"random panning" bug tracked down

Today i did it. Having a free week i decided to finally catch that bug. Ok was no big deal, although i had to record for over 7 minutes until it popped up the first time.
So heres what i did:
Put a single sine wave on a pad, set to one shot (i tried with "note on" and looped, but nothing came up, i didnt ran it that long though). Recorded a sequence, full level on and tap tempo pressed, on the 16th.
Recorded the whole stuff till i heard the phasing.

When i looked at the waveform i noticed the mpc did something to the waveform:

Left the original waveform, to the right the recorded one. Is this the decay the mpc applies to one-shot samples? Why does it take longer to reach the zero? Doesnt look like a normal decay to me, more like a pitch decay.

And heres the error the mpc makes, the "panning" bug (actually its a phase shift):


I tried to determine which channel gets delayed but had to dig deeper, as the distance between the waveform before the phase-shifted one and the beginning of the earlier (the right channel) waveform was the same as the distance later waveform (left channel) to next recorded waveform. Soeey if i cant describe it better, i´ll try to draw it:



So no channel gets delayed, just some extra time inserted? I measured the distances of waveforms before and after and couldnt really interpret this, it seems the mpc adds extra time at certain moments
Heres the list:
5504
5505
5536
5504
5505
5504
5504
5537
5504
5504
5536
5505
5504
5504
5537
5504
5504
5505
5536
5504
5504
5502
5536
5504
5505
5504
5504
32 <- phaseshift
5504
5505

seems pretty periodic in the second half, maybe that event occurs around each quarter note and thus sometimes before sometimes after the beat. and if its at exact 0.0.00 it cant play the 2 sample channels at the same time (just noticed it was a stereo sample...) - just some wild speculation though..

I will try to explain to JJ, anyone knows if he knows bout this?

thanks for your deep research ! I hope it gonna help JJ to correct it.

well i didnt understand much of this post, but fixing this bug would be MAJOR!!!

i feel like a nerd because i did understand that!

- very cool that you took time to do that, good karma methinks

Nice work. If you need any help with it, be it explaining it to JJ or whatever, PM me.
JJ knows about the problem, but I think not in this detail. I guess the problem will be to determine WHY this happens. But now we know WHAT happens, which is good.
Have you already mailed it or maybe got a response?

Nice discovery. Much props for the research. That shyt is annoying!

can you upload an audio clip maybe so that I could better understand this problem?

Great analysis!

Thx for the feedback. Still like to know about that decay, anyone?

I didnt send this to JJ so far, can anyone pm me his email-adress?

I don't think it's the decay. Decay does this:

Sample without decay:
--------------------------
Sample with decay:
------------------______

Meaning decay doesn't make the sample longer, which seems to be the case with your observation.

delete

did you get an answer from JJ?

No not yet, but i sent the mail right now... i wonder if he understands what i wrote...

sorry to speak on this so late. really good work, here.

did you send the images/audio with your email? that would help a lot.

i've never captured the phase issue "on tape" so to say
but the audio and visuals would help him to understand what you are saying.