I could be wrong, but weren't those arranged/made before General MIDI was published ?

Hi,
No Instrument means invalid bank + program combination. Contrary to original SC-55 the mkII has no Capital Tone Fallback feature so it remains silent on channels where invalid combination is found.
There are multiple versions of canyon.mid/passport.mid wandering on the net so it would have been useful if you had attached your versions.
I have already met with 2 problematic versions of canyon.mid and one of them can cause the exact problem you described.
It sends Bank MSB 5 + Bank LSB 87 + Program 90 to channel 1 that is an invalid combination on any known Roland gear.
the 2nd one uses Microsoft's proprietary dual authored Base + Extended channel layout thus wants to use channel 16 as drum channel.
Both problems can be cured by using FSMP as Midi player.

More info on using software Capital Tone Fallback with FSMP:
Re: Heads Up: Roland Sound Canvas VA VSTi Plugin!

More info on MS specific Base+Extended dual layout Midi files:
Re: Falcosoft Soundfont Midi Player + Munt VSTi + BassMidi VSTi

I have also attached the 2 problematic versions of canyon.mid for testing

In addition to what Falcosoft mentions, I am wondering what you were using to play the MIDI files on these computers. Was it always Windows 3.1, for instance? If so, there might be something off with the MIDI Mapper configuration if it happens to be the same between all of them. Also, in the Windows 3.1 timeframe there were different MIDI channel standards and the one set up by default by most sound card driver installations was not always the one that really makes sense to use with a GM MIDI device. There were some which used channel 16 as percussion vs. the more standard channel 10, some which only gave you channels 1-10, others which only gave you channels 11-16, etc.

Lazer42 wrote on 2021-05-22, 02:09:

In addition to what Falcosoft mentions, I am wondering what you were using to play the MIDI files on these computers. Was it always Windows 3.1, for instance? If so, there might be something off with the MIDI Mapper configuration if it happens to be the same between all of them. Also, in the Windows 3.1 timeframe there were different MIDI channel standards and the one set up by default by most sound card driver installations was not always the one that really makes sense to use with a GM MIDI device. There were some which used channel 16 as percussion vs. the more standard channel 10, some which only gave you channels 1-10, others which only gave you channels 11-16, etc.

What you described is exactly the same that I mentioned above related to the 2nd problematic file.
Just a little correction:
Base setup used channels 13-16 (not 11-16) and channel 16 was the drum channel.
Extended setup used channels 1-10 and channel 10 was the drum channel similarly to GM.
The 3rd option was the full setup, used all channels 1-16 and it was GM compatible:
https://indafoto.hu/zbacsko/image/26278493-f47e68eb/user

But none of these channel configurations can cause No Instrument errors on mkII. This is a completely unrelated issue.

won't setting the unit to GM mode (power on while holding right PART button) do the trick?

Thanks all for the replies. Very informative!

Falcosoft wrote on 2021-05-22, 01:29:

Hi, No Instrument means invalid bank + program combination. Contrary to original SC-55 the mkII has no Capital Tone Fallback f […]
Show full quote

Hi,
No Instrument means invalid bank + program combination. Contrary to original SC-55 the mkII has no Capital Tone Fallback feature so it remains silent on channels where invalid combination is found.
There are multiple versions of canyon.mid/passport.mid wandering on the net so it would have been useful if you had attached your versions.
I have already met with 2 problematic versions of canyon.mid and one of them can cause the exact problem you described.
It sends Bank MSB 5 + Bank LSB 87 + Program 90 to channel 1 that is an invalid combination on any known Roland gear.
the 2nd one uses Microsoft's proprietary dual authored Base + Extended channel layout thus wants to use channel 16 as drum channel.
Both problems can be cured by using FSMP as Midi player.

Yeah I knew about the lack of CTF on the mkII, I was just rather surprised to encounter this with these MIDI files in particular. The files I'm using are the ones that came with the Win98SE CDROM in the C:\Windows\Media folder. Checking out canyon.mid in a sequencer, it has the exact setup you describe with the bank selects and program change.

Lazer42 wrote on 2021-05-22, 02:09:

In addition to what Falcosoft mentions, I am wondering what you were using to play the MIDI files on these computers. Was it always Windows 3.1, for instance? If so, there might be something off with the MIDI Mapper configuration if it happens to be the same between all of them. Also, in the Windows 3.1 timeframe there were different MIDI channel standards and the one set up by default by most sound card driver installations was not always the one that really makes sense to use with a GM MIDI device. There were some which used channel 16 as percussion vs. the more standard channel 10, some which only gave you channels 1-10, others which only gave you channels 11-16, etc.

I've actually never used Windows 3.1. I was using Win98SE with a Music Quest MPU clone card, and also my Win10 PC with a Roland USB to MIDI cable (UM-One mk2).

auron wrote on 2021-05-22, 05:55:

won't setting the unit to GM mode (power on while holding right PART button) do the trick?

You know what, that did it! Everything got selected properly once I did that combination and it sounded perfect. Guess that's all it needed for both of these MIDIs. I'm still learning these button combinations across a few synths and I have to say, finding them is a chore as I can't find any good references for them. The manual for the SC-55 and mkII do mention some of them at least.

