MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Sub Osc and DCO questions
MUFF WIGGLER Forum Index -> Modular Synth General Discussion  
Author Sub Osc and DCO questions
raccoonboy
Hi all

Regarding sub-oscs which have a fixed phase to the master osc (most hardware suboscs using comparetors).

Does this pretty much give it the same sound as a DCO?

A big complaint about analogue DCOs is that the oscillators, as they are digital clock controlled and not voltage controlled, they may sound thin, icy, even muddy (I love my Matrix1000). The reason for this I've heard is that multiple oscs won't go out of sync from eachother.

Surely a sub osc (with VCOs) is the same? Unless because their start positions are the same i.e. 0 degree then this is okay. Same goes for hard sync I guess. Will a DCO even sound thin (to some) in isolation even without a 2nd DCO? If so, why?

Does the issue only arise with DCOs when the two are phase locked but start out of phase? I.e. one at 0deg and another at 30? Is that even a thing?


Also. When folk describe muddyness of sub oscs. Surely this is due to the square wave. Surely a sine wave sub wud be perfect. Or do you prefer the sound of a sine wave out of sync from the main osc? I've heard folk argue a subosc of a pure sine is the best.

Opinions?

Thanks
Dave Peck
1. DCOs can easily be 'out of sync' with each other, you just detune one of them slightly. Same as other oscs.

2. Sub osc signals can be generated several different ways. One way is to run the main osc signal through a S/H and flip flop logic to create a square wave one octave below the main osc (for every cycle of the main osc, the sub circuit changes state between positive & negative, creating a square wave at half the frequency of the main osc). But that's not the only way to do it. In any case, assuming the sub osc signal is derived directly from the main osc in some way, it will always be phase locked to the main osc - you can't detune it from the main osc. You COULD run it through an inverter to invert the phase/polarity relative to the main osc, but that typically doesn't accomplish anything since they are an octave apart.

3. A square wave is far more bright and buzzy than a sine wave. So a square sub osc will be much brighter than a sub sine osc, not more muddy. Which one is preferable or 'better' is a matter of personal taste and what sounds 'best' in a particular patch.
raccoonboy
Dave Peck wrote:
1. DCOs can easily be 'out of sync' with each other, you just detune one of them slightly. Same as other oscs.

2. Sub osc signals can be generated several different ways. One way is to run the main osc signal through a S/H and flip flop logic to create a square wave one octave below the main osc (for every cycle of the main osc, the sub circuit changes state between positive & negative, creating a square wave at half the frequency of the main osc). But that's not the only way to do it. In any case, assuming the sub osc signal is derived directly from the main osc in some way, it will always be phase locked to the main osc - you can't detune it from the main osc. You COULD run it through an inverter to invert the phase/polarity relative to the main osc, but that typically doesn't accomplish anything since they are an octave apart.

3. A square wave is far more bright and buzzy than a sine wave. So a square sub osc will be much brighter than a sub sine osc, not more muddy. Which one is preferable or 'better' is a matter of personal taste and what sounds 'best' in a particular patch.


Thanks.. I'm more looking for opinions though on the sounds, rather than the technicalities as I understand the principles.

If the DCOs are NOT detuned, i.e. tuned via software, does this cause any issues if they are not phase locked? I mean, it all sounds good to me, I just heard a lot of folk complaining about their sound and forget the reasoning.

For example, I know a square wave and how it sounds compared with a sine. I'm just wondering why some folk think suboscs are too muddy, is it because of the phase lock or just because it's a square. Is it less muddy if it's phase locked? Same goes for sine. Do you prefer locked?

I'm more interested in the principles of phase cancelation etc, as it's something I'm a bit less familiar with.
Dave Peck
Two oscillators that are at exactly the same frequency can either be phase-synchronized or NOT phase synchronized. You can force them to be in phase by using hard sync. If they are NOT in phase, the sound produced will vary depending on exactly what their static phase relationship is (i.e. out of phase by five degrees, or by 95 degrees, or by 180) and also what the waveforms are and whether both oscs are producing the same or different waveforms.

And all of this is true regardless if they are VCOs, DCOs, DSP generated, or something else. None of this is specific to DCOs (although pairs of analog oscs rarely produce exactly the same pitch without using hard sync).

