Forum rss-feed


Help: Eigenharp, MainStage and MIDI Loop Feedback error

Most Recent

written by: geert

@dhjdhj, the global one is 'minimal decimation' it will take precedence over the local ones unless the local ones are larger.

written by: dhjdhj

Thu, 10 Mar 2011 19:02:40 +0000 GMT

(This is a copy of a problem I just posted on the Apple MainStage forum but in case someone in Eigenharp land has seen it, I'm posting it here)

I have a very simple concert setup (and I have the IAC device disabled). I am playing an Eigenharp and MIDI in from the Eigenharp is being received both directly by MainStage (on one channel strip) and also by a MaxMSP patch.

MainStage, on another channel strip, is receiving the MIDI output from the MaxMSP patch.

I can play my song 10 times in a row with no problem. Then on the 11th time (say), in the middle of the song, MainStage will just popup the **** MIDI Loop Detected message and everything stops until I press enter at which point it's business as usual.

Because this is completely random, I have no idea whether this will happen the first time or the 10th time. I'm supposed to be performing in public tonight with this setup and I'm just petrified by this issue.

There are no MIDI cycles (of which I'm aware) and I'm actually skeptical whether the actual problem is a MIDI loop or something else.

Anyone have ANY idea how to kill this?

written by: john

Thu, 10 Mar 2011 19:21:05 +0000 GMT

Hi David

That sounds like a data rate issue. Loop detection is usually managed in software by looking at the data transmission rate and the assumption is that MIDI has a pretty low maximum rate. This is true for most setups, but the Eigenharp generates a lot of data so tends to trigger the detection algorithm.

I'm not sure if you can disable this detection in Mainstage, I'll ask Dave to look into it. Alternatively you can do the horrid thing and turn the data decimation up in the new MIDI routing matrix, there's a new parameter in the GUI just for this purpose - this will reduce the performance resolution but might get you out of a hole and depending o what you're doing might not be that noticeable.


written by: mikemilton

Thu, 10 Mar 2011 19:44:36 +0000 GMT

Well, this isn't based on any actual knowledge, but perhaps the loop detection might be triggered by 2 similar messages being received within a time window by code ahead of routing to a channel strip.

So, in addition to (or instead of) throttling can you ensure that the two paths never send the same message, perhaps by putting both streams through MaxMSP which would then ensure this and have an output for each strip?

Just a thought


written by: dhjdhj

Thu, 10 Mar 2011 20:14:36 +0000 GMT

Actually, I already turned the decimation on --- but it wasn't clear whether I got it everywhere --- there's a global setting and I can't figure out whether the local ones override it or it overides the local ones.
I set it to 10 ms, it didn't seem to have any impact on my vibrato but of course I can't tell whether it fixes the problem as it only breaks when it breaks (sigh)

written by: dhjdhj

Thu, 10 Mar 2011 20:15:50 +0000 GMT

Well, I normally play several keyboards in parallel, all fed through MainStage and have never seen an issue. I certainly could reroute everything in principle but I'm not going to have time to do it for the gig tonight, so keeping my fingers crossed.

written by: mikemilton

Thu, 10 Mar 2011 21:20:54 +0000 GMT

Best of luck, and have a great evening

written by: geert

Fri, 11 Mar 2011 07:51:34 +0000 GMT

@dhjdhj, the global one is 'minimal decimation' it will take precedence over the local ones unless the local ones are larger.

Please log in to join the discussions