I think I get your setup. You really don't want to be using Merge in->out. That's really only for people without the luxury of a patchbay. Since you do have one, use it to route cc messages from the XTk to *both* the XL7 and the Pulse. You'll get the same behavior as Merge I>O without the MIDI loop.
That said, the assumption in the second paragraph is what's not right. Per the MIDI spec, EVERY time you get a start message you are supposed to return to the beginning, even if you're already playing. By looping the start message over and over, you're essentially telling the XL7 to restart the sequence every millisecond (yipes!). As you've no doubt experienced, the XL7 doesn't like that. Amusingly, if you let it go till it processes every message that came in, it'd probably recover. That might, however, take quite a while.
Cheers,
-Aaron
Zsolt Szabo <Zsolt.Szabo@...> wrote:
Hi Aaron,
I wasn't creating a MIDI loop. At least not intentionally.
I took great care that every kind of MIDI message (except
realtime) is properly routed and doesn't get back to XL7.
My MIDI interface/patchbay is an Emagic AMT-8.
I had to merge the MIDI In to Out because in the mentioned
case I wanted to tweak parameters on a Waldorf Pulse (CC50
as Cutoff) while recording movements in XL7, using a
Waldorf XTk as masterkeyboard (CC assignments same as on
the Pulse where applicable). Since CC33-63 is nonexistent
on the XL7 this was the only solution.
Anyway, since I can overcome the problem is no big deal
but I assumed realtime messages are handled correctly (for
example, after an initial MIDI start following start msgs
are discared until a stop message received/transmitted).
Regards:
Zsolt
> Von: Aaron Eppolito
> An: xl7@yahoogroups.com
> Betreff: Re: [xl7] (not) a nasty bug
> Datum: Wed, 21 Dec 2005 20:24:47 -0800 (PST)
>
> So wait, creating an intentional MIDI loop is a bug in the XL-7?
> Hardly! If you create a MIDI loop, every single product on the market
> will do MIDI loop things. If there is a bug (other than in your setup)
> it's the fact that MOTU (I'm presuming that's what you have) patchbays
> don't route realtime messages, they multicast them. It's intentional
> on their part, but it's really annoying.
>
> Anyway, the solution is don't merge in to out, and filter reception of
> start/stop on devices you don't want to recieve it.
>
> Happy holidays,
> -Aaron
>
> --- Zsolt Szabo wrote:
>
> > Hi !
> >
> > XL7, OS 2.0.
> > Connect both in/out to a MIDI interface or patchbay which doesn't
> > have
> > the MIDI realtime msgs filtered. Enable "Merge MIDI In to Out" on the
> > XL7, it doesn't matter which out port, A or B which is even more
> > weird
> > than what happens next. Start XL7 by pressing PLAY. Now XL7 will
> > nicely
> > transmit MIDI Start message. If the MIDI interface transmits it back
> > to
> > the XL7's MIDI In port a nasty MIDI loop will occur which will kill
> > the
> > XL7 in approx. 10 seconds. Complete freeze. If you are quick enough
> > to
> > pull the cable the XL7 is able to cope with it (buffer size big
> > enough
> > to hold approx. 8-10 seconds worth of MIDI msgs ?)
> >
> > 100% percent reproduceable.
> > Workaround - filter the outgoing MIDI realtime on the MIDI
> > interface/patchbay towards the XL7's MIDI In. Keep an eye on
> > it in live situations - I learnt it the hard way ...
> >
> > Regards:
> > Zsolt
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
[Non-text portions of this message have been removed]Message
Re: [xl7] (not) a nasty bug
2005-12-22 by Aaron Eppolito
Attachments
- No local attachments were found for this message.