Lost Notes

[PiersW]

I have been having problems with lost notes. I appear to have finally sourced the route of the problem to either copperlan or the al88. I have two tracks playing midi on studio one and one of them misses occasional notes when the other is playing. If I turn the midi track sending to v3 off the midi notes play ok on v5 where they were originally missed. I had hitherto thought it to be a Studio one problem but to day rigged up midiox monitor to the studio one output channel on v5 and noticed that the notes were being sent by studio one, all note on and note offs were accurate, but some notes were missed in the audio I was hearing. This is a slow 4 note pattern repeated at intervals, very simple to test the studio one output. I tried with autosensing for the relevant output switched both on and off. It is perhaps an al 88 problem. Perhaps it might be helpfull to add that the notes were on quarter notes ie quantised. It is not an insurmountable problem as I can record to audio one track at a time but it is not ideal either. I love the speed of copperlan and the ability to for example switch midi routings between two computers without having to un-plug a single wire It is Stunning. I expect these are the normal teething problems in what is a stunning system.

[Copperhead]

Since there are no clear indication that CopperLan is at fault, I can only suggest the following as possible causes: Have you tried replicating the same setup but going through a USB/MIDI interface, not using anything CopperLan related? Does it give you the same results? If yes it might be due to an inadequate merging of notes from the two tracks to the same destination.
We have not heard of any problem in CopperLan like the one you describe. Please see with Presonus and Alyseum if they can help.

[PiersW]

Perhaps I did not explain this very well. The two tracks are not merged. They are routed to different copperlan inputs in this case v3 and v5 which in turn are routed to corresponding inputs on the al 88. By putting a midi monitor between the v5 out which is the track the notes were disappearing from and al 88 interface. I monitored the midi coming out of my Daw and it clearly showed notes I wasn't hearing. I had already been in contact with Presonus on this problem, I have now proved it is not Presonus from my test. As I said I suspected it was the al 88, in particular because the midi note on and note offs are appearing in v5 copperlan before it goes to the midi monitor. I did not know whether you were somehow linked to alyseum hence mentioning this on the forum. Perhaps the automatic firmware update made me think that way. The fact remains that I am losing notes somewhere between copperlan and the al88 midi output. I will contact alyseum and ask them.

[Copperhead]

In order the find where the problem comes from, could you please connect the faulty MIDI stream coming out of the AL-88 to a MIDI-USB interface that will be monitored via MIDI-OX?
You could then tell for sure if and what messages are modified or missing. You could also check if there is any interleaving (multiple Note-ON (or Note-OFF) of the same note value sent consecutively.
Hoping to be helpful...

[PiersW]

I shall do it but it will take me time to actually set up as I do not have usb set up on my system anymore and don't we all just love rummaging around in dark inaccessible corners changing wires around, - but I do see the necessity, so please bear with me! You will hear the result tomorrow. I hope I won't be putting the embarrassed emoticon at the end of my next email

[PiersW]

I did what you suggested and sent a screen shot and description to Willi at Alyseum asking him to forward to you but now I see I can upload a screen shot here so here it is. The out of al88 to laptop monitoring midi and on to the synth. The sequence of 4 notes is repeated but the second time only shows note offs. Actually sometimes I get neither note ons or note offs and then both in the right place within the sequence of four notes which is as I recollect a cut and paste repeat. I actually removed note overlaps to make it easier to read the midi data on the screen.
Also I don't know what all that controller reset stuff is. When I routed in copperlan I tried just 1 channel out as needed but for some reason didn't get a signal so ended up routing all channels hence the 16 midi channels of controller data.
Hope this helps.

[Copperhead]

What you describe could be due to the temporary loss of the network connection.
In which case the AL-88 generates a string of All Notes Off + Reset controllers on every channel.
To validate that hypothesis, please do the following:
Put your monitoring program (via USB) to the MIDI out of the AL-88.
Establish a single CopperLan connection (one channel) between Studio One and the MIDI out of the AL-88.
Do not play Studio One - it must be stopped.
If you see any Control Change with 7B (All note Off) or 79 (Reset Controller), you have a faulty connection between your computer and the AL-88.
Try connecting the AL-88 straight on the Ethernet connector of the computer.

[Copperhead]

Besides a possible physical link problem, there is something weird in your short MIDI listing:
We see that every note-on gets two corresponding note-offs:
- one just after, with velocity zero,
- a second one, after all notes have played (on+off), with a velocity corresponding to the note-on!!!
The CopperLan networking is not memorizing anything like notes and their velocity.
There is no chance to recreate deliberately or accidentally a four note sequence whose velocities correspond to past note-ons. Therefore that aspect is clearly out of CopperLan responsibility.
You could check by playing Studio One to the USB out, use a MIDI cable to loop-back the USB MIDI out to the USB MIDI in, and monitor the result in your program.

[PiersW]

As I say It is a repeat of the earlier sequence of notes, except this time the note ons were not sent but the note offs were. Nothing to do with Copperlan remembering anything, but the not ons not being sent to the synth. I suppose I could try another midi lead, but as it is intermittent it seems unlikely that its that. What is interesting that you have pointed out which I hadn't noticed was the velocity being the same as the note ons, sort of as if the velocity was sent but not the note on. I agree its weird. Unfortunately its not very satisfactory I don't know how midi works but I did genuinely do the test as described yesterday and your recommended test today, I suspect its the al88 as copperlan definitely received the data in yesterday on v5 in the tests as I listened to the notes missing in the audio, and as shown in the midi monitor in todays tests rather unsurprisingly ie when the audio is missing the data hasn't been sent. I do not think that this proves copperlan or the al 88 absolved but I hope they are and that there is another explanation, I want everything to work properly as much as you. Incidentaly I tried to repeat the problem to show a friend today and was confused that despite several plays that I wasn't getting it and then realised that the other midi track was muted. as soon as I unmuted the other midi track going to v3 the problem re occurred. Also I might add that the notes are not fast eight to half a bar in length.

[PiersW]

Have just read your earlier post, I only read the last one sorry. Do you mean Studio one to the Midi in of the al 88 (you said midi out) (one channel) and then on out to the midi monitor via an alternate usb connection.Please let my know in case I am missing something. ethernet dropout could be a cause that I hadn't considered.
Also I shall connect the track to the usb midi out i fortunately still have on my computer and on to the synth. If I get the problem, the ethernet, copperlan and al 88 will be completely absolved!!