Midi all input Omni mode with multiple keyboards seems to have some quirks for me

In the course of setting up an experiment for my post about CPU useage on my multi-core system with X1c expanded I ran across another issue which hopefully someone can enlighten me on. 

I have two midi keyboards I could play instruments with and both were enabled in the devices preferences. They were both sending midi output. They were both sending on midi channel 1. The first device (my Kurzweil 2600) was being used to control my CPU useage experiment and was fine. All the VST's were responding to it's keyboard input with no issue. Interestingly, NONE of the VSTs responded to the other device (my Axiom 61 keyboard, also on midi channel one) when I was using the 2600. All VST midi channels had input echo enabled to allow them to all play at once for my experiment.

I paused to look in my X1c preferences for the setting of some init file values. I changed nothing and cancelled out of the preferences screen. My 2600 no longer could control the VSTs even though it was clearly still sending midi messages. The Axiom 61, however, now COULD control everything. Mind you both devices were on midi channel one and all the VSTs were supposed to be paying attention to any midi input on any midi device. Except they weren't...

I am probably going to slap myself in the forehead when someone much wiser gives me the one simple thing I am missing but for right now I am stumped. Any ideas?

When you have both keyboards set to channel 1, most times whichever track is selected is one you'll hear and get input from. Other times, both keyboards will trigger the soft synth. So, it depends on the soft synth most of the time. If you're using two different soft synths, one for each keyboard-controller, the response through Sonar will be varied. Again, depending on the soft synth, the controller, and your setup. There are several variables I can't see from here without knowing your keyboards. Omni mode should allow both controllers to trigger but the situation I described above can over ride. Change one of the keyboards to a midi channel other than 1. I use two straight keyboard controllers, one set to channel 4, one to channel 6. Set up your track inputs to the specific channel the controller is on, not omni. Have a seperate soft synth for each controller or a multi channel soft synth that will accept multiple controller input, ( Fantom VS is one ).
You can also put both controllers through a Midi Router and select the Midi router for your input.

Stone House Studios


Interesting.  Are both devices using the same midi driver?  Is the Axiom USB?

Brian

All devices are using midi connections. I purposely did not want to use USB. The Kurzweil does not have a USB port.

I would think, while you did not change any settings, it still checked your connections. I do not use "all" inputs unless I have to. I know it works in Sonar 8, because I used to have a problem to separate the drummer from keyboard player because I set their both channels to "all". I would suggest to close everything, restart computer and see what happens. There is no reason for it to not being able to respond to both boards. That's the struggle with Sonar, while many users try to deny it and claim that everything works for them all the time.

Crg


I don't know why you would use such a setup to examine CPU scheduling. CPU scheduling is not going to change based on what midi port you use as far as I understand it. CPU scheduling will change based on the amount of information you throw at the CPU in relation to how many tasks it is already handling and whatever scheduling protocals are programmed into the Motherboard-PC. As to the behavior changes, if you have midi out ports enabled to both controllers when ever you access preferences, Sonar is going to send controller messages to the keyboard controllers. Exactly what messages and what the result would be I couldn't say. If you're not driving a keyboard controller with recorded midi from tracks in Sonar, you should disable your midi outs. In the same vein, if you have midi outs enabled, when you play one keyboard it can send controller messages to the other keyboard. Many times different brands of keyboards don't match exactly in the messages they send. When you have all this going on on the same channel, controller messages can clash. It becomes a driver conflict at some point and Sonar ( or any DAW host ) will have trouble processing the mixed controller messages.

I used this simple setup to insure that all of the VSTs are triggered at once from a single controller. Once again, this was not supposed to be a forum topic on whether this was a sensible way to set up midi keyboards because I know it was not. I just assumed that an extremely logical load balanced system would assign each VST a separate execution thread. I do not see evidence of that from my experiment. The experiment was prompted by a REAL setup in which I did NOT have keyboard controllers on the same midi channel and did NOT assign the same controller or midi channel to any of the VSTs.

There are no midi messages being sent back to my keyboard controllers. I don't see any sort of driver conflict going on because only the MOTU driver comes into play for midi routing. As I stated previously, I purposely did NOT use USB nor would that have been an option for my Kurzweil 2600. This whole setup is done purposely LIVE so Sonar has no reason to even try to send control messages to the keyboard controllers.

In creating this forum topic my intention was to just point out an anomaly that seemed not to make sense to me. I did not want it to devolve into a discussion about whether using VSTs set to receive any input on any midi channel is a good thing to do. It clearly is not and I understood that from the start. I still think the setup was fine for demonstrating the CPU scheduling anomaly which was the original intention. 

I do appreciate all attempts to understand this issue and help so hopefully my replies don't come off as being arrogant or rude.

brundlefly


I don't know what you're overlooking, but there must be something. I can't replicate this problem using my RD-300s and QS8 on independent MIDI DIN ports (both on the E-MU 1820m), both sending on channel 1. Both can play any combination of hardware and software instruments simultaneously if I set inputs to All Inputs - Omni, and enable Input Echo on them.

I have a MIDI Express XT that's not in the mix right now. I could test it, but I'm fairly certain there would be no issue there, either.

Thanks for taking the time to check this out. I agree with you that the Midi Express XT is unlikely to make any difference. In any case, unless we had the exact same clockworks routing setup it would not matter.


After all is said and done perhaps I will just ascribe this to "it just is that way...oh well" and move on since as a practical matter I would really never set my system up this way. Just the engineer in me likes to have those annoying loose ends explained. One day I may trip over an explanation but others are probably correct in that it may just be some sort of VST dependent behavior.

brundlefly

In any case, unless we had the exact same clockworks routing setup it would not matter.

If you have a ClockWorks setup that does anything other than treat all the ports as independent I/O (i.e. using one of  the "Sequencer" presets) so that SONAR has full control of the routing, that would be the first place I'd look for a problem.


I'm curious to know what the issue is as well - if only so there isn't another unresolved thread floating around that chalks up an issue to "General SONAR Strangeness" when there's actually a logical explanation to be found.

Yes this is a vanilla setup. 30 fps sequencer. I have other presets for ClockWorks that take advantage of the routing capabilities but I wanted to minimize the variables here.


I had an interesting unrelated issue a little while back when first getting my MOTU MIDI Express XT. It was replacing a Midisport 8x8 that had died. I was getting all sorts of crashes that made no sense in ClockWorks. It turned out that I had not uninstalled my Midisport driver. I guess that driver and the MOTU driver were fighting like two trolls in the basement and my system was the collateral damage :-) Things cleared up nicely after I uninstalled the Midisport drivers. Figured someone might find that interesting. I am definitely liking the MOTU unit a lot these days.

I do appreciate all attempts to understand this issue and help so hopefully my replies don't come off as being arrogant or rude.

 
No problem. I won't flame on you. I'm just running some scenarios by you. If you have your Midi outs enabled and hooked up, messages are being sent. Not only to each keyboard but also to Sonar. Have you looked at your Event List? You'll see note events, controller events, etc. With both controllers set to channel one, there is going to arise, depending on the way and timing of the input of events that controller messages might not be entered in the right order per keyboard. So, if windows enters data at a set point and Sonar records it at that time stamp, and the controller messages don't link correctly to the proper note events, you get weird behavior. I have had this happen on projects where two controllers were sending on channel one. It won't matter if you are using a router or not. I had to go into the Event list and decipher the control events and remove some of them to save the project. With that scenario in mind, you're not going to get an accurate picture of CPU usage.