Personally, I have not heard about this complaint of DCOs sounding muddy. I have heard some people say they prefer to have more of the 'motion' and unpredictability in the tuning relationships of oscillators (analog, typically) when the amount of detuning constantly changes slightly over time and across different voices in a polysynth, and that simply setting the 'detune' parameter on DCOs doesn't give the same behavior as typical analog oscillators. This is often true, the DCOs will typically maintain the same amount of detuning over time and across multiple voices. But there are ways to create that same randomized detuning behavior in DCOs using additional pitch CV modulations.
Blairio
If you have a modular setup, then an easy way to derive a sub oscillator is to patch an oscillator through a clock divider. It works best with simple waveforms, such as sine or square wave. Most euro VCOs with sub outs offer a square wave sub - perhaps because the sound retains its definition better, and you need less level to hear square wave sub frequencies in a mix. My Makenoise STO has a great square wave sub.

Equally, my WMD SSF Spectrum offers both sine and square subs, either 1 or 2 octaves down.

Neither of the Spectrum's sub waveforms sound muddy, but then again it is a very clean sounding oscillator.

Perhaps the muddy-ness folk hear from some oscillator sub outs is because the oscillator itself does not sound that crisp or clear in the first place?

Each to their own. Not everyone wants clean / clinical sounding oscillators!
dubonaire
Blairio wrote:
Most euro VCOs with sub outs offer a square wave sub - perhaps because the sound retains its definition better, and you need less level to hear square wave sub frequencies in a mix. My Makenoise STO has a great square wave sub.


The reason is it's easy to do. You just feed the oscillator output to a bistable flip-flop circuit.
Blairio
dubonaire wrote:
Blairio wrote:
Most euro VCOs with sub outs offer a square wave sub - perhaps because the sound retains its definition better, and you need less level to hear square wave sub frequencies in a mix. My Makenoise STO has a great square wave sub.


The reason is it's easy to do. You just feed the oscillator output to a bistable flip-flop circuit.


Ah, I see, thanks . My experiments with using sine wave subs in tracks have not yielded great results - square waves sound better. But, I guess I could take a sine wave sub and run it through a wave folder and come up with a range of sub 'timbres', with a bit more 'bite' than a straight sine. That'll be tonight's homework.

<Blair edit> ...and with a bit of modulation of the wave folder's Fold & Symmetry parameters, this could make for a pretty interesting sub.
dubonaire
Blairio wrote:
dubonaire wrote:
Blairio wrote:
Most euro VCOs with sub outs offer a square wave sub - perhaps because the sound retains its definition better, and you need less level to hear square wave sub frequencies in a mix. My Makenoise STO has a great square wave sub.


The reason is it's easy to do. You just feed the oscillator output to a bistable flip-flop circuit.


Ah, I see, thanks . My experiments with using sine wave subs in tracks have not yielded great results - square waves sound better. But, I guess I could take a sine wave sub and run it through a wave folder and come up with a range of sub 'timbres', with a bit more 'bite' than a straight sine. That'll be tonight's homework.

<Blair edit> ...and with a bit of modulation of the wave folder's Fold & Symmetry parameters, this could make for a pretty interesting sub.


One thing I have done is run the saw, the double saw and the sub out from the Rubicon into a Triatt which gives a very rich sound - awesome for filtering.

Low freq sines are good to add to kicks as well.
mskala
Are there any modules with sub-oscillators that produce sine waves? Normally, a sub-oscillator works by putting the main oscillator signal through a digital flip-flop to divide the frequency down, and that naturally produces a square wave. About the only ways I can think of to do frequency division with a sine output would be to try to shape the divided-down waveform, either by filtering it or by locking a PLL onto it; or (in the case of a digital oscillator) to just generate it directly as an additional output that really is separate in the firmware and just happens to be at the right frequency.

From a musical perspective it does make sense that you would want a sine-wave "sub-oscillator," because that way its harmonics won't interfere with the harmonics of the main oscillator. That may relate to the perception of "muddiness" - a non-sine sub-oscillator will screw up the harmonics of the main oscillator signal when you mix them. For something like a bell, piano, or pipe organ where a lot of the sound is in harmonics but there's also a subharmonic partial in the bass, it might make sense to supply that subharmonic by throwing in a sine wave from somewhere. But as a practical matter it may be best to supply that low sine wave with a separate oscillator tuned appropriately, rather than by trying to derive it as a "sub-oscillator" signal.
luketeaford
mskala wrote:
Are there any modules with sub-oscillators that produce sine waves? Normally, a sub-oscillator works by putting the main oscillator signal through a digital flip-flop to divide the frequency down, and that naturally produces a square wave. [...]

From a musical perspective it does make sense that you would want a sine-wave "sub-oscillator," because that way its harmonics won't interfere with the harmonics of the main oscillator.


