Which controllers with motorized platter does Traktor support?

[nem0nic]

Did this capture happen in the period of a couple seconds?

It looks like B0 is counting off "ticks", but I don't know what the E0 (pitch bend) message is doing. Looks like it's sending the tick packet and the E0 packet together. If I had to guess based solely on this capture, I would guess that the B0 packet is an absolute message simply counting off each count of resolution, and the E0 packet is some kind of higher resolution positioning data - maybe converted from gray code or something like that.

The first thing I would try to do to determine if that E0 message is positioning data is to filter out the B0 message, then see if the E0 packet shows the same value in the same place on the platter. I would also see how each packet behaves when the platter is being manipulated. That might reveal that in fact the B0 message is the platter location message (for instance).

[Afterhour Ali]

That happened all within maybe a 10th or even 100th of a second.

I'll check for the absolute positioning idea.

EDIT: Nope it doesn't seem to be absolute positioning data. After more than a full revolution the same pitchbend value doesn't occur twice.

[nem0nic]

EDIT: Nope it doesn't seem to be absolute positioning data. After more than a full revolution the same pitchbend value doesn't occur twice.

Does the B0 message repeat after a single rotation?

[Afterhour Ali]

Since that one is only one byte long it repeats after a small part of a rotation. I'd say it repeats 50-100 times per revolution, I'll have to count that though when I'm home.

[zestoi]

maybe converted from gray code or something like that.

this would make perfect sense given the situation.

[nem0nic]

Since that one is only one byte long it repeats after a small part of a rotation.

That's interesting. Then I DO think it's counting off ticks. Have you tried moving the platter very little back and forth over the same spot? Does the E0 message report similar values in that case?

Also, can you see a pattern in E0 if you move the platter faster or slower in a single direction?

[nem0nic]

Since that one is only one byte long...

It's 3 bytes long. It's just that bytes 1 and 2 aren't changing in this case.

[Afterhour Ali]

These are the binary-pairs for the first four Pitchbend-messages of my data above:

00110011 01000011
00011001 00101110
00111000 00011010
00001000 00001001

Doesn't look like gray code.

I can't see any pattern in the Pitchbend-message no matter how I move it.

[nem0nic]

If you filter out everything except B0, is there roughly 3600 counts of resolution in a single rotation?

[Afterhour Ali]

We're on something here!

I tested four times to capture a full rotation and filtered out everything but B0.
The first revolution was spinning at normal speed, had 3001 messages.
The second one was spinning very fast -> 1833 messages
The third one was very slow (maybe 15 seconds for the revolution) -> again 3001 messages
For the fourth one I started and stopped the motor after one revolution -> 1820 messages.

This message alone should suffice to map the platter properly with some coding. Still I don't like the idea of discarding the other two bytes without knowing what they mean.