stevecowley

I know it doesn't make any sense for it to be related to Diva, but since updating it, Logic has been able to save again. I haven't tried re-installing the previous version to see whether the problem reoccurs. Just for the sake of keeping this topic updated, I hadn't tried a different OS X account, or saving to another location. The behaviour was that if I pressed Command-S or selected Save from the File menu, nothing happened at all. If I then Saved As, Logic would crash immediately after saving.

I recently installed Logic 10.1.1 on a new MacBook Pro, and also all of my 3rd party plug-ins, including Diva. I know that there is a known issue with the GUI of u-he's Diva not opening in 10.1. Does anybody know if this also affects Logic's ability to save? Before moving to the latest beta of Diva (which fixes the GUI issue), Logic wouldn't save - but it would Save As. Since installing the latest beta of Diva, Logic seems to be saving ok, but I'm not 100% confident that the 2 things are related. Has anybody had the same experience? Is there anything else that is known to cause Logic to have problems saving?

The subject line says it all really, but here's a bit more detail: I am setting up a MainStage session for a band. They want to use it for keyboard sounds and backing track playback. I have used MainStage many times before for keyboard sounds, but never for playback. I have exported CAF files from Logic for the different tracks of each song and have them all playing back as expected, with the CPU meter reading nice and low and never getting close to peaking, even with lots of notes being played on both keyboards. Unfortunately, during the first rehearsal day, playback stopped without warning - twice in 6 hours. I am using a Novation Launch Control to remotely Play and Stop as well as Previous/Next Patch and Previous/Next Marker. Also 2 Novation Impulse 61 controller keyboards, connected via a MOTU microlite MIDI interface. I have checked the routing of all controllers from both keyboards and the Launch Control. In MainStage, I have deselected "Send all unassigned MIDI to Channel Strips" and nothing from either keyboard is affecting playback. I plan to try: - Monitoring the MIDI output of the Novation Launch Control. It is possible that the Launch Control was sending a Stop message without that pad being pressed. Seems unlikely but it's worth looking into. - Trying WAV files instead of CAF files, and also turning Sync off within the Playback plug-ins. Now that tempos have been decided on, I don't need to be able to sync the playback to the tempo of the song - I can just let them all play at their natural speed. And I can live without having the markers. Just being able to play a song from the beginning to the end without stopping is all I really want! If anyone has any bright ideas of other things I can try, or any experience that may be helpful, then I would flippin LOVE to hear from you! Cheers, Steve Mainstage 3.0.2 Macbook Pro running OS 10.9.1 MOTU microlite MIDI interface Focusrite Scarlett 18i20 USB audio interface

I have the same issue here with 3.0.2 - and the worst thing is, even though MainStage still seems to be saving all assets as part of the file, when I load this file into MainStage on a different computer, it can't find any of the playback files or samples, even though they appear to be embedded!

Thanks Michael. I solved the sustain pedal issue - it was creating a MIDI loop because Mainstage was set to pass any unassigned MIDI input to ALL channels. The sustain pedal had not been assigned to a specific instrument, so it was being sent to all instruments.

Hi, Re. Plogue Bidule - I ran it as a standalone program, hosting Kontakt as a plug-in. It's been a few weeks since I did this, but from memory, I used the IAC bus to carry the MIDI from Mainstage to Bidule. Then set up Mainstage audio channels with the Bidule outputs set as inputs, to route the audio from Bidule back into Mainstage. Unfortunately this idea ran aground because everything I did resulted in latency being introduced. The idea I mentioned above, to set up a dummy channel strip to absorb the MIDI messages that were coming back in on the IAC bus, didn't work. Then I found out about going to the Assignments and Mappings tab, selecting the "Unassigned" assignment at the bottom of the list, and unchecking "Send unassigned MIDI to all channel strips". This negated the need for a dummy channel strip, but I still had the latency problem after running for around 30 minutes. It seems that either: - There is a fundamental problem with the IAC bus (unlikely as I haven't found anyone else with the problem) - There is a problem with Mainstage that makes the IAC bus unusable (possible) - There is a problem with my specific system that makes this not work (hopefully unlikely as I am using a brand new MacBook Pro with all the latest software and drivers) Having said all of the above, I'm not even sure that using the IAC driver is the solution to my problems, as I *think* I've had the CPU spikes even when using only stock Logic/Mainstage plug-ins. I need to do more research on this. Please keep us updated with how your IAC bus adventures go! Steve

Ah ok... perhaps I should give that a go. I can always revert to a backup if it screws something else up! Cheers, Steve

