i'm having a problem in win98se that i think i've isolated to midiplayer6. first, i use midiplayer6 extensively with midiyoke/midiox and roland/yamaha serial drivers as a central hub for all my midi interconnections as well as being my defacto vst/sf2 engine. i've never had a problem except when it comes to connecting 2 softsynths: invision cybersynth 3.0 and seer systems reality v1.55 (vsc, s-yxg50/100, munt, timidity, sw-10, wingroove, wavecube, websynth d77, fpd98 all give no issues).

seer systems reality is probably the easier of the 2 to tackle. it publishes its own entry in the windows midi mapper and also can be configured to accept input on a midi-in port. when i try to play a mid file in midiplayer6 using reality all i get is a smeared mess of gibberish for output. it doesn't matter if i connect using the windows midi mapper or via midiyoke.

here is the interesting part...if i use dosbox to connect to reality using the windows midi mapper or via midiyoke--playback is normal. additionally, i can use vdmsound for win9x and get the same normal playback i get in dosbox. when i go back to midiplayer6 it's unintelligible. so that's 2 independent midi clients that can interface successfully with reality.

midiplayer6 is very well authored. with all the use cases i throw at it i can't imagine it's doing something non-standard protocol-wise. i surmise that reality is expecting input that adheres to a non-conventional format that apparently dosbox and vdmsound are able to squeak by with. there is next to nothing to re-configure in reality...what can i try to fix this from the midiplayer6 end of things?

stanwebber wrote on 2026-07-10, 14:24:

i'm having a problem in win98se that i think i've isolated to midiplayer6. first, i use midiplayer6 extensively with midiyoke/mi […]
Show full quote

i'm having a problem in win98se that i think i've isolated to midiplayer6. first, i use midiplayer6 extensively with midiyoke/midiox and roland/yamaha serial drivers as a central hub for all my midi interconnections as well as being my defacto vst/sf2 engine. i've never had a problem except when it comes to connecting 2 softsynths: invision cybersynth 3.0 and seer systems reality v1.55 (vsc, s-yxg50/100, munt, timidity, sw-10, wingroove, wavecube, websynth d77, fpd98 all give no issues).

seer systems reality is probably the easier of the 2 to tackle. it publishes its own entry in the windows midi mapper and also can be configured to accept input on a midi-in port. when i try to play a mid file in midiplayer6 using reality all i get is a smeared mess of gibberish for output. it doesn't matter if i connect using the windows midi mapper or via midiyoke.

here is the interesting part...if i use dosbox to connect to reality using the windows midi mapper or via midiyoke--playback is normal. additionally, i can use vdmsound for win9x and get the same normal playback i get in dosbox. when i go back to midiplayer6 it's unintelligible. so that's 2 independent midi clients that can interface successfully with reality.

midiplayer6 is very well authored. with all the use cases i throw at it i can't imagine it's doing something non-standard protocol-wise. i surmise that reality is expecting input that adheres to a non-conventional format that apparently dosbox and vdmsound are able to squeak by with. there is next to nothing to re-configure in reality...what can i try to fix this from the midiplayer6 end of things?

Hi,
I do not have Seer Systems Reality v1.55 so I could only test the still available v1.56 demo version that I could download from here:
https://seersystems.com/products/downloads/
According to my tests there is no problem with Midi player. The problem is that the default bank of Seer Systems Reality is absolutely NOT a GM/GS/XG compatible bank but a rather custom one. You should also test other Midi players that can select the Midi Out port explicitly such as Van Basco Karaoke Player. They should sound the same. Here is my test video how Grabbag.mid from Duke Nukem 3d sounds with MidiPlayer and Van basco when Seer's Reality is used so you can compare to yours:
https://youtu.be/vcdU62In6TQ

The real question is how it sounded good with yours DosBox. I do not know the answer since it should not. But I suspect that in reality they did not use Reality at all 😀.
Most likely they used another Midi output port.

reality has gm banksets available. on that same download page as the reality demo you can obtain the surreal demo which uses reality as a backend (folder structure almost identical and overlapping so it might overwrite). the surreal demo has 2 gm banksets located in program files\...\banksets\ that you can plug into reality.

i have all the demo banksets (both reality & surreal) and have tried them without any change in outcome. obviously, the gm banksets will sound correct, but even with the reality demo bankset the sounds may be incorrect, but at least they should still sound like instruments. when i say a smeared mess of gibberish i really mean it. it's nigh impossible to pick out any individual notes or identify any musical instruments known to man.

the fact that you got the reality demo to play and could at least recognize the instruments were wrong is disheartening as it probably means something localized on my system. midiplayer6 is flawless in every other regard on my system...i can't for the life of me figure out what could be wrong. i will try the install on some alternate hardware and see what happens.

stanwebber wrote on 2026-07-10, 17:19:

reality has gm banksets available. on that same download page as the reality demo you can obtain the surreal demo which uses reality as a backend (folder structure almost identical and overlapping so it might overwrite). the surreal demo has 2 gm banksets located in program files\...\banksets\ that you can plug into reality.

i have all the demo banksets (both reality & surreal) and have tried them without any change in outcome. obviously, the gm banksets will sound correct, but even with the reality demo bankset the sounds may be incorrect, but at least they should still sound like instruments. when i say a smeared mess of gibberish i really mean it. it's nigh impossible to pick out any individual notes or identify any musical instruments known to man.

the fact that you got the reality demo to play and could at least recognize the instruments were wrong is disheartening as it probably means something localized on my system. midiplayer6 is flawless in every other regard on my system...i can't for the life of me figure out what could be wrong. i will try the install on some alternate hardware and see what happens.