Can you explain how a flip-flop works in this way? There is a flip-flop patch in the Make Noise Maths manual and I tend to use it in a clumsy way as a switch (the EOR/EOC outputs open or close VCAs...) When I use Maths to create a subharmonic, I do it by simply patching into cycle input, increasing the cycle times, and using EOR/EOC output as the sub osc. I'm wondering if the patch I'm describing as a flip flop (the incoming wave handles resetting itself unlike the one in the manual) or if there are other applications of a flip-flop patch that goes over the heads of non-engineers like myself.

I agree that a sub harmonic sine is nicer, but earlier this week I was experimenting with using an octave down sawtooth through a fixed filter bank. One of the interesting things about that patch is it can kind of sound like it's changing the timbre by the way it adds harmonics to the main oscillator but in somewhat unconventional parts of the spectrum.
fitzgreyve
mskala wrote:
Are there any modules with sub-oscillators that produce sine waves?


Yes, Freq. Central Waverider (using Electric Druid VCDO chip)
"sub"-osc can be -2, -1, 0 or +1 octaves (octave above)
mskala
There are several different kinds of flip-flops and they don't all do the same thing, but the defining feature is that any kind of flip-flop has two states and will remain in either state indefinitely until some external input causes it to switch. For that reason a flip-flop is basically the same thing as a "bistable multivibrator," as opposed to an astable or monostable multivibrator which is at least some of the time switches states on its own without external input. The differences among flip-flops have to do with how the external control signals operate.

For frequency division you'd use the kind of flip-flop that toggles between its two states each time it receives an external signal. In digital logic terms this would be a "T flip-flop," which can also be simulated by a "D flip-flop" with its inverting output fed back into the D input, or by a "JK flip-flop" with both J and K held high. If you feed a signal of 1000 Hz into the clock input of such a circuit, it changes state 1000 times per second, which means it goes through a complete cycle of turning on and then back off (two toggles) 500 times per second. The output is 500 Hz, half the frequency of the input.

But because it's switching between discrete states, with exactly the same time between the two halves (assuming the input was frequency-stable itself), the output has to be a square wave and has to be a perfectly 50% duty cycle square wave, not some other kind of rectangle or pulse.

The flip-flop patch in the Maths manual is a little different because it has separate "set" and "reset" inputs and no separate clock - it's more like what in digital logic would be called an "SR flip-flop." To make one of these work for frequency division would require some additional logic and filtering to generate separate triggers for "set" when the input goes high while the output is low, and "reset" when the input goes high while the output is high.
mskala
fitzgreyve wrote:
mskala wrote:
Are there any modules with sub-oscillators that produce sine waves?


Yes, Freq. Central Waverider (using Electric Druid VCDO chip)
"sub"-osc can be -2, -1, 0 or +1 octaves (octave above)


Looks like that's a built-in feature of the chip's firmware, so other modules using this chip would also be able to do it.
raccoonboy
mskala wrote:
Are there any modules with sub-oscillators that produce sine waves?


I'm trying to find a square wave to triangle converter. It's apparently easy to do for one frequency, but maybe not for various frequencies.

Because then I could use .COM for square subs and I'd also have the option of triangle, saw and sine if I combined with https://www.modulargrid.net/e/bubblesound-instruments-cvws and then filter to taste!
mskala
raccoonboy wrote:
mskala wrote:
Are there any modules with sub-oscillators that produce sine waves?


I'm trying to find a square wave to triangle converter. It's apparently easy to do for one frequency, but maybe not for various frequencies.


True, and it's easy to understand why... the square wave only switches between two voltages. When it's high, there is no way to tell from the voltage where it is in its cycle. So the converter can't just look at the voltage, determine where it is in the cycle, and then send the appropriate voltage to the output to represent that same point in a triangle (or sine) cycle. The only way a square-to-triangle converter can possibly work is by measuring the time since the last transition of the square wave, and then it needs to know the frequency of the desired output, either from an external adjustment or by measuring the times over at least one full cycle (which is what a PLL does). All this is much more complicated than converting in the other direction, which can be done just by looking at the voltages without reference to timing.
raccoonboy
Just found this from an old post on here:

I kinda remember reading about this but forgot!
You can get a triangle from a SQUARE SUB..

IF and only IF the SQUARE SUB is a SUB for a SAW wave.



"If you have a ring modulator and a square sub-octave generator, you can get a triangle sub-octave.
Patch the VCO saw into the ring mod X and the sub-octave square into the ring mod Y,
and then the ring mod output will be a triangle wave one octave below the VCO frequency."

Then you could also get the saw sub and sine sub. Which is kinda sweet! I use a lot of SAWs, so that's all good.

credit to Stereotactixxx
MUFF WIGGLER Forum Index -> Modular Synth General Discussion  
Page 1 of 1
Powered by phpBB © phpBB Group