Weird midi crossed notes

[BrunoParticipant]

Hi guys!

I’m starting to try the zaquencer on Linux with software and a simple midi device, and have been having some issues that popup when I start to change too many steps at once.

The issue is I start with various software connected through ALSA on the 4 different channels. I change notes and paremeters around, and after some event I can’t tell I start getting notes from different Zaquencer channels arriving to different channels on the software. This happens to the point of having notes on some of the synths that only show up if I have those 2 channels on on the zaquencer. If I solo each I get notes on channel 1, for example, but if both are on I get notes on both channels.

I can’t tell if this may be due to some bug in ALSA, my MIDI interface (it’s a cheap one), or something that goes crazy on my BCR when I change to many things at once. Normally happens when using the change all steps at once function.

Anyone else using this on Linux? Or with a similar problem?

Cheers!

[BrunoParticipant]

Hmmm is it possible that these simple midi interfaces can’t handle 4 channels simultaneously, right? Could this be the problem?

[BrunoParticipant]

OK, I tried to get what the zaquencer was sending to the interface from a clean init, and just outputting one note on each channel on first note of each row (muted all others, ch1 in row 1, ch2 in row 2, etc):

Waiting for data. Press Ctrl+C to end.
Source Event Ch Data
0:1 Port subscribed 129:0 -> 134:0
20:0 Note on 0, note 63, velocity 127
20:0 Note off 0, note 63
20:0 Note on 1, note 63, velocity 127
20:0 Note off 0, note 63
20:0 Note on 2, note 63, velocity 127
20:0 Note off 0, note 63
20:0 Note on 3, note 63, velocity 127
20:0 Note off 0, note 63

The notes on are arriving on a channel below the expected (is this just an error from the software interface?) but notes off come on the same 0 channel.

So, if channel 0 here is the omni channel that goes to all channels, I’d get notes off for notes on different channels, so they’d clash, right?

Now if this is the case, is this an issue with ALSA or my MIDI interface or something on the zaquencer?
Maybe I flashed it wrongly?
Should init reach the bottom row of encoders? Here it stops on the previous to the last.

Thanks!

[ChristianKeymaster]

Hi Bruno,
ok I think I´ve understood your problem.
Unfortunately I haven´t heard of it before nor do I have a quick solution, but maybe I can help you get to the bottom of this.
First, to address your questions:
- the init does not reach the end of the encoders but only the first 3 rows, that´s correct
- midi interfaces should be able to handle all channels simultaneously, even the cheapest ones. if not, there would be something wrong with your interface
- the midi output looks a bit strange to me, especially the note offs coming in on channel 0.
By the way, in many midi monitor softwares channel 0 is actually channel 1. I´m not sure there is something like an omni channel. Especially not in the Zaquencer, one midi message has only one midi channel.

This is what I get when I record the output of a freshly inited Zaquencer V1.0 with all four tracks enabled in MidiOx (Windows), watch out, it´s HEX, so 3F=63 and 7F=127:
Type–Note–Vel—Chan
90—-3F—-7F—-1—-Note On
93—-3F—-7F—-4—-Note On
92—-3F—-7F—-3—-Note On
91—-3F—-7F—-2—-Note On
93—-3F—-00—-4—-Note Off
92—-3F—-00—-3—-Note Off
91—-3F—-00—-2—-Note Off
90—-3F—-00—-1—-Note Off

I´ll write some ideas in a new post.

• This reply was modified 10 years, 1 month ago by  Christian.

[ChristianKeymaster]

Some ideas for troubleshooting:

- can you switch around Midi Interfaces/Computer/Hardware?
Maybe go to a friend´s place for trying out different setups?
Or try another Midi Monitor software? As I said the output seemed strange to me.

- just to be safe, try reflashing the whole firmware (not only reinit, but uploading the firmware)

- can you reproduce it more exact in a way, saying “If I do xyz it happens”. I could try out your exact steps here and see what happens

That´s all I got for now. Please keep me posted on this!

[BrunoParticipant]

The same as my first experiment from a different interface (the same kind of cheap one, I have two), reflashed the firmware and initialized the parameters:
$ amidi -d -p hw:1,0,0
90 3F 7F
90 3F 00
91 3F 7F
90 3F 00
92 3F 7F
90 3F 00
93 3F 7F
90 3F 00

As your suggestion (4 channels on from a fresh init):
90 3F 7F
93 3F 7F
92 3F 7F
91 3F 7F
93 3F 00
92 3F 00
91 3F 00
90 3F 00
90 3F 7F
93 3F 7F
92 3F 7F
91 3F 7F
93 3F 00
92 3F 00
91 3F 00
90 3F 00

(this seems to be right, no? notes on and off on different channels)

I used 4 instances of a synth (amsynth) and it worked perfectly after the init, until I started to mess around with the notes, mostly using the wave generator and change all notes, so things that messed with the patterns a lot, for notes and mutes. After a few of these changes the channels become all messed up. There are some notes on the correct channel only if I have more than one channel ON, if I have each channel ON solo I only get notes on channel 1. I can recreate this every time.

Oh, and this is after I restarted both the USB interface and the bcr (storing the pattern and recalling it), the behaviour continues.

And it gets weirder… It only works perfectly as I said above if I have 4 channels running. This way the routing to each synth works well, if I solo a channel everything starts going to channel 1. If I mess with the patterns things start to get weirder, because if I have notes on the same step in different channels they go to the right channel, if notes are solo on any on channel, they go to channel 1.

Is there a way to dump the firmware or the pattern so I can send you? Could you understand what’s the problem from that?

Here’s another try with 8 steps sequences:
ch1 solo
90 3F 7F
90 3F 00

ch2 solo
91 3F 7F
90 3F 00

ch3 solo
92 3F 7F
90 3F 00

ch4 solo
93 3F 7F
90 3F 00

ch1+ch2
90 3F 7F
91 3F 7F
90 3F 00
90 3F 00

ch1+ch2+ch3
90 3F 7F
92 3F 7F
91 3F 7F
92 3F 00
91 3F 00
90 3F 00

ch1+ch2+ch3+ch4
90 3F 7F
93 3F 7F
92 3F 7F
91 3F 7F
93 3F 00
92 3F 00
91 3F 00
90 3F 00

ch2+ch3
91 3F 7F
92 3F 7F
90 3F 00
91 3F 00

ch3+ch4
92 3F 7F
93 3F 7F
90 3F 00
92 3F 00

ch1+ch3
90 3F 7F
92 3F 7F
90 3F 00
90 3F 00

ch1+ch4
90 3F 7F
93 3F 7F
90 3F 00
90 3F 00

Does this help at all?

• This reply was modified 10 years, 1 month ago by  Bruno.

[BrunoParticipant]

(one note on the first step of 8 steps sequences)

[ChristianKeymaster]

Hi Bruno,

I´ve looked at all the data and there is something fishy going on with your individual BCR and/or setup. This is the first time I hear from this error. Also changing the notes cannot be the culprit by itself as that has been tested 1000s of times. I´ll send you an email to discuss your further options.

[Author]

Posts