Agent of the BSoD wrote on 2021-05-22, 06:43:

You know what, that did it! Everything got selected properly once I did that combination and it sounded perfect. Guess that's all it needed for both of these MIDIs. I'm still learning these button combinations across a few synths and I have to say, finding them is a chore as I can't find any good references for them. The manual for the SC-55 and mkII do mention some of them at least.

That's funny. On my mkii I don't have to specifically turn on GM mode for it to work. It plays Canyon just fine in the factory default.

Lazer42 wrote on 2021-05-22, 14:50:

That's funny. On my mkii I don't have to specifically turn on GM mode for it to work. It plays Canyon just fine in the factory default.

I just double checked the attached MIDIs above. The first one does indeed say No Instrument in normal operation but is fine in GM mode. Have you run the first attached one through your module? And what media player or sequencer are you using for playback? Curious as to why it would work on one module and not another. In Win98 I just use Windows Media Player (updated one that came with IE 5.5), and in Win10 I use Sekaiju.

Agent of the BSoD wrote on 2021-05-22, 17:43:

Lazer42 wrote on 2021-05-22, 14:50:

That's funny. On my mkii I don't have to specifically turn on GM mode for it to work. It plays Canyon just fine in the factory default.

I just double checked the attached MIDIs above. The first one does indeed say No Instrument in normal operation but is fine in GM mode. Have you run the first attached one through your module? And what media player or sequencer are you using for playback? Curious as to why it would work on one module and not another. In Win98 I just use Windows Media Player (updated one that came with IE 5.5), and in Win10 I use Sekaiju.

Well I am just talking about the file that came with my Win3.1 installation. I didn't download this one in particular and check.

Lazer42 wrote on 2021-05-22, 19:48:

Well I am just talking about the file that came with my Win3.1 installation. I didn't download this one in particular and check.

Hmm, maybe that version of the file isn't the same as the one with Win98SE so it doesn't have this same issue. Still, good to know that this was just an issue with these particular versions of these songs and not with my unit.

Agent of the BSoD wrote on 2021-05-23, 00:29:

Lazer42 wrote on 2021-05-22, 19:48:

Well I am just talking about the file that came with my Win3.1 installation. I didn't download this one in particular and check.

Hmm, maybe that version of the file isn't the same as the one with Win98SE so it doesn't have this same issue. Still, good to know that this was just an issue with these particular versions of these songs and not with my unit.

Yeah, the version of canyon.mid that came with Win3.1 is the 2nd one (also attached above) that has the dual channel layout (Base+ Extended) and thus has the drum on channel 16 problem 😀
Also it's worth mentioning that Windows Media Player recognizes such files by detecting 'Sequencer specific' meta event with a value of '000041h' and plays only the Extended channels (1-10) so no drum channel problem in case of WMP.

Falcosoft wrote on 2021-05-23, 08:36:

Agent of the BSoD wrote on 2021-05-23, 00:29:

Lazer42 wrote on 2021-05-22, 19:48:

Well I am just talking about the file that came with my Win3.1 installation. I didn't download this one in particular and check.

Hmm, maybe that version of the file isn't the same as the one with Win98SE so it doesn't have this same issue. Still, good to know that this was just an issue with these particular versions of these songs and not with my unit.

Yeah, the version of canyon.mid that came with Win3.1 is the 2nd one (also attached above) that has the dual channel layout (Base+ Extended) and thus has the drum on channel 16 problem 😀
Also it's worth mentioning that Windows Media Player recognizes such files by detecting 'Sequencer specific' meta event with a value of '000041h' and plays only the Extended channels (1-10) so no drum channel problem in case of WMP.

Does this include WMP back to 3.1, 95, or 98?

Lazer42 wrote on 2021-05-23, 20:12:

Falcosoft wrote on 2021-05-23, 08:36:

Agent of the BSoD wrote on 2021-05-23, 00:29:

Hmm, maybe that version of the file isn't the same as the one with Win98SE so it doesn't have this same issue. Still, good to know that this was just an issue with these particular versions of these songs and not with my unit.

Yeah, the version of canyon.mid that came with Win3.1 is the 2nd one (also attached above) that has the dual channel layout (Base+ Extended) and thus has the drum on channel 16 problem 😀
Also it's worth mentioning that Windows Media Player recognizes such files by detecting 'Sequencer specific' meta event with a value of '000041h' and plays only the Extended channels (1-10) so no drum channel problem in case of WMP.

Does this include WMP back to 3.1, 95, or 98?

In case of Win 3.1 the situation is different: WMP in Win3.1 respects the channel layout settings of Midi mapper. That is it plays the Extended part (ch 1-10) when Extended setup is selected in Midi mapper, Base (ch 13-16) when Base is selected, and plays all channels when Full is selected. So drums on channel 16 can cause problems in case of GM devices when Full setup is selected.
I have not tested Win95 but from Win98 up to Win10 WMP always plays only the Extended part (ch 1-10). I suspect it's because in later Windows versions the dumber Midi mapper is always in 'Full setup' mode by default.
If you remove the above mentioned 'Sequencer specific' meta event from these files all channels are played on all versions of WMP.