Oscilator bug?

Voltan · Post by **Voltan** » Sun Nov 18, 2007 9:26 pm

This happens to me every time:

-Select patch INT-C 045 - Mega Sync Lead
-Play a not-very-high note, say, a D3
-then play legato to a pretty much higher note (D5 or higher) - just hit D5 while holding D3
-the sound gets awfully detuned from there

It seems that while playing that lead legato up, one of the voices stops at some point at doesn't get any higher - it remaines at its "top-note" while the other voices of the program go higher - causing a nasty off-tune sound

I experimented with soloing the multisamples of that particular patch - looks like one of them is ok but all the others will stop at some note - different for each of them.
The "top note" also seems to be different depending on the note you start with - not for every note tough - more like for groups of them (guess this has to do with the multisamples' zones or something)

It does not happen when portamento is on
It does happen when portamento AND fingered are on

It also happens on some other patches (like C-061 Espress Lead)

Can anyone reproduce this on their units?

r12 · Post by **r12** » Mon Nov 19, 2007 9:07 am

yes, I'm not sure but it seems like multisamples zones can't be transposed more than 43 semi-tones...

maybe fixed in future OS

Lorenzo · Post by **Lorenzo** » Mon Nov 19, 2007 9:26 am

can't hear the problem here cos I don't have an M3... but maybe it's an aliasing problem...

r12 · Post by **r12** » Mon Nov 19, 2007 11:29 am

I don't think it's an aliasing problem

some multisamples are affected and others not...

maybe samples compression

Lorenzo · Post by **Lorenzo** » Mon Nov 19, 2007 11:39 am

maybe you can post an mp3... I don't know what you mean with "alising" but if the samples are not good enough if you try to play them pitched on higher notes you can here digital noises from some subtle ones to heavy aliasing that doesn't let you to here the sound anymore... if it's the case the problem is with the samples themself and no one can fix it...
regards, Lorenzo

Voltan · Post by **Voltan** » Mon Nov 19, 2007 12:40 pm

It's not producing noises when shifting higher - that would be ok as there's nothing much to do about it really.
The thing is the samples' pitch can be shifted to a certain point and no further - they play the same note regardless of the key you hit.
I'll post a recording later.

Voltan · Post by **Voltan** » Mon Nov 19, 2007 1:24 pm

Here it is: http://download.yousendit.com/3639AC4D0A7C97DA
I recorded it twice:
- with the factory preset
- with 1 multisample on each OSC, one of them playing properly and the OSCs set to seperate channels (left and right) - so you can here it clearly.

r12 · Post by **r12** » Mon Nov 19, 2007 1:55 pm

yes that's right, and if you pitch down to -12 the OSC that is pitched up to +12, the effect disapear...

Lorenzo · Post by **Lorenzo** » Mon Nov 19, 2007 2:11 pm

uops... not certainly aliasing... this sound really strange... maybe it's really a bug.
Regards, Lorenzo

Voltan · Post by **Voltan** » Mon Nov 19, 2007 2:49 pm

Well I hope someone at Korg will take a look at this

lcmorley · Post by **lcmorley** » Mon Nov 19, 2007 3:19 pm

Voltan wrote:Well I hope someone at Korg will take a look at this

You may be better off reporting the issue on the Oasys forum as I know the Korg bods seem to look at that forum a lot.

Diego · Post by **Diego** » Mon Nov 19, 2007 3:32 pm

OMG!!!!

It's the same was happening on the Trinity 12 years ago!!!
Try the monster lead patch on the Trinity and do a note-jump higher than 2 octave and you will ear the same result...incredible!!!

regards

r12 · Post by **r12** » Tue Nov 20, 2007 11:49 am

I've tried it on triton classic and...working perfectly

you can see the problem without portamento : try to transpose up at max some multisamples :

1 : osc page : pitch up at +12
2 : pitch page : set sw1 at +12 and put sw1 on
3 : set pitch env at +12 and pitch env parameters at 99/99/99/99...
4 : set lfo1 mod to pitch at +12 and lof1 at square / frequency at 0

normally you have a +48 semitones transpose and you will see the problem immediately
some multisamples work fine and others bug

Lorenzo · Post by **Lorenzo** » Tue Nov 20, 2007 11:55 am

I think that the one you write about is the aliasing problem I was talking about, a different thing than the problem of the m3... I still have to see if the problem could be found on the triton too.
regards, Lorenzo

Timo · Post by **Timo** » Thu Jan 31, 2008 9:24 pm

The way it appeared to occur on the Trinity was if your hand strayed into another sample-region (of the multi-sample) to that of the original key, it couldn't seem to be able to deal with it as it couldn't pitch up or down the sample that far. (It doesn't play the sample from the new region, but uses the original sample from the original region, but this can only be pitched up/down so much).

http://www.korgforums.com/forum/phpBB2/ ... ght=legato

It's sad to hear that the M3 has the same problem that the Trinity had all those years ago. I can't understand why any new note can't trigger a new sample from the new sample region, while just holding the EGs from the first key pressed. I can only think that it's due to needing to somehow bypass the attack portion of the new sample, or to start the sample at a later point in time so as to match that of the sample from the first key-press, which may be complex to execute and require extra DSP overheads in the form of sample analysis/tracking and could possibly affect polyphony if the legato function was enabled in this way.

I've always tended to use my Moss, Virus or Radias to do this stuff, as the bug is related to multisample-based ROMpler usage. I'm unsure as to whether it's limited to just Korg's ROMplers, as I don't have any others (other than Roland MC505, which can't use legato)? Does the Motif or Fantom have legato, and if so do they suffer from this?