Oscillators always oscillate?

[joeboy]

So I know lots of LFOs are free running, where they are always cycling through their waveform regardless of whether or not a note is being played. This behavior is easy to observe since the F of the O is so L, and the frequency is usually independent of any pitch that is played.

I think I've read before that V/DCOs behave in a similar manner. If this is the case, I'm curious about how this free-running behaves when pitched output is expected. When I depress the middle A key, the VCO's frequency is 440Hz, but what "fundamental" frequency is the oscillator running at when I take my finger off it? Without any sense for electronics, I'm guessing it's a much higher one that is whittled down into a musical frequency through the circuitry? Am I somewhat on target here?

[commodorejohn]

It depends entirely on the oscillator/synth design, but general practice in analog synths is to have a sample-and-hold circuit for the pitch CV which latches the correct value when a key is depressed and holds it after (so you don't get a sudden change in pitch during the release phase.) Nominally this means that the oscillators will continue at their last frequency until a new pitch CV is latched, although an analog S&H will eventually decay from the "correct" value.

[joeboy]

That's pretty cool. What's the oscillator doing before any keys are pressed though?

[Broadwave]

That's pretty cool. What's the oscillator doing before any keys are pressed though?

Most, if not all, analogue VCOs oscillate continually - Pressing a key only changes the control voltage to vary the pitch, the oscillator will stay at that frequency until a different key is pressed.

The same goes for filters and amps... you just reduce the input the voltage until you can't hear anything - nothing in the analogue signal path is actually "switched off".

[joeboy]

Sorry, I meant "what frequency is a VCO oscillating at before any key is pressed and a sample and hold circuit locks it into the last played pitch's frequency?" I'm guessing something like this varies between hardware, but is it generally higher/lower than the keys' pitches?

Also, when simulating a VCO digitally, is it important to keep track of this free-running oscillation? If I play a middle C, let go, and then press another key 5 seconds later, should that new note's audio begin at the wavetable index indicated by 261.6 * 5.0 % <wavetable length>, or is it safe to just start playing from the beginning of the wavetable again? Is there an audible difference?

[synthRodriguez]

Sorry, I meant "what frequency is a VCO oscillating at before any key is pressed and a sample and hold circuit locks it into the last played pitch's frequency?" I'm guessing something like this varies between hardware, but is it generally higher/lower than the keys' pitches?

Good question and again, it depends on the design. As before, most synth oscillators are always running in the background, but the amplifier's faucet is closed. The amplifier chip's ADSR settings command how fast it opens and and what levels to let out the sound upon a key closure.

I never investigated it too much, but the oscillators in the OB-Xa don't start running at all until a note is played, activating them. You need to play some notes and get them running for it to really warm up properly. The Yamaha CS-50 defaults to 2.111 VDC on startup which is around the middle of the 0.25 to 4.0 VDC range of the keyboard.

[madtheory]

I never investigated it too much, but the oscillators in the OB-Xa don't start running at all until a note is played, activating them. You need to play some notes and get them running for it to really warm up properly. The Yamaha CS-50 defaults to 2.111 VDC on startup which is around the middle of the 0.25 to 4.0 VDC range of the keyboard.

That is very cool info! We could hazard a guess that digitally controlled synths mostly work like the OB, whereas analogue controlled ones work like the Yamaha.

If I play a middle C, let go, and then press another key 5 seconds later, should that new note's audio begin at the wavetable index indicated by 261.6 * 5.0 % <wavetable length>, or is it safe to just start playing from the beginning of the wavetable again? Is there an audible difference?

It is audible. A lot of VAs simulate this. The Novation KS lets you control the relative oscillator phase, and start points. Think about it- the VCA is just gating the VCO, which is free running. So it can open the gate at any point in the waves cycle. The onset (or attack) of a sound gets more attention from the brain than any other part. This has been extensively tested, using sounds with attack only, decay and sustain removed. Humans can still tell what it is. Essential for fast reaction time to predators in the jungle

[joeboy]

Thanks for the answers. I have a much clearer idea of where to go in my programming now. I already have a variable that increments for every sample-writing cycle and "timestamps" a given component with the last cycle it was active, so it should be easy to calculate the phase of waves without actually iterating through their tables during moments of silence. And it seems that not "activating" the oscillators until the first note is played would be fine here, since programs don't need to warm up.

