Osc Phase Issue

[Chrono]

Im stumbling on some problems when comparing the playability of the NMG2 and the Access Virus A.

so imagine 2 osc's connected to a Xfade. And in this case using 8 polyfonic Voices. OSC 2 is very slighlty detuned for the phase changes to become noticable.

NMG2: if u trigger a melody with just 1 key pressed at a time
the phase jumps around like changing its position everytime you press another or the same key.

Virus A: if u trigger a melody with just 1 key pressed at a time or repeat the same note. the phase moves without jumping to different positions.

-----------------------------------------------------------------------------
(for reference)
This problem won't happen in the NMG2 if voices are set to MONO.

When holding 1 key on the access virus and pressing a 2nd, then the phase changes its step. Noticable if you release the 2nd key again.
------------------------------------------------------------------------------

test patch is in the attachement
progress so far:
if the patch is reinitialized the technique works on 1 key
but after pressing more different notes the phase starts jumping to
different possitions again.

(solution should work with polyfonic setting)

I hope someone can help with this one!

---

NMG2 _ Access Virus A Phase.mp3
Description:	Audio Reference of the problem 1st: NMG2 ,2nd Virus A, 3th NMG2 detuned, 4th Virus A detuned.	Download (listen)
Filename:	NMG2 _ Access Virus A Phase.mp3
Filesize:	572.97 KB
Downloaded:	1187 Time(s)

---

---

OSC Virus Phase-test01.pch2
Description:	simple test structure	Download (listen)
Filename:	OSC Virus Phase-test01.pch2
Filesize:	1.46 KB
Downloaded:	4456 Time(s)

---

[ian-s]

The G2 only has the one poly key assign mode, which will always trigger a new 'voice' every time a key is pressed.
Most early voice allocation synthesizers offered a choice of 3 poly modes. The virus might to, or perhaps they just chose a different fixed mode.

You can cheat in a limited way by using a single OscShpB with the double saw mode and a saw lfo modulating the shape 100%. The lfo rate then becomes your detune amount and you can select the Mono mode on the lfo. This will force all voices to have the same phase offset while the phase is still free moving.
Only good for that two saw 50:50 sound though.

[Chrono]

i see! , but still its strange the Logic "NAND"GATE manges to do it the right way for a moment.
_____________________________________________________________
You can reset the emulation by selecting GATE and then PHASE again in the path FX section.
__________________

just after a couple notes the more jumping occurs. I think that has something to do with the amount of voices, so the phase does not know wich one to choose and just jumps around.

It also seems to jump around quicker if u press notes further away from eachother.

I just cant think of an obvious solution to make it fixed.

It might need some serious brainwork to find this solution. but im sure there should be one.

i do not doubt your knowledge, ill have a try with the Shape osc. But it's main purpose is to work on every osc with a sync input.

[Roland Kuit]

And if you take an osc with a sync input. The keyboard module gate out to sync in osc.? Also LFOB module can by reset by the gate out of the KB module. So you can leave it on Poly On, when modulating the Shape osc.

[varice]

The problem that you describe is caused by the polyphonic “rotate” voice allocation mode of the G2. As Ian pointed out above in his post, the G2 *always* assigns each new note event to another available poly voice. But, just like a real poly analog synth, the free running oscillators in each voice of the G2 are phase independent. So, for a voice allocation scheme that always selects another voice for each new note, it will force the revelation of the phase differences of the oscillators in each poly voice, which you (along with others, including myself) notice as a problem.

The solution would be for the G2 to have an option for a more natural poly voice allocation mode “repeat”, where repeated notes are assigned to the same poly voice. Apparently, the Virus synth sound that you posted is based on a “repeat” mode.

Many poly synths have an option to select between voice allocation modes. I sure wish that the G2 also had these options. The default G2 poly “rotate” voice allocation mode limitation was discussed long ago. The last post by Dave Peck in this linked topic describes *very well* why it is *not* desirable to be limited only to the “rotate” mode:

http://electro-music.com/forum/topic-15459.html

[Chrono]

Thanks for that!
makes it more clear to understand the voice allocations compared with eachother.

I will post my G2-> Virus A patch soon.
just need to reduce some Cycles/Memory to fit all functions.
i will put that in the Synth section.