Thanks Maddcow and mmm42 for getting involved. Maddcow, I would be very interested to know whether you find anything in your Console log that correlates with your overloads. I have run several tests at the request of someone at Apple, both with and without my audio interface connected. Still getting spikes that correlate with the same "HI_WAT_ALERT" collection of messages in the log. What I did to test it: Using an external computer, I set up a Logic session to send every note at every velocity into every MIDI input for every song, with program changes to select each song in MainStage. In a set of 15 or 16 songs, I'm finding that I get a CPU overload roughly once every 2 runs through the set. Next, I'm going to create some EXS24 instruments that take up the same amount of memory as the Kontakt instruments. My next test will be to remove all traces of Kontakt from the MainStage concert, replace them with the new EXS instruments and run the same MIDI test from Logic. I won't be able to do this until Saturday, but I'll let you know how I get on. mmm42, thanks for the heads up, but I'm not going to update to 10.8.4 at this point. This MainStage concert is critical to a large touring production and although the CPU spikes are annoying, I'm able to flip to a back-up rig as soon as they happen. This is a real pain, and does mean that I have to constantly have my finger on the A/B selector button, but at least I know that's the ONLY problem I'm having. So I'm sticking with 10.8.2 for now as the release notes for 10.8.4 don't mention anything about audio drop outs, and for 10.8.3 they say it fixes "an issue that may cause audio to stutter on 2011 iMacs", which isn't the problem I'm experiencing. If you know more info than is mentioned in those release notes, please let me know! Steve

Exactly - whatever is sent on the IAC bus appears to be coming back into Mainstage. And of course Mainstage assumes that any incoming MIDI that you haven't routed anywhere should be sent EVERYWHERE! This is why I created a dummy keyboard in Layout mode - to absorb this incoming MIDI on the IAC bus. Unfortunately this has the effect over time of introducing a massive amount of latency - I'm not sure why. To be honest, I think I'm about ready to give up on the IAC bus idea, as I need it to work in one direction only, or for Mainstage to be able to ignore anything coming in on the IAC bus. I've tried setting up a session using Plogue Bidule as a place to host instruments outside of Mainstage, and this seems to be working. For now, I'd recommend this as a workaround.

I'll keep the IAC Driver discussion to the other thread. CCT, thanks for suggesting that I have a look in Console for the cause of the CPU spike. Sure enough, something did happen at the exact time of the last CPU spike: 02/06/2013 21:15:55.000 kernel[0]: (default pager): [KERNEL]: ps_allocate_cluster - send HI_WAT_ALERT 02/06/2013 21:15:55.000 kernel[0]: macx_swapon SUCCESS 02/06/2013 21:15:59.000 kernel[0]: (default pager): [KERNEL]: ps_select_segment - send HI_WAT_ALERT 02/06/2013 21:15:59.000 kernel[0]: macx_swapon SUCCESS There doesn't seem to be a "problem" as such - there are no error messages - but clearly whatever is happening here is causing Mainstage's CPU meter to peak. I have been monitoring the CPU usage window in Activity Monitor during the gig, and none of the 4 cores are peaking there. I'm going to do some research into what this HI_WAT_ALERT message means, and whether it's a process that can be disabled without doing any damage. If anyone has any hints, please chime in! Steve

Thanks guys, CCT, you may be on to something there. I'm on tour at the moment so I don't really have time to sit and play with a test concert, as the spikes only seem to happen very intermittently - perhaps once every 2 or 3 days. However I do know when the last one happened, so as soon as I get to the venue tomorrow I'll have a look at the Console log and see if it shows anything. If there ends up being no problem running Kontakt inside Mainstage, and I can solve the CPU spike issue, I'll be very happy... Steve

Hi Jordito, many thanks for your reply. I am running a MainStage session with 3 MIDI keyboards, each routed to inputs of my MOTU micro lite MIDI interface. In the layout page, there is a keyboard representing each of these, and all have the relevant hardware settings. As far as I know, all MIDI inputs are set correctly, but I'm now going to set up a brand new concert, very simple, to see what happens: First of all, Jack is running to provide 32 channels of audio. Jack is not set up to handle MIDI. I plan to route MIDI from Mainstage to an external instance of Kontakt, on channel 1 of IAC Driver bus 1. The audio will then return to MainStage via Jack channels 31-32. MainStage preferences: Audio Output on MOTU Ultralite mk3 Hybrid Audio Input on JackRouter Layout mode - set up one keyboard, with the MIDI Out of the keyboard plugged into Port 1 of the micro lite. Keyboard 1 Hardware Input: Device: Port 1 micro lite Channel: Channel 1 Sustain pedal 1 Hardware Input: Device: Port 1 micro lite Channel: Channel 1 Type: Press/release Number: 64 (sustain) MIDI Thru: Automatic Send Value to: None Edit mode - set up one external instrument In the New Channel Strip dialog box: 1 External Instrument MIDI Input: Port 1 micro lite MIDI Output: IAC Bus 1 MIDI Channel: 1 Format: Stereo Input: 31-32 (these are Jack channels) Output: 1-2 Play notes on the keyboard - fine. Press sustain pedal - MIDI loop detected. So there are no keyboards in Layout apart from the one I have created, which is set to receive on micro lite Port 1 only. There are no channel strips within the concert other than the external instrument I just created, which should only be receiving MIDI from the keyboard on Port 1. Clearly MainStage is responding to MIDI coming back through the IAC bus. At this point I don't even have Kontakt open, but if I do have it open and set to receive MIDI on IAC Bus 1, and set to not pass any MIDI through, the result is the same. Have I missed something obvious?! Steve

