adding oscillators

Otto · Post by **Otto** » Tue Dec 21, 2010 12:57 am

Weird problem: If two RADIAS oscillators are added with the same waveform, frequency and amplitude (analog tune set to zero), each time a key is pressed it will sound differently. It might be the same sound a few times before changing, and some notes seem to show more variance than others - no way to predict. This seems to be caused by phase shifts between the two oscillators: Two sine waves (from two timbres - the sine of OSC2 is not very pure) result in anything from zero to double amplitude, two sawtooths result in anything from double frequency to double amplitude. This also applies to different wave forms, amplitudes, and frequencies (one an intger multiple of the other, e.g. an octave), but then it is less audible. Does anybody know how to disable or work around this not-so-useful feature? Or does anybody know why I should love it?

To recreate the problem: Choose an unsed user program with just one timbre being "TimbreInit1". Press "Edit/Yes", select the Pitch page, reduce Analog Tune from 5 to 0, and turn up the volume of OSC2 to 127 with the knob or in the Mixer page (OSC1 already is at 127, OSC1 and OSC2 are set to sawtooth). Play a note repeatedly. If you try it with two OSC1's of different timbres, don't forget to st analog tune to zero in the Pitch page of both timbres.

Best
Otto

Timo · Post by **Timo** » Tue Dec 21, 2010 1:02 am

http://www.korgforums.com/forum/phpBB2/ ... hp?t=36629

The effect is caused by the initial phase of an oscillator waveform when it's triggered. With the Radias, the initial phase of each oscillator is always random, hence why they cancel/boost each other randomly.

Otto · Post by **Otto** » Tue Dec 21, 2010 1:23 am

Ah yes - I see: at least I am not the only one to notice it. A pity they don't fix it.

Best again, Otto

tpantano · Post by **tpantano** » Tue Dec 21, 2010 2:06 am

Otto wrote:Ah yes - I see: at least I am not the only one to notice it. A pity they don't fix it.

Best again, Otto

Note to Korg, should they see this, i shall never ever ever buy a Korg again until this is fixed. Thank you and good night.

ultravvox · Post by **ultravvox** » Tue Dec 21, 2010 4:36 am

And what about old MS2000 and MicroKorg? I don't have them right now, just interesting, if they have problem with "phase memory" or not. They sound much more focused that R3 and Radias.

X-Trade · Post by **X-Trade** » Tue Dec 21, 2010 10:24 am

It is not considered a problem. In fact in many circles it is considered desirable because this is how most real analogs operate. The oscillators are still free-running whilst they are off. They do not 'stop'.

The MK and MS2K do this too.

If you do want the oscillators to stay in phase when set to the same tuning (but why would you want that if they are the same waveform?), you could try turning sync on.

Also the oscillator 1 detune mode forces that oscillator to start at a given phase. So you even have control over the attack characteristic. You can always set 'detune' at zero and you've got a simple note-synced oscillator.

ultravvox · Post by **ultravvox** » Tue Dec 21, 2010 5:17 pm

How analog oscillators could be "free-running" in terms of random/synced phase? They are "on" only when control voltage (hence a name "Voltage Coltrolled Oscillator") is applied. At which frequency should they oscillate when no note is pressed and no control voltage is applied? Check old synths schematics, keyboard is just array of resistors that are connected to VCO when you press a note.

I think it's more correctly to say that analog oscillators are INDEED synced (not hard-synced as in Sync mode on Radias!) like on Virus, as every time you press a note both oscillators start always from 0 degrees phase, as AFAIK analog oscillators just cannot operate otherwise.

X-Trade · Post by **X-Trade** » Tue Dec 21, 2010 5:27 pm

Well, in modular systems and in later digitally-controlled polysynths, the oscilaltors would continue to oscillate at the last pitch allocated to that voice, but the envelopes would be closed.

Similarly, on a hammond organ for example the tonewheels do not stop spinning when you are not playing any keys (well, because this is a physical system, if they did stop you would get a portamento effect from zero with every note-on, whilst the tonewheel speed ramps up). You get a different attack on every note.

Even if the control voltage is 'zero', the oscillator will still be oscillating just at a very low level, I don't believe they can enter a static mode.

Anyway, the only time this becomes a problem is when your oscillators are the same pitch, and in that case you can synchronise them without any other adverse effects because there is no pitch difference between the two.

I also found this which is an interesting read:
http://waveformless.blogspot.com/2010/1 ... lator.html

Not saying that it wouldn't be nice to have an option to simply turn it on or off for each oscillator though.

Otto · Post by **Otto** » Tue Dec 21, 2010 7:27 pm

Dear Korg,

ere is my wish for Christmas: I want the fundamental component of all tonal wave forms of all oscillators in the system running at the same frequency to be perfectly in phase - this includes PCM waves, of course.

Things probably get complicated if pitch is EG controlled. As I don't want varying sound in this case, too, a more realistic wish would be: The fundamentals of all tonal waveforms should start with the same phase, and they should start perfectly synchronised for each note-on-event. Independent of frequency.

Please.

meatballfulton · Post by **meatballfulton** » Wed Dec 22, 2010 12:44 am

ultravvox wrote:How analog oscillators could be "free-running" in terms of random/synced phase? They are "on" only when control voltage (hence a name "Voltage Coltrolled Oscillator") is applied

As X-Trade wrote, VCOs stay locked at their last pitch until the CV changes to indicate a new pitch. If the CV actually drops to zero between key presses, the oscillator simply drops to it's lowest pitch. The VCO itself is always on, it's the VCA envelope that controls whether any sound is output. On synths that let you keep the VCA open all the time, the pitch stays constant until you press a new key. This has been true for every CV analog I have ever owned.

Typically on vintage analog monosynths the lowest key on the board is set to send 0V to the VCO. Then scaling the keyboard is simply a matter of checking the voltages at every octave; on a V/oct synth each octave should generate a CV exactly 1V greater then the previous octave.

Modern monos with onboard MIDI-CV conversion usually "learn" the pitch by having the user play the lowest key on the board, this is then mapped to 0V. The actual pitch generated is controlled by tuning the VCO to the desired pitch.