop MIDI messages when
dealing with a large amount of data (in the same way that a 16450
UART will overflow its 1-character buffer if an interrupt is missed,
occasionally the PAS-16 will drop MIDI control messages)."

PAS-16 is half-baked?
"Supporting only MPU-401 uart mode only is ok, but it occasionally
drops a byte here or there. Their reason? They really didn't expect
we would USE it... All it takes is one dropped byte to create one
stuck note and ruin the entire sequence."

Pro Audio Spectrum-MIDI?
"Well, not to tell the truth here or anything, I used to work at MV,
have LOTS of experience with the PAS16. The MIDI FIFO (buffer) on the
thing is fundamentally busted and typically results in lots of hung
notes."

Stuck Midi Notes/PAS-16
"I've also had similar problems with 'stuck notes' using the PAS-16's
external MIDI so you aren't alone. I've talked to MediaVision tech
support twice--they're aware of the problem but don't have a solution."

PAS 16 and MIDI
"I tried using PAS16 thru a Midimate(their interface). Got a lot of
hung notes (probably some missed notes as well, not as easy to hear.)
Mediavision's "midi specialist" admitted he used an MPU-401 for the
same reason."

PAS16 MidiMate Internals
"Unfortunately, the PAS 16 MIDI port simply doesn't work right. It
generates random notes that hang. After calling at least a dozen
times, I finally got someone to duplicate the problem at MV. Their
engineering department is currently investigating the problem."

Problems with PAS16 Midi connection
"The PAS-16's MIDI port is notorious for dropping MIDI events."
"This is a known problem with the chip that does the MPU401 emulation on
the PAS16 board. It will occasionally lose bytes in the midi stream
causing wrong notes, missed note ons, missed note offs, wrong velocity,
etc."

Any WaveTable daughter board??
"Media Vision produced a sound card (initially intended to be the Pro
Audio Studio XL) called the Pro Audio Spectrum 16 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."

Thank you! That helps a lot, especially reading the messages posted back when they were still being supported by MV.

What a massive blunder that was to have broken MIDI UART support on something that ended up having SO many accessories and versions released which focused on MIDI and multimedia features.

For some reason I just have this inexplicable desire to see these cards redeemed after all these years and for a workaround to be found for this... whether it be a hardware mod, a system spec change or a software solution. I can't shake the feeling that this must not have existed during testing and that it cropped up later once the system got out to more users. The PAS16's entire lifespan was only a couple years, during which time many many things were happening with the company that could have prevented them from finding a real solution to it. I highly doubt that the company's engineers were allowed to invest a huge amount of time into fixing this since they basically just kept pumping out new products and variations until the doors were shut due to bankruptcy.

I think my interest in MediaVision products stems from a few things:
1. Their products themselves (cards, external devices, etc.) are extremely nicely built and the quality seems far higher than most sound cards produced during the time. For example, tantalum caps are often used in favor of cheap aluminum eletrolytics.
2. The underdog story of a company that was doing okay but just fell apart due to incompetent\corrupt management and stiff competition from Creative.
3. The fact that they were the precursor to Aureal, which was, yet another underdog story of a company being ruined by Creative.
4. Because of 2 and 3 I see them sort of like 3dfx... or maybe like Nvidia if the tables were turned and 3dfx somehow stomped them out early on because of market dominance.

If I had any of the needed skills to work on a fix I'd love to take a stab at it... I just don't have anything of the sort, sadly.

What I do have is a fairly large variety of Media Vision cards and a huge variety of hardware (motherboards, CPUs and MIDI modules) on which to test them. If anyone has the skills and desire to even attempt a software or hardware-mod solution to work around this issue, I would help in any way that I could to narrow down any hardware + software combinations that exhibit the problem more or less frequently.

Do any of the sound card collectors of Vogons have any old not-intended-for-the-public software or documentation from MV that may be of some help with this? Or, does anyone have any really odd engineering sample cards or very late models that may not exhibit the issue?

Along those same lines, Cloudschatze, do you know if any of MV's other products aside from the PAS16 have the same hanging note issue? Like, the Jazz16 for example, or the pre PAS16 cards like the PAS Plus?

... just thinking out loud, but has anyone ever tried using SoftMPU to see if it had any impact on a PAS16 having hanging notes?

Ozzuneoj wrote on 2025-03-24, 05:33:

Along those same lines, Cloudschatze, do you know if any of MV's other products aside from the PAS16 have the same hanging note issue? Like, the Jazz16 for example, or the pre PAS16 cards like the PAS Plus?

The MVD101 chip found on the PAS Plus, PAS16 rev. C, and CDPC models lacks the MPU-401 UART-mode functionality of the MVD101D, so it's a bit more difficult to determine if, or how affected it might be. The required use of Media Vision's API, SDK, and/or native drivers for MIDI playback with the MVD101 would presumably have minimized or circumvented any hardware-related issues.

The MVD1216 is a similar grey area. There are fewer reports of MIDI issues associated with cards bearing that chip compared to those with the MVD101D, but still more than zero. Furthermore, the co-designer of the Pro Audio 3D, Media Vision's John Neary, mentioned the following in some correspondence I'd had with him many years ago:

"The Pro3D was the pinnacle of Media Vision before the engineering group was whacked after Paul Jain sank the company.... which was quite a nice card. 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."

So, purportedly problematic, but perhaps to a much-lesser degree. I can't recall having personally experienced any MIDI issues with the Pro 3D, for what it's worth.