Hi guys, I have kind of solved this issue, but it has created another problem! I've detailed it in a new post as it's rather a long explanation - see viewtopic.php?f=17&t=96042 Hope we get this one sorted eventually...

If you don't want any background to my issue, please skip the next 5 paragraphs and start with the heading "IAC DRIVER PROBLEM" BACKGROUND: I have been running a MainStage concert which has quite a few instances of Kontakt in it. I am using Alicia's Keys in most songs, and around 15 other Kontakt patches in various songs. The only other third-party plug-in I'm using is Massive - everything else is stock MainStage plug-ins. I have experienced problems with sudden CPU spikes - the CPU meter within MainStage stays below 50% the whole time, apart from seemingly random spikes. I have monitored the CPU meter in Activity Monitor, and when MainStage tells me there's a spike, all 4 processors are still each running at under 25%. No spike is visible in Activity Monitor. This spike is not repeatable, and happens maybe once every 2 or 3 shows. Unfortunately that's too often as I am working for a major label artist who is playing to a lot of people every night, who I really don't want to hear a nasty crackle from MainStage. It doesn't necessarily happen when there's a lot going on - it can be during a period where there's only piano and nothing else putting pressure on the CPU. I have done plenty of gigs without using Kontakt or Massive, and haven't experienced these sudden spikes. Maybe I've been lucky. But I'm working on the assumption that there is something about using Kontakt and/or Massive that doesn't agree with MainStage. I have searched for a solution to this, but I haven't found anything. I'm using the latest versions of all the software, so if anyone has any suggestions that don't involve using third-party plug-ins outside of MainStage, then I'm all ears! IAC DRIVER PROBLEM: I've set up Kontakt as a standalone instrument, receiving MIDI from MainStage via the IAC Driver. This works as expected, and Kontakt passes audio back into MainStage via Jack (or Soundflower, both seem to work). I have set Kontakt to not pass MIDI through - however MainStage still seems to be receiving MIDI via the IAC Driver as I get a message that alerts me to a MIDI loop. This can be worked around by creating a new dummy keyboard in MainStage's Layout page, and assigning the IAC Driver as the input MIDI channel. I then create a dummy/empty instrument channel within each song, with that dummy keyboard as the MIDI controller for the dummy instrument, and that extra MIDI that's coming in is no longer a problem. Everything is working fine, with one MAJOR problem - after the system has been running for 10 or 15 minutes, the latency of Kontakt instruments has noticeably increased. After an hour, the Kontakt instruments are unplayable, with the latency lasting about half a second. It's definitely the MIDI side of things, rather than audio, as I can see the meters in Kontakt showing the audio being triggered late. So this seems to be either an issue with Kontakt dealing with the MIDI via IAC Driver, or an issue with MainStage receiving a load of MIDI info via the IAC bus which it shouldn't be getting, and getting "clogged up" with it. Can anyone confirm this behaviour on their system? Has anyone successfully used the IAC Driver with MainStage (whether with Kontakt or any other standalone instrument)? [Edited title to reflect the way this conversation is going...] PS. Please ignore the gear list in my signature - for this system I'm using the latest MacBook Pros with solid state drives, the most recent OS 10.8, Mainstage 2.2.2, the latest version of Kontakt, MOTU Ultralite and Micro lite interfaces

I'VE FIGURED IT OUT!!!!! Perhaps I should've known this, but the solution is to make sure that there is at least one instrument assigned to each keyboard. If there isn't, then the controls on the keyboard which isn't being used will affect all the other keyboards. So for example: Keyboard 1: piano sound Keyboard 2: lead synth sound Keyboard 3: nothing In the above example, if I press the sustain pedal of keyboard 3, it'll affect keyboards 1 and 2. But with this set-up... Keyboard 1: piano sound Keyboard 2: lead synth sound Keyboard 3: dummy sound (this can be mapped to just one note, or muted) ...everything works correctly. Mainstage needs the dummy instrument there to route the MIDI to, otherwise it'll send it all over the place! Phew. Hope this helps anyone else with a similar issue. Steve