For anyone having probs with notes being cut off check this

Ok so I think I found out whats causing this but I'd like people who are experiencing it to try something.

Altho Nym's suggestion to chop your samples up instead of using long loops works it still didn't satisfy my need to understand what was going on. The problem I had with his solution is that it should still uses the same amount of voices whether you play a 16 bar loop sustained or cut it into two 8 bar loops and trigger one after another. So although it works it didn't make sense that it should...

I went over the song I had problems with and started experiementing. Made sure all my sounds werent overlaping, were set to mono trigger and turned the default release from 5 down to 0. I had also previously shorted my 2-4 bar loop notes by a tick (old trick from the day to help optimize your sequence, altho the mpc shouldnt need it).

So then I tried taking off a few extra ticks at the end so the sample stops about 4 ticks before the next bar. On some samples you can hear a gap when this happens (which proves that the sample itself is not playing on top of it self, in other words not using any extra voices).

After doing this my sound would not get cut off anymore. I tried this back and forth on various loops I had just to double check. There was a repeatable effect which proved that this was in fact helping.

Thing is it shouldn't make a difference. I already had my samples set to mono and they already had a very small but slight gap between them. All I can gather is that the MPC is not releasing the voice used for a note immediatly after you tell it to stop playing. By taking of about 3-4 ticks from the note it gives it a chance to before the next note comes in. Otherwise what was happening is that when the first note stopped playing and the second one came in was that you had for a split second two voices being used. Now multiply this across all my sequences (because in this case pretty much all my notes were 2 or 4 bar's) then you have double the polyphony being used at that one moment. This is why I think the sustained pad (in my case a 16 bar one) got cut off.

Now why this happened at bar 9 and not 5 is still a mystery to me. I have the same turnaround of notes at 5 that I do at 9, so it should actually have got cut of then. At some points it did early in my constructing of this song.
I also noticed it didn't always get cut off. I could make it kleep playing more often if I was in one of the other screens than the main one. Like if I was in step edit while things were playing my 16 bar note would not get cut off as much. So there is some element of randomness.

The reason why Nym's idea works is that even tho your using the same amount of voices by retriggering the second half of the full sample with a new note you are forcing the MPC to allocate it a voice. New notes have priority over old ones. So this overides its problem with holding onto voices from notes that have just had a note off event.

So if anyone wants to try this take any of your loops you have and shorten a few ticks. Especially at the point where your longer notes are being cut off. Altho Nym's advice works I don't always want to chop my samples becuase sometimes you can gear a glitch in the transition point, mainly on pads. Drum loops won't have this problem. But for a sustaining type sound you might get it (altho using patch phrase with some extra release set works sometimes to hide smooth the glitch). Shortening notes by a few ticks (ones that bascally played right up to the next bar) doesn't also make sense either as sometimes you'll hear the gap that is created. It depends on the sample. If your sample has a bit of silence or fade near the end of it anyways then you won't miss those extra ticks being gone.

Basically we have two options now. And for me I am at least satisfied in knowing whats actually causing this which was my main goal in experimenting more. Now I can get on with some serious work not having to worry about it being halted.

I don't always want to chop my samples becuase sometimes you can gear a glitch in the transition point, mainly on pads.

what i would suggest, if this is possible, is the following. say you were sampling pads from reason that you'd programmed on your mpc. i do this kind of thing a lot.

i'll record them onto reason via midi and then separate them, so instead of

___----___---- (w/ the pads overlapping/touching) i make it

___ ---- ____ ----

length is preserved, but this way i don't f' up the edges and there's no glitchiness.

granted, this isn't always possible. with string samples from vinyl, for instance. however, the workaround for the string scenario is to either extract each slice instead of exporting all the new samples to pgm

or, even better, to timestretch the entire loop 5 bpm or so SLOWER than you intend it to be. then chop without worrying about edges. now, turn the pgm to POLY and they will overlap without glitch. the overlap may not be pleasing to the ear, but you can fix that with decay/attack. this works wonderfully for pads/strings.

akai should not have made it so chopping fades out the edges. that's what attack/decay is for. if we want that, we'll put it on ourselves.

I'll have to try that. Only thing that stopped me from doing it before is that since they overlap you are using two voices and my whole idea was to only use one total for that particular sound. I'm afraid they wont end up overlapping anyways becuase of the poly issues I've been experiencing. The idea being that if it didn't play across measure 9 as one sample then why would it fade out across measure 9 as well? Yes the second half will trigger fine since its going to get a voice assigned to it, but the first half will probably just get cut off as it fades (blends).

Then again when I made a patch phrase of my 16 bar loop into 2 8 bar peices and then applied a long release in the convert page it worked out. I'm not sure if the slices in patch phrase can play poly and overlap. But maybe that is what made it sound seamless so in that case I can do as you suggested.

By the way the case of this pad I had it was one long sustained chord so there was no easy transition point to chop from. But the patch phrase method did work. Also the fact that convert fades out didn't affect the sample, as I had release set to like 50 anyways so the slices actually came out longer (timestretched). Even if I used extract I still got a glitch. It seems like the MPC won't play those two slices perfeclty adjacent to each other. It's not sample accurate timing. But thats fine, I like its feel and don't want it "perfect" like a DAW would be. So this is just one of the allowences you have to make considering the design.

What I'm basically going to do now is adjust my other sounds so that the pads have their polyphony ready when needed. Drums and other rhythmic elements are easier to do the workarounds it seems at this point.