Do you get the wrong sound also when you play notes on the virtual keyboard of Midi Player with your mouse? Or only when you playback Midi files?

stanwebber wrote on 2026-07-10, 17:19:

... I will try the install on some alternate hardware and see what happens.

You have not even mentioned what hardware you use so it's hard to tell what the problem can be.
BTW, the only thing that Midi Player does that other Midi software usually do not that can effect wave out quality on old hardware is that it starts a Directsound capture thread in order Volume meter an wave recording capabilities to work.
You can try to disable this by opening the Device Settings dialog and press the minus sign (-) next to the 'Input Port' label at the 'Recording Source' section.

i installed everything on another win98se system and there were no issues with playback so, as i feared, it's something localized to my setup. the problematic system in question is a compaq armada e500 laptop (850mhz p3 256mb ram 440bx) with an ess maestro 2e (es1978s). given this soundcard, i have a strong suspicion that directsound will have something to do with the problem or solution or both. to that end i've already switched from vxd to wdm ess drivers to see if there would be any change, but alas none.

the smeared sound occurs in midiplayer6 whether playing a mid file or pressing notes on the keyboard. i also ticked the minus sign in the recording source section of the settings, but it was no help.

i did discover something new that is interesting. when i test the reality midi connection in dxdiag i get the exact same smeared output as i do in midiplayer6.

i don't know if there will be a solution for this if it all comes down to a driver/hardware incompatibility.

edit: i worked out all my issues with seer systems and cybersound softsynths. midiplayer6 was not involved at all (other than having to increase the audio buffer for my wretched soundcard to a ridiculous level).

Hi Zoltan )
Me need your help , as you can see , me very like XMPlay I think what you this understanded when see this topic Re: XMPlay Extended version , and because what https://www.un4seen.com/forum/?msg=147042 me need plugin whitch will can normally do output to external or virtual MIDI ports , I long time asked and continue asking author XMPlay , but this doesn't seem to make much sense , I offer you write plugin for XMPlay whitch will work as your FSMP , but without include SF2 synthesizer , me need plugin whitch do reset synthesizer before send MIDI stream and whitch will simply send MIDI stream to selected device as is without feedback in soundsystem XMPlay, you are can help with this ? I offer named this plugin as FSMP MIDI trannsmitter )

Why I ask you and author of XMPlay ?
This because I see what both man available source and known WinAPI for work MIDi devices , I think what this very simplifies task )

Roland User wrote on 2026-07-15, 09:28:

Hi Zoltan ) Me need your help , as you can see , me very like XMPlay I think what you this understanded when see this topic Re: […]
Show full quote

Hi Zoltan )
Me need your help , as you can see , me very like XMPlay I think what you this understanded when see this topic Re: XMPlay Extended version , and because what https://www.un4seen.com/forum/?msg=147042 me need plugin whitch will can normally do output to external or virtual MIDI ports , I long time asked and continue asking author XMPlay , but this doesn't seem to make much sense , I offer you write plugin for XMPlay whitch will work as your FSMP , but without include SF2 synthesizer , me need plugin whitch do reset synthesizer before send MIDI stream and whitch will simply send MIDI stream to selected device as is without feedback in soundsystem XMPlay, you are can help with this ? I offer named this plugin as FSMP MIDI trannsmitter )

Why I ask you and author of XMPlay ?
This because I see what both man available source and known WinAPI for work MIDi devices , I think what this very simplifies task )

I'm sorry but I'm not interested. I have many things to do nowadays and I have never written an XMPlay plugin anyway.

Thank you ) I understanded you ) , maybe some day you create plugin for XMPlay ) API in XMPlay fully open , but I not known how this do and I not known Windows API )

I inform you , this problem fixed in XMPlay version 4.1.0.13 , I was very busy when make transcoder and that's why didn't pay attention this fix )
I whish you good luck )

Hi Zoltan )
Plrease , fix bug in your WinMM Multi Port VSTi , if I long time play MIDI via your router and next want rechange synthesizer , I click on port and try do reselect other synthesizer , but when I do this , plugin freeze and MIDI stream not work , this is happening only if very long time play MIDI via selected MIDI device on port )

Roland User wrote on Yesterday, 21:00:

Hi Zoltan )
Plrease , fix bug in your WinMM Multi Port VSTi , if I long time play MIDI via your router and next want rechange synthesizer , I click on port and try do reselect other synthesizer , but when I do this , plugin freeze and MIDI stream not work , this is happening only if very long time play MIDI via selected MIDI device on port )

I could not reproduce this problem. Maybe this is port specific. I do not think I can do anything without exact steps how to reproduce the problem.
BTW, you should always stop the Midi stream before changing ports.

I think what you not do tested for talking about it , when I say about long time usage , I mean very very long time 3-6 hours , I doubt what has already passed time go from my report )

How much I can understand , this bug availabled cumulative structure and if test in short time , bug not , or this bug not replay )