Nord User Forum

Hello ye NL3 fanatics and experts, I've got a good one for ya here.

So I've got this new-to-me NL3, and I've noticed a very odd thing: between alternating keypresses (when in polyphonic mode), the phase relationship between the oscillators is not constant at the onset of the keypress - even if the oscillators have exactly the same pitch and shape! For instance:

1. Set up a simple init patch with OSC MIX set dead centre to 0. This ensures that no LFOs or other modifiers are adjusting the oscillator frequencies behind your back, so to speak.
2. Strike any single key for any single duration, then release and strike again. Notice how the "quality" of the note has changed, this is because the phase relationship between the two oscillators has shifted. This is easily seen when using an audio oscilloscope, but even more easily heard with ordinary ears.

This can be heard with chords too.

I notice that in general they seem to alternate between two relationships, with the double sine alternating between a much wider variety (from nearly 180-degree cancellation to nearly 0 degree coherence and several steps in between).

This difference can be very un-subtle and effectively ruins the patches/sound.

Is this a known thing with the NL3, never to be fixed? Is it something odd with mine? Is there a workaround so I can guarantee synch between, e.g. a square and triangle?

I don't know specifically about the NL3, but in general, going back to a purely analog synth, the oscillators are free running and hence at the onset of a given note the phase relationship will be different each time. For a monophonic synth the notes played will have a certain phase relationship to each other as it's the same voice playing all the notes, in a polyphonic synth one tends to get a cyclic behavior if the synth reallocates a new voice for every keypress.

In the digital world, there is a design choice, sometimes even with a user parameter to control the behavior, because the software has full control over the oscillator behavior. In general though, if the oscillators start at the same phase for every new trigger, the result is the sound becomes static. sounding the same every time, like playing back a sample, so I think that behavior is more often avoided than not. Still as you say, sometimes it is desirable to have a fixed phase relationship, especially when using the oscillators in complex ways where the phase difference between the oscillators greatly influences the sound.

So in the end, I'd say that it's probably a well known and intended design choice when it comes to the NL3, in order to give it a more 'analog' character. As I mentioned, some synths, such as the Access Virus, make this a user controllable parameter (in the Virus it is called 'Phase Init'). Clavia tend to be not so hot on allowing the user to configure things on this level, so it is in line with their style not to make this controllable.

Hi Ricard,

Yes, I considered that possibility - that the system might be simulating free running oscillators, but the data does not show that this is the case. A free running oscillator tuned to exactly the same frequency (plus minus the margin of error in analogue circuitry) should phase, if at all, only very slowly with respect to the other oscillator at the same pitch. In this case, the phase relationships are drastic, and clearly alternating in the case of square and saw waves - this neither aligns with voice allocation nor with the idea of free running, in which all phase relationships are possible, depending on the interval between the two note presses.

Instead, the system seems to randomly set the phase of the two oscillators per "voice" and leave it static, the worst of all worlds - you get neither a convincing or useful analogue-style simulation nor the practical repeatability and sonic adjustment potential of key-triggered phase reset.

I'm loving my NL3 but this neither sounds like my analogue Moogs and Prophets nor like the other Nords I own, either. Quite frustrating for such an otherwise powerful synth.

If I am not mistaken, the oscillator waves start at a random point in the phase on NL3 and NL4, while they always start at the same point on the NLA1. I only own the 4, but I think that's one of the subtle differences between A1 and 4(3). It has pros and cons. Same point is better for instant attack basses, random point makes for example fast arpeggios more alive. On some analog dco synths you also have that effect when you load a patch with two osc on the same pitch, but it's gone as soon as one osc is detuned for the first time