Ruby MIDI switch/glitch oddity

[HughBanton]

Don't know if anyone can shed any light on this ... I've devised a schematic to illustrate a strange finding :

The module in my schematic is a Ruby MIDI switcher; if you hold down a chord and click 'MIDI On/Off' - true/false - you should get a seamless switching of the MIDI stream; all the notes in the chord you are holding should start and stop, just as though you played them that way. It achieves this by storing the current notes in the '@vel' table, (which you can 'watch' at the bottom) and restores them to be playing again when the MIDI switch comes back on.

However ....

Now change the 'MIDI 2 String Monitor' switch at the top to 'true'. This merely introduces a string into the MIDI line, in order to monitor its status, like wot it sez in the user manual.

On my setup, at least, I now find that the MIDI switch becomes completely glitchy, with clicks and audio disturbances every time it's switched on or off.

Is everyone getting that, or is it just me?? If I turn off the MIDI string monitor connection everything's fine again.

Does it mess with the envelope stages or something? Whaaa? It's baffling me.

I realise this looks like a pretty trivial problem (I don't really need to monitor the MIDI anyway) but I've been getting a related issue in a much more elaborate setup, and I think this may well at the route of it. Every little bit helps ...

H

[tulamide]

I suspected a back trigger issue, before downloading your example. Now, on my setup (v3.0.6) it is not completely glitchy, but I do hear a very noticable drop noise everytime I switch from true to false (and only then). I could solve it by preventing a back trigger to mess around, by connecting a trigger blocker with the selector's output and the blocker's output with the (midi) string prim (instead of directly connecting selector and string).

Please try that, it may solve your issue as well.

[Spogg]

Yes I got the same glitch (also using 3.06).

Tulamide is correct (of course!). In my case I connected a green sample and hold prim on the selector output and triggered it from the bool which operates the selector. No more glitches.

BTW I'm envious of your Ruby skills!

Cheers

Spogg

[HughBanton]

Thanks guys, but try as I might I can't get rid of them glitches! The methods you describe all seem to disable the MIDI-string-monitor function; it's certainly true that as long as it's not monitoring MIDI then you don't get glitches.

I'm guessing the provision of the MIDI to String 'conversion' is just a handy one-off convenience facility anyway. It only seems to work if it's directy connected to MIDI, or via the selector. Anything else in between seems to disable it and instead it states things like [#<Midi:0x4f38758>, 1, 40, 20]. (Dunno quite what that means ...)

Well, as I said I don't actually need a MIDI monitor, it was just that connecting it up in this way seemed to cause exactly the same kind of glitches I've been experiencing elsewhere, so I wondered if there were clues to be had. Maybe not.

I've been working with complicated setups with several of my Ruby MIDI switches feeding several individual sound generators - Organ Stuff ( - there's a surprise!! Not..) where I want to be able to turn the generators on & off at will. It all works great .. but I get these clicks. The more complicated my schemes the worse they seem to be - most frustrating when everything else works so well!

I know there are other methods to turn sounds on & off, but switching the MIDI is so much more natural because it's completely identical to taking your hand off the keyboard.

Hey ho, I'll find a solution. Onward!

Hugh

[HughBanton]

BTW I'm envious of your Ruby skills!

Oooh - nice of you to say so but as you'll see I'm afraid my skills don't go much beyond if..then..else and 'for..next' loops!

My so-called C++ attempts are very similar. Endless trial and error, and error, and yet more error ..

Mind you it's quite amazing what you can achieve with just if..then..else and for/next loops

H

[RJHollins]

Hi HughB

Was just testing the first schematic you posted.

I'm NOT getting any 'glitches'. I'm on FS 3.04 [if that matters]

The display readout is showing the [Note, Vel, Ch] as per MIDI spec.

The only change I made to the schematic was switching to the DS OUT module, rather than ASIO out.

[HughBanton]

Ah, that's interesting. Thanks RJ.

If I switch over to DS it greatly improves it, but doesn't cure it completely. So I went back to ASIO, and discover that it's affected by the ASIO settings, certainly much worse at the high-performance end of the spectrum. Which is where I need to be, hmm.

Curiouser and curiouser ... !

[tulamide]

Thanks guys, but try as I might I can't get rid of them glitches! The methods you describe all seem to disable the MIDI-string-monitor function; it's certainly true that as long as it's not monitoring MIDI then you don't get glitches.

If that's what you experience, your setup is completely messed up. What I described does not disable the string-monitor-function at all. It just prevents a back trigger.
The same is true for Spogg. The sample and hold also prevents a back trigger, but not at all the functionality.

[Spogg]

Hmmm…

I’ve found that when I use strings for monitoring that if the string is long, or there are a lot of them being updated, that it seems to cause a slowdown in the FS edit environment during the update.

In your real schematic are there lots of such monitors? Certainly in the uploaded example it gets fixed by stopping back-triggers.

Cheers

Spogg

[HughBanton]

In your real schematic are there lots of such monitors? Certainly in the uploaded example it gets fixed by stopping back-triggers.

No, none at all. I was just an accidental discovery that the 'monitor' arrangement produced similar sounding clicks and I wondered if there might be a common cause.

They sound to me very like the audio clicks you can get when changing a preset, and I seem to recall a conversation on the forum wondering if that might be caused by some sort of interruption to the contents of the poly stream? Which I guess is a mechanism hidden to us. As far as I can see there is no disturbance to the envelope stages 0-1-3-4, it literally sounds like the sound stops for a couple of milliseconds and then resumes. Looks like that on my 'scope. (Analogue man, moi).

Can I be a pest and ask that one of you puts up the schematic for your back-trigger fix? Since I don't seem to be able to replicate it I evidently must have messed up as charged!

Ta

H