can oscillator sync input be used as a phase reset/retrig (ie from a gate)?

[cnt]

I want to reset the phase of the oscillator with each note hit (ie input from a gate that triggers the oscillator to restart). Not sure if Sync Input works the same way in all VCO either. Any tips?

I'm trying to do it in both Eurorack hardware and VCVrack.

[S0lo]

It should work. But no guarantee on all VCOs. I mean, some manufacturer might be doing things a bit differently.

[foosnark]

I want to reset the phase of the oscillator with each note hit (ie input from a gate that triggers the oscillator to restart). Not sure if Sync Input works the same way in all VCO either.

Some VCOs will reset phase to 0 when it receives a sync pulse. Some will reverse the direction instead, and more rarely, some will do other things. Intellijel Shapeshifter has a whole bunch of different sync modes, some of which are really odd.

In most cases you should be able to use a gate just fine. But in hardware modular, some digital VCOs will have latency at the sync input. This doesn't matter for syncing to another VCO, but if you're trying to synchronize a gate and the attack of an envelope, it's a problem. (This is one reason why a few digital VCOs, like Plaits, have built-in envelopes.)

[S0lo]

It is true that some/many digital VCOs struggle with sync. And it's really mainly because of aliasing. Sync is a very aliasing prone operation digitally. Specially when the slave is a none standard waveform (like wavetable or some complex form). And thats probably why some digital VCO manufacturers choose to do only internal sync or even abandon sync all together.

To sync immediately is sometimes problematic because the master is sampled. So some manufacturers might wait a few microseconds to figure out the master frequency precisely. And then do the sync internally and independently. That is probably the reason for the latency.

Obviously in analog, all of this is not a problem since analog is aliasing free, and time resolution is extremely high. When a trigger comes, just reset the slave and all is good.

[foosnark]

It is true that some/many digital VCOs struggle with sync. And it's really mainly because of aliasing.

I guess you're assuming an oscillator design that is bandlimited at the core, rather than oversampling and using an aliasing filter at the DAC. Which is not universal by any means.

In fact, the one I know of that does things that way is Odessa -- and it's additive, it simply doesn't bother to generate partials above 21KHz. You can make it alias with strong enough TZFM. But it lacks a sync jack (which is unfortunate, because you can detune the partials and/or voices and force-syncing them with the Voices button sounds really cool )

Anything with wavetables (especially custom ones), phase modulation, waveshaping etc. is much more likely to oversample and use an antialiasing filter at the DAC. Sometimes to extremes -- Shapeshifter has an internal sample rate of 25MHz, downsampled to 98KHz at the DAC, so you literally cannot make it alias.

Ataraxic Iteritas is one that I have with a sync input that will alias at high sync rates.... but it's an inherently, intentionally glitchy and noisy oscillator in the first place, so that's more of a feature than a problem.

BTW, analog oscillators also have trouble with high frequency sync: they lose a lot of volume because the core never has time to get very far from 0V. The square output of my Synchrodyne will go completely silent at high sync rates because it never reaches the threshold for its comparator.

[S0lo]

I guess you're assuming an oscillator design that is bandlimited at the core, rather than oversampling and using an aliasing filter at the DAC. Which is not universal by any means.

Anything with wavetables (especially custom ones), phase modulation, waveshaping etc. is much more likely to oversample and use an antialiasing filter at the DAC. Sometimes to extremes -- Shapeshifter has an internal sample rate of 25MHz, downsampled to 98KHz at the DAC, so you literally cannot make it alias.

It is quite remarkable that Shapeshifter can internally run at 25Mhz!!. That is way, way higher than most of the digital modules I've seen. I've seen such things only done with FPGA and so. I'd say that this is not the norm.

Take Synthtech E370 for example. That is a capable wavetable VCO, runs at 96KHz. Same with Klavis twin waves MKII at 96Khz. I highly doubt they're running higher rates internally, if they did, docs would probably brag about it. Mutable plaits at 48Khz, docs explicitly state it's band limited. Endorphin.es Godspeed+ runs at 80Khz. And that one has sync in

But true, if you run in the >10Mhz, then you seldom need to worry about aliasing. In software if you oversample x32 at 44.1Khz. Thats 1.4Mhz, and you may still have aliasing if you don't bandlimit, specially with waveshaping. So I don't know what intellijel is doing if not FPGA.

BTW, analog oscillators also have trouble with high frequency sync: they lose a lot of volume because the core never has time to get very far from 0V. The square output of my Synchrodyne will go completely silent at high sync rates because it never reaches the threshold for its comparator.

True, but this happens with all sync be it digital or analog. The master has to always run at a lower frequency than the slave. Otherwise, the slave will never complete a cycle and you'd hear that as lowered volume

Good discussion

[foosnark]

Shapeshifter's using an FPGA, yeah. It also has a 55-bit phase accumulator, which... is just not something a sane developer would do on normal architecture But the result is SUPER clean FM.