Cloudschatze wrote on 2025-03-24, 18:08:

The MVD101 chip found on the PAS Plus, PAS16 rev. C, and CDPC models lacks the MPU-401 UART-mode functionality of the MVD101D, s […]
Show full quote

Ozzuneoj wrote on 2025-03-24, 05:33:

Along those same lines, Cloudschatze, do you know if any of MV's other products aside from the PAS16 have the same hanging note issue? Like, the Jazz16 for example, or the pre PAS16 cards like the PAS Plus?

The MVD101 chip found on the PAS Plus, PAS16 rev. C, and CDPC models lacks the MPU-401 UART-mode functionality of the MVD101D, so it's a bit more difficult to determine if, or how affected it might be. The required use of Media Vision's API, SDK, and/or native drivers for MIDI playback with the MVD101 would presumably have minimized or circumvented any hardware-related issues.

The MVD1216 is a similar grey area. There are fewer reports of MIDI issues associated with cards bearing that chip compared to those with the MVD101D, but still more than zero. Furthermore, the co-designer of the Pro Audio 3D, Media Vision's John Neary, mentioned the following in some correspondence I'd had with him many years ago:

"The Pro3D was the pinnacle of Media Vision before the engineering group was whacked after Paul Jain sank the company.... which was quite a nice card. 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."

So, purportedly problematic, but perhaps to a much-lesser degree. I can't recall having personally experienced any MIDI issues with the Pro 3D, for what it's worth.

Man, you are a wealth of Media Vision knowledge... 😅

I'm speaking with near-zero knowledge of the subject of buffering MIDI data, but I wonder if it's possible to circumvent this issue with a hardware mod or add-on to the MVD chip. Since the company was being driven into the ground very quickly, I doubt that any post-production modifications would have ever been feasible back in the day even if it turned out to be relatively simple. Of course, a software solution would be cool too, but hardware mods are nice and permanent.

I also have a Pro Audio 3D I could test at some point, along with the various PAS16 cards.

Ozzuneoj wrote on 2025-03-24, 05:33:

What a massive blunder that was to have broken MIDI UART support on something that ended up having SO many accessories and versions released which focused on MIDI and multimedia features.

Hi, I assume that the MPU-401 wasn't being supposed to be really used in first place.
Or at least not in a sophisticated way, but rather as a means to hook up, say, a Casio keyboard and run a MIDI composer such as CuBase.

The PAS16 itself initially was more of a rival/challenger to the SB Pro/SB Pro 2, -which competed with PAS/PAS Plus-, which didn't feature an actual MPU-401 port but had SB-MIDI only.
The PAS16 later became an SB16 rival, obviously.

This sometimes reminds me of the relationship between Sega Mega Drive (Genesis) and the Super Nintento in the 16-Bit era.
Most people today compare the MegaDrive with the Super Nintendo and see the inadequacies.
However, the Mega Drive was older and was meant to compete with the 8-Bit NES in first place, before the Super NES was out.
In its role it was an updated version of the Sega MasterSystem, rather.

Ozzuneoj wrote on 2025-03-23, 06:27:

Wow... it has been FIVE YEARS since my list post in this thread and over SIX since I started it. Kind of surprised there haven't been any more posts from people wandering in here after messing with a PAS16.

Well i probably missed this thread, but in my case, and owning a Pro Audio Studio XL i admit i never tried to use it for MIDI, daughterboards or not, because i was already aware (also courtesy of Cloudschatze) of its buggy MPU-401 implementation.

Jo22 wrote on 2025-03-24, 20:03:

Hi, I assume that the MPU-401 wasn't being supposed to be really used in first place. Or at least not in a sophisticated way, bu […]
Show full quote

Ozzuneoj wrote on 2025-03-24, 05:33:

What a massive blunder that was to have broken MIDI UART support on something that ended up having SO many accessories and versions released which focused on MIDI and multimedia features.

Hi, I assume that the MPU-401 wasn't being supposed to be really used in first place.
Or at least not in a sophisticated way, but rather as a means to hook up, say, a Casio keyboard and run a MIDI composer such as CuBase.

The PAS16 itself initially was more of a rival/challenger to the SB Pro/SB Pro 2, -which competed with PAS/PAS Plus-, which didn't feature an actual MPU-401 port but had SB-MIDI only.
The PAS16 later became an SB16 rival, obviously.

This sometimes reminds me of the relationship between Sega Mega Drive (Genesis) and the Super Nintento in the 16-Bit era.
Most people today compare the MegaDrive with the Super Nintendo and see the inadequacies.
However, the Mega Drive was older and was meant to compete with the 8-Bit NES in first place, before the Super NES was out.
In its role it was an updated version of the Sega MasterSystem, rather.

That is a good point, for sure.

Still, MediaVision's chips were used on several products with dedicated MIDI playback hardware. For example, the Jazzwave (Jazz + ICS Wavefront), the Pro Audio 3D apparently had an optional KORG daughterboard and the Logitech Soundman Wave was a PAS16 with a Yamaha OPL4. They also sold dedicated MIDI breakout boxes and the Memphis multimedia kit also had full size MIDI in\out connections.

It is, of course, totally feasible that no one ever got any bug-free MPU-401 MIDI out of an MV product, it just seems strange that they would go to so much effort to put out so many products while knowing they didn't work right. At least for all of Creative's faults they didn't put out a bunch of first-party SB16 cards that were specifically built to around the bugged MPU-401 playback.

Those were definitely crazy times though. Between sound cards and video cards, the 90s were the wild-west era of PC hardware. 😀