Ruby MIDI switch/glitch oddity

[tulamide]

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.

Since your experience is so totally different from mine, I almost think that it is due to your ASIO buffer being too small.

About Midi: When you connect a red Midi line with a green string, a conversion takes place and a back trigger causes a dropout, whenever you switch the selector. I don't know why this is not the case for you, but I thought, a system that takes away the string conversion from red and leaves it to Ruby could help.

This also does work perfectly fine for me. I'm eager to hearing what you experience!

Oh, and test this schematic, BEFORE increasing your ASIO buffer, please.

[Spogg]

First thing is for me to apologise!
For my solution, with the sample and hold, the glitch disappears but so does the monitoring function. I was too quick and didn’t check the function.

I’ve experimented more and found that the string prim behaviour is not “normal” with a MIDI input. Have a look at the schematic.

capture.png (211.28 KiB) Viewed 15103 times

The MIDI message displayed in the first string prim can never be extracted from it, even with triggering.
Below that is normal string prim behaviour where a string in the first one is also displayed in the second one.

I shall conclude that this MIDI monitoring with a string prim is a “special feature” to offer a quick aid and the prim’s functionality is modified with this connection in place.
It’s been interesting to find this out.

Tulamide's Ruby solution is the best method.

Cheers

Spogg

[RJHollins]

instead of a STR prim ... have you tried a TEXT box ?

[HughBanton]

Tula's Ruby monitor is superior to the FS string method without a doubt. I've already made a cutdown version and put it in my toolbox for future use, attached.

Thanks for that. I've contemplated making a 'scrolling' MIDI monitor for a while, this will definitely be a good starting point for when I (or someone!) gets round to it.

Well, the news is that for me it functions without any glitches. Substitute a String prim and they're back, so certainly looks like you've hit the proverbial nail. However changing my ASIO buffer size has no noticable effect, so nothing very conclusive there.

Just for the record, I'm currently working with my HB3 Organ Generator in the attached form. Since early versions of HB3, (which I think is on Spogg's Flowstoners site) I've split the entire graphical programming section off into a separate entity; it saves the 'voicing' in a '.vcg file' that I've devised (loads much faster than a regular preset file) and an organ, when I finally get that far, will just run pre-loaded Generators, as per the attached. They're a bit of a Tardis - massive on the inside - and if it looks over-complicated bare in mind it can accurately do anything from a Contra Trombone 32' to a 1' Penny Whistle!

It is these that I've had trouble reliably switching, without the audio glitches, nothing actually to do with MIDI monitors ..

Thanks for your interest!

H

[Spogg]

I like your organ Hugh!

Cheers

Spogg

[tulamide]

Spogg: Interesting finding. Never realized it!

Hugh: Is the HB3 where you have those issues? I ask, because it runs just fine here. (Wanted to learn from the modules, but it's all Assembler )

[HughBanton]

I like your organ Hugh!

Flattery will get you everywhere.

Hugh: Is the HB3 where you have those issues? I ask, because it runs just fine here. (Wanted to learn from the modules, but it's all Assembler )

Sure, HB3 runs just fine in single instances, it has done in most of its incarnations. But a proper organ will need to run umpteen (I'm not even sure yet how many a regular PC can handle) .. and it's when I multiply it all up that the switching starts to get glitchy.

Speed considerations are also why I've been converting everything I can into Assem. Don't for one minute imagine I know how to write this stuff (!!), but I've had some success over the months doing the editing that's described on Flowstone Guru, I've kind of got the hang of it after a while.

Happy to answer any questions and explain any of the insane workings inside HB3!

Hugh

[tulamide]

Here's the MIDI monitor again, but optimized for both, speed and code readability.

[HughBanton]

.... and following your lead here's a streamlined version of my MIDI switch. Didn't realise I could lose the input array.

I have a niggling doubt about the 0..71 loop - how fast does Ruby run this? Could this in fact be the bottleneck at the route of my problem? (And is there a better way, maybe some kind of note list ??)

H

[Spogg]

I have a niggling doubt about the 0..71 loop - how fast does Ruby run this?

I’m just beginning with Ruby so it might be best to ignore this

I made a basic loop that counted from 0 to 100,000 in steps of 1. It completed in about 3 seconds, which means each iteration took around 33uS. This surprised me because I was thinking of it like DSP which is based on sample rate timing, and I took the 100Hz thing to mean the whole code would be “clocked” every 10mS. Obviously this isn’t the case; it’s only the elapsed time that seems to do this.
In addition when I doubled the loop size to 200,000 the duration increased from 3 to 11 seconds so I didn’t see a proportionality. It suggests that shorter loops may run even faster, but hey, what do I know? Very little so far.

Cheers

Spogg