BUG - not all channels transposing

[oldgearguy]

My goal - input notes on DIN In 1, output same notes transposed 3 octaves higher on DIN Out 2.  Do this for all channels being sent from DIN In 1.

Set up Transpose MOD, any channel, +12
Insert 3 copies of TR 1 into route from DIN In 1 to DIN Out 2
Observe that channels 1-5 transpose correctly, channels 6-8 does not transpose.  

If I do just 1 transpose, channels 1-7 transpose an octave and 8 does not.
insert TP1 twice and channels 1-6 transpose and 7,8 do not.

This is kind of annoying and I really needed this functionality.
OS version 1.1.052


Have not tested channels 9-16 yet, but would assume they need attention as well.

[Jesse Johannesen]

Sounds like an off by one error in the modifier code. I'll see if I can get Steve to fix it in the next update. Thanks you for finding this!

---

Can you try sending with 1 modifier, but instead of ch 1-8 try sending on higher channels and see if signal gets modified on say 9 or above?

[oldgearguy]

Sounds like an off by one error in the modifier code. I'll see if I can get Steve to fix it in the next update. Thanks you for finding this!

---

Can you try sending with 1 modifier, but instead of ch 1-8 try sending on higher channels and see if signal gets modified on say 9 or above?

I'd need to hook up some other machine.  The current device (Grex MXF8 - kind of hard to find) is hardcoded channels 1-8, note number choice 26 to 33.

It may not be immediately - things have been a bit more than hectic around here.

[oldgearguy]

yeah, something is really not right.
Today, I can't get any transpose happening from 2 different sequencers on any channel.
In addition, if you have a Velocity map set up on a route and you send channel pressure through that route, your dest max is set to -128 in that Vel Map.

No idea what's going on here and I've run out of time for the day.

If all this is legit, it should be very easy to reproduce with an OXI or any sequencer that supports multiple channels.
It would be nice to get confirmation that it's not something specific to my setup.

[oldgearguy]

MIDI note on is indicated by a 0x90 byte followed by the note value and velocity
There are two ways to transmit MIDI note OFF - 0x80, same note value, release velocity (or whatever you want). --- OR ---
0x90, same note value, 0x00 velocity

Running status might be what is causing the problem 

More data -- it appears that the transpose MOD doesn't work if the transmitting device sends MIDI Note on and then MIDI note off as Note On, velocity 0 (using 0x90 exclusively).
Devices that send Note on (0x90) and the Note off (0x80) transpose correctly

Also - once you 'hose up' the route with the transpose on it by using the second Note Off method, the MRCC is all confused - you see note On in one octave and note off in a different one.

Bottom line for the CL guys - set up routes to a computer, use MIDI monitor, use a bunch of different note generating devices and watch what happens with transpose MOD and different methods of Note OFF.
Likely other MODs are broken by this as well and then there's still the multi-channel thing to debug as well.

[Darryl]

Hey Tom, we're looking into this. Using NDLR, we can set Note On to zero velocity so even if we don't have other equipment that does it, at least we have a convenient way to generate those messages. When we tried replicating the issue without concern for the Note On zero velocity issue it seemed that everything was working as expected. We're working through a list of issues, so we might have more questions when we get to this one.

[oldgearguy]

Hey Tom, we're looking into this. Using NDLR, we can set Note On to zero velocity so even if we don't have other equipment that does it, at least we have a convenient way to generate those messages. When we tried replicating the issue without concern for the Note On zero velocity issue it seemed that everything was working as expected. We're working through a list of issues, so we might have more questions when we get to this one.

Ok, so here's more data now that I'm home and have some time.
The key to causing failures seems to adding 4 or more MODs.

The device in question (Grex MXF8) is regularly sending out a MIDI note (33d in this case) on channels 1-8.
BUT - the notes are being sent like this -
Note OFF - 80h 21h 00h
Note ON - 90h 21h 7Fh  (for channel 1)

Start with a rebooted MRCC, blank preset.

So, the note OFF is first, then note ON.  The purpose is triggering loops in a sequencers/sampler at regular intervals.
Now, I create a Transpose #1 -- channel Any, Amount 12 and a Transpose #2 -- channel Any, Amount 3

Set up a route between the MXF8 and the computer.  Insert TP 1 as a MOD in slots 1, 2 and 3.
Start playing -- nothing unusual happens.  The notes on all 8 channels are transposed correctly.
Stop play.  Add a 4th TP1 MOD.  Start play.
Now channels 7 and 8 aren't transposing.
Stop play.  Add TP2 as the 5th MOD.  Start play.
Now channels 5 - 8 do not transpose.

Even if you delete the last two MODs, the transpose is still messed up on some channels.  You have to remove all the TP MODs and then it's OK, but as soon as you add some back, it gets bad again.

I just tried it again with another sequencer and if I set the step length long (like 2 to 4 bars) so that the Note ON time is long, it also messes up (and this sequencer uses Note on = 90h, note off = 90h, velocity 0)

[Jesse Johannesen]

Great info! I just sent you an email before reading this, saying we weren't able to repro. This info may be what we needed, so I will make sure we get it tested and see if we can reproduce with this. Let you know what we find.

[oldgearguy]

What's the status of this bug?  It is preventing me from using at least one of the MIDI devices I have here.
Easy to reproduce -- bring up any multi-track DAW.
Create 8 tracks.  Insert a single MIDI note on each track (say C1).  
In the MRCC, set up a Transpose MOD - any channel, +12.
Create a route from the DAW to any multi-timbral device (alternate option later).
Hit Play in the DAW - all 8 notes sound.
Add a TP1 MOD to the route.  Still all 8 play.
Add two more TP1 MODs (now transposing up 3 octaves).  All notes play.
Add a 4th and a 5th and you'll start to see channels 7 and 8 not being transposed.
I don't know if you had 16 channels in parallel how soon you'd start losing notes (maybe after 2nd TP1?)

Alternate approach (no audio required)
Set the DAW up same as before.  Route the DAW out to a DIN channel (say channel 7).
Cable up Channel 7 out to channel 2 in on the MRCC.
Route channel 2 to the PC port.
Use MIDI Monitor/MIDI-Ox to watch what happens when you keep adding TP1's

[Jesse Johannesen]

I think last we tried this we couldn't reproduce it, I will sit down and give it a shot with the above steps and see if that does it.