![]() |
|
#1
|
|||
|
|||
|
Hi Jeff,
I wrote about this problem a long time ago on the yahoo board, but was traveling and the problem was intermittent so I was not able to catch it using dbwin until recently. ANy time I get a few layers of overdub going on mobius as a vst in a pretty complex and resource intensive bidule layout, mobius cuts out and switches to bypass. Sometimes I can just reset bypass and the loops will return but usually it just keeps resetting itself. The problem happens if I have several layers of overdub on one track or several single layer tracks going. Is this the problem described in http://www.circularlabs.com/forums/showthread.php?t=105 ? The DBWin log output for one of the crashes is here: http://way.net/tmp/~DBWIN2-mobius.log If it is the loop flettening problem, could you let me know? Or is it something else. The layout I use uses a lot of resources, so my guess is that I am just running out of memory (I have 3 gigs plus virtual) but I am not sure. The relevant lines seem to be:
|
|
#2
|
|||
|
|||
|
Wow, I could have written this exact posting myself. This problem is THE thing keeping me from taking Mobius/Bidule on my laptop out to gig. I've been too busy with other musical adventures to begin a dogged pursuit of a solution, so it is a very welcome sight to see someone else taking the time to document it.
I say it is the same prob;em, but the truth is I have not isolated it to being about number of layers - I can really only say that I have the same symptom of Mobius switching to bypass at some point during complex, operation-intensive loop creation. Same setup, Mobius as a VST within a complex and resource-intensive Bidule layout. Thank you, rcrath, for your post. I look forward to some enlightenment from Jeff. Hal |
|
#3
|
|||
|
|||
|
Same issue here!
because I often rerecord The output from mobius through fx and back to mobius, there is an extreme feedback when mobius goes to bypass. really catastrophic, if this happens on stage... unfortunately I couldnīt figure out, what conditions are leading to this behavior in my case. But I can say that this happened to me, even with relative simple looping. I was just playing to a loop for a half an hour and suddenly there was this screaming feedback (rerecord routing was enabled, but no record) and mobuis was bypassed. To me, this is a rare event, which is not a good thing as itīs making it very hard to isolate. So Iīm happy, rcratch has managed to take a log! Last edited by Piesi; 02-12-2010 at 10:37 AM. Reason: tried to make my post more accurate |
|
#4
|
||||
|
||||
|
Hard to say what this is, unless you've got 16 or more layers this
probably isn't the SUSSubstitute bug that we were discussing in a different post. People regularly run all 8 tracks with 10 or so layers with no problems. It is more likely something in the way the tracks were made, how they're synchronized, etc. As you say, it could be running out of memory, but unless your loops are minutes long that seems unlikely. When Bidule puts Mobius into bypass it probably means Mobius crashed. Mobius will never put itself into bypass but some hosts will catch crashes and disable the plugin. The "Recorder::interrupt reentry!" message means that Mobius is doing something that is taking a really long time (like maybe crashing) and it isn't keeping up with requests from Bidule to process the audio stream. The interesting part of the trace is what led up to this, so if you still have the file or you can reproduce it I'd like to see more of this. Sorry I can't shed more light on this, it really could be anything. If it's hard to reproduce, I'd recommend just leaving debug trace on all the time whenever you practice, then if it crashes zip up the entire thing and send it to me at jeff@circularlabs.com. |
|
#5
|
|||
|
|||
|
Hi Jeff,
Thanks for your response. The lines prior to the ones I posted should be in the full log above, which was for the entire session. The link again is http://way.net/tmp/~DBWIN2-mobius.log. I just posted a few lines of it in the post, but the rest is in the linked file. I'll keep the dbwin on for a while and will send any more logs directly to you. Thanks for looking into this. Rich |
|
#6
|
|||
|
|||
|
Quote:
Thought maybe this info could help narrow things a bit more. ~Rich |
|
#7
|
|||
|
|||
|
Hi Jeff, was the log linked to above enough or would you like more of them or any other information?
Best, Rich |
|
#8
|
|||
|
|||
|
My workaround for this problem is hosting Mobius in a Kore2-instance within Bidule. Havenīt had a crash since then:-) If you donīt have Kore2, maybe another VstPluginHost-Plugin will do the job as well. There are even free ones, you likely find them at KVR.
|
|
#9
|
||||
|
||||
|
Sorry for the delay on this guys, I'm just really swamped at work right now.
I'll try to look at the linked log this week. |
|
#10
|
|||
|
|||
|
Hey, I've just checked back and found that some discussion is underway about this. I'm enmeshed in several projects right now that won't allow me to spend serious time on this until after 3/21, at which time if it is useful I can attempt to duplicate my episodes and log them, and post the results here. For now I can offer this:
I use 6 tracks, 4 loops, as a basic template. Presets have all tracks set to Group 1, then after initial material is recorded I switch individual tracks to Group 2 for altering via 1/2-speed, rev, divide, etc. Usually before I hit NextLoop I have everything back in Group 1. I use a Peavy 1600X, FCB1010, a gamepad, and a another expression pedal to send commands to Bidule, VST instruments and effects, and Mobius. Without a doubt one way to crash is to send too many cc messages too fast, but I don't think that has ever resulted in the bypass phenomenon - that is a much cruder and nasty, screeching crash. I've learned the appropriate restraint and this hasn't happened for many months. The bypass problem happens when I have a fair amount of material spinning, sometimes deep into the number of layers on some tracks. As I noted in my previous post, I have not been scientific about this as Rich has been. One last thing- every once in a while when this happens, I can successfully change the status of Mobius within Bidule; usually, though, any such change lasts a split second before it returns to bypass. I can't tell you how happy it would make me for this to go away. I've invested a LOT of time in setting up Bidule and Mobius to work together, and other than this, it is like walking into a dream to work with these tools. Jeff, whether this works out or not, bless you for Mobius. |
![]() |
| Thread Tools | |
| Display Modes | |
|
|