[meatballfulton]

In CV analog systems, a VCO has a base pitch set by the tuning controls. Once a non-zero CV is applied the pitch changes and remains the same until that CV changes. To deal with long EG releases, the VCO has to maintain that pitch so the CV source (keyboard, MIDI/CV converter, etc.) output has to remain constant until a new note is played. Every modular I have used works this way, as do CV controllable synths I've owned like the Arturia Brutes.

For a hardwired synth typically 0V represents the lowest key. When using MIDI/CV conversion it can mean pretty much anything. I've owned a couple of modulars and each one had a different way of setting the MIDI/CV to 0V. A common method is a power up mode where the first note-on sent over MIDI determines 0V. Note that 0V does not refer to a pitch! There is no standard for what pitch 0V should produce, unlike MIDI note numbers which are precisely defined (although affected by any transposition/tuning features of the synthesizer).

Some VCOs will respond to negative CVs, some will not.

Some MIDI/CV converters have limited ranges. Commonly that is at least five octaves (61 keys) and behavior outside that range may be any of the following:

-- gates do not fire
-- pitch is fixed to the lowest or highest note of the range
-- each octave outside the rnage repeats the last octave inside the range

The only MIDI/CV converter I have used that covered all 128 MIDI notes came with the pAiA 9700.

[pflosi]

Sorry, I meant "what frequency is a VCO oscillating at before any key is pressed and a sample and hold circuit locks it into the last played pitch's frequency?"

The frequency is determined by the sum of the manual offset controls ("coarse" and "fine" tune, usually) and all CV at any control inputs (that includes the pitch CV). Once any of those change, the pitch changes accordingly. It will stay on that pitch until any input changes.

There's usually no S/H circuit in CV keyboards, analog S/H is tricky to design without "droop" (unstable CV that gets lower and lower over time). The pitch CV is separated from the gate: The gate just opens the envelope, which opens a VCA, which makes the oscillator audible. The pitch CV is more like a variable DC "offset" that just stays at the respective offset ("key") until you change that (by pressing another key). Just like the tuning controls, but controlled from a keyboard sending stepped values instead of a pot.

I never investigated it too much, but the oscillators in the OB-Xa don't start running at all until a note is played, activating them.

I don't think that's the case, more likely it is oscillating at a very low frequency before any key is pressed.

Also, when simulating a VCO digitally, is it important to keep track of this free-running oscillation? If I play a middle C, let go, and then press another key 5 seconds later, should that new note's audio begin at the wavetable index indicated by 261.6 * 5.0 % <wavetable length>, or is it safe to just start playing from the beginning of the wavetable again? Is there an audible difference?

Yes, there is a difference, particularly on waveforms with less harmonics (sines, triangles). Well, at least you hear it better on them. To restart an analog osc on every key press, you patch the keyboard gate to the hard sync input of the VCO. That can be helpful for some patches to get rid of the irregular clicks. It will still continue to oscillate after you release the key, but get restarted (hard synced) on every note.

The same goes for filters and amps... you just reduce the input the voltage until you can't hear anything - nothing in the analogue signal path is actually "switched off".

Unless you use a switch

Cheers

EDIT: Looks like meatballfulton beat me to it

[JayEm]

It depends entirely on the oscillator/synth design, but general practice in analog synths is to have a sample-and-hold circuit for the pitch CV which latches the correct value when a key is depressed and holds it after (so you don't get a sudden change in pitch during the release phase.) Nominally this means that the oscillators will continue at their last frequency until a new pitch CV is latched, although an analog S&H will eventually decay from the "correct" value.

you can try this with an ms20.
patch the square output of the mg to the trig in and it'll start generating a very low tone when the env is triggered. press a key and it'll start playing at that pitch until you press another key.

i've used the monotribe sync out to act as a trig of sorts giving straight sixteenths, then just change key as needed.