#14

@ngwese i’d love to merge this into the main firmware!

4 Likes
#15

I don’t have an Oplab I can test with but if you could record the MIDI stream which is generated when you change modes and send it to me I can attempt to find and fix that problem. Recording the MIDI output from your SubPhatty as you play (key plus pitch bend) could also be of use in tracking down these bugs.

The MIDI spec has some finicky parts around running status which might not be getting handled properly. The pitch bend support was tested using a Push 1 (in user mode) as well as a Novation SL 61 but not as heavily as other parts of the code. I’ve found each controller can have it’s own quirks - the Zendrum controllers (I have a Zendrum ZAP) send note on and note off together in one burst which certain other devices didn’t work well with (it uses running status and zero velocity note on(s) in place of explicit not off(s) IIRC).

#16

Took me couple attempts but I did install it eventually. Works like a charm with Qnexus. Thanks you ngwese!!
BTW, I though of it for a bit, maybe you can do something like built in arpeggiator and replace tracking knob with speed of arpeggios.

1 Like
#17

@ngwese i’m absolutely loving this with the op-1. i was using the oplab to sequence, but this is waaaay more convenient. one thing i’m missing is a midi clock, so i can synchronize my other sequencers (meadowphysics & teletype) with the 4-track tape playback on the op-1. would it be hard to add this? perhaps holding the front panel button for a few seconds could switch output 1 (currently mod wheel) to be a clock out?

3 Likes
#18

Anyone had any luck getting this to work with a Korg Microkey? I’ve tried both the latest HEAD and the linked hex and neither are working properly (though it is working a bit, if I wiggle the pitch bend it elicits a response).

#19

so, is qunexus a “go” with this?

#20

it might be a good idea to make sure the leds are turned off (via editor) or use switch in between, to prevent whine. but it seems yes.

need to test myself…

#21

I’ve had bad luck with the QuNexus:

#22

out of curiosity, did you disable the LEDs? did it really brick the ES so you had to reflash it?

#23

No, I left the LEDs on. Yes, it was causing issues that wouldn’t disappear until a reflash was performed. I didn’t end up going through the trouble of fixing it, since the QN already has gate, pitch, and two CV outputs on the unit itself.

It was time consuming to troubleshoot variations in setup, so I’m sticking with using QN as standalone, and using the OP-1 with ES.

1 Like
#24

sounds like an interesting idea. i hope to have some time toward the end of this month to investigate this and try to make things more robust.

#25

Some general updates:

  • there looks to be some low level bugs in the midi receiver code which may well be the source of some of the erratic response. I’ve started testing some changes in that area which should improve robustness (both on the modules and aleph)
  • I’m able to get earthsea to overflow its internal event queue when receiving midi from a DSI Evolver set to transmit all parameter changes over midi. This is most likely the what is going on when the earthsea lights stop changing and the module appears to have locked up. Hopefully I can now get to the bottom of the lockups - oddly my early tests a few days ago were lockup free but I’m now wondering if I in fact had slightly different firmware from what was released.

Regarding adding midi clock support. I’ve started to work on that using an Evolver desktop as the source. @shellfritsch can the OP-1 send midi program changes or sysex? I was wondering if it might make sense to have different output configurations selectable be midi program change.

In an attempt to remain modeless I’m going to do a little experiment with using the current velocity output as a gate with variable height. If that works then midi clock could be edge…

1 Like
#26

If it helps I can confirm that it works really bad on Korg microKey-25. The only thing tha works is the modulationstick (pitch bend).
It works ok on a Samson carbon-61.

1 Like
#27

…if there is anyone in the los angeles area with a microKey-25 they’d be willing to loan me for testing PM me.

#28

i’m pretty sure it does not send either of these. there is an “ableton live mode” that turns the op-1 into a control surface:

https://www.imaginando.pt/products/op1/help/usage

though i have no idea what type of data op-1 is sending when in this mode…

#29

I do have a microkey-25, and I have had the problems described, but I’m not even in the right country to loan it to you. But I can definitely help with any debugging if you need it.

#30

I might not need it after all. I believe I’ve found a combination of bug fixes, timer tweaks, and rx buffer size which makes this much more reliable. I hope to have a build for people with a QN or microkey to test this weekend.

1 Like
#31

how goes?

any progress with clock out?

also, i was playing with earthsea the other day and it started experiencing the slow-down again. after sitting for a while (no reboot) it eventually returned to normal response.

#32

I’m confident that I’ve found and fixed the source of the midi receive problems reported by @sam and @barksten. I just need to clean up the change and get that merged in.

On the clock front things are much more involved. I’m made two different attempts at using the incoming midi messages to build a “clock” signal but I was unhappy with the amount of jitter in the cycle, small amounts of drifting relative to the note gate on/off, and instability when running at high BPM values (the module was fine but the software clocking mechanism started to miss beats).

I have started looking into the possibility of adding a completely different clocking mechanism to the module code base - one which is built directly on top of the remaining unused timer/counter hardware. Ideally I can get the hardware timer to drive the “edge” output directly for a really stable hardware based clock (then use the incoming midi data to change the phase and period as needed to stay in sync). Doing this requires delving into hardware level details on the avr chip.

4 Likes
#33

thanks for the update!