Understanding VCOs, DCOs & Digital Oscillators

In the world of synthesizers, there are many opinions & misconceptions. The differences in oscillator designs are no exception. The oscillator is where the music path starts in a synthesizer. Its primary responsibility is getting waveforms onto the music palette so they can be adjusted, refined and modulated into the sounds and tones that you want.  To assist those new to synthesis and to clarify how different types of oscillators work, I built this page in hopes of giving the reader a general backgrounder.

As a preface, you may wish to visit my backgrounder on Analogue Vs Digital synthesizers. For the examples used on this page, please note that there are several different ways in which VCO & DCO circuits can be implemented. The underlying concepts presented here however are similar and will outline the important concepts.

The VCO

VCOs have been around for decades. If you are really interested in leaning the techie bits, Moritz Klein has a great video series on how they can be constructed. However, I am just going to try to keep it simple. VCO stands for Voltage Controlled Oscillator, which implies we are going to control an oscillator with a voltage source of some kind. For the oscillator examples here, Voltage Control (VC) means voltage will be used to control the pitch on an oscillator. For the musician, what is needed in a VCO is to create various waveforms that:

  1. When you press a musical note (key) on a piano type keyboard, the frequency of the oscillator (pitch) will change to match the theoretical frequency of a musical note to be played. Specifically, if I press the A key, on the second octave (A2), I should get a waveform that is exactly 110 cycles per second or 110 Htz.
  2. The VCO should also be able to scale within an octave. There are 12 semitones in an octave, and each key in a scale should output the appropriate frequency,
  3. The VCO should scale appropriately over octaves. When I press A3, the frequency should double to 220 Htz, when I press A4, I should get 440 Htz. This means the frequency should double with each successive octave.

Notice that a piano keyboard, generally has 5 to 9 octaves of 12 semitones each. This implies then, we can use a frequency ratio of one to the twelfth root of two ( 12√2 or 21/12). In simpler terms, means that given any semitone, we can then calculate the frequency of the following semitone when we multiply the first notes frequency by 1.05946. Early synthesizer designers like Robert Moog, developed many of the concepts for this idea.

The implementation idea: For a voltage source, we can use a piano keyboard, and incorporates an analogue voltage divider that adds 1/12 (0.0833) volt for each successive key pressed. Example: Press the lowest (left most) key, let’s call it C0, and it will output 0.0833 volts, hit the next key and it outputs 0.1666 volts, hit the next key and you are at 0.2499 volts, and so on. When you complete 12 notes, or one octave, and hit the C1 key, you will be at 1.083 Volts

The musical scale as it relates to voltages

And that, is the basic idea of controlling a VCO in a synthesizer. It is an analogue design , that uses 3 major components:

  • A Control Voltage (CV) source
  • A circuit of converting that voltage into a raw waveform that is musically usable
  • A final circuit that builds other waveforms that match the behaviour of the raw version

Early synthesizers were built from individual or ‘discrete’ electronic components. These components have a tendency to drift in value due to temperature, age or simply bring made with slight imperfections and loose tolerances. To compensate, VCOs of this nature require time to reach a stable operating temperature and they also have several variable resisters called trimmers, which assist in calibrating the VCO so the raw waveform will be as close as possible, to the desired frequency. But the originating voltages in the keyboard also have drift, which can also effect the oscillator. In short, analogue voltages are never exact and therefore frequencies can drift between semitones. Frequencies can drift so much that many older designs are completely out of tune after 3 or 4 octaves.

You can hear this drift in VCOs. Imperfect or fluctuating voltages can provide a slight modulation to the waveform. In some cases, minor variances in drift give a pleasant width or warmth to the sound. Other times, VCO drift can be a pain in the butt, as it may require constant tuning. Scaling over octaves could also be limited to 2 or 3. This is especially noticeable when multiple VCOs are used in a synthesizer. It is often the case where one VCO will be out of tune before another. This can limit the octave range of a synthesizer before it requires retuning as well as be very annoying to a musician.

To solve this drift and tuning issues, engineers turned to digital devices to assist them. By taking the control voltage from the keyboard (see diagram below), and converting them into discrete numerical values, adjustments can be made to correct any keyboard drift, and output a highly accurate frequency or clock signal.  This clock signal will change to the appropriate frequency whenever a key is pressed (voltage is input), from the keyboard. This method is so accurate that initial tuning is not effected by temperature and scaling can easily be maintained over 7 to 9 octaves.

DCO

A Digitally Controlled Oscillator (DCO), takes can take the same source CV but converted into discrete numerical information that a say a microprocessor will use to create a highly stable clock signal. Then the analogue portion of circuit takes that clock signal as an input, converts it into a frequency source and further converts it into a raw analogue sawtooth waveform. Just like the VCO, the raw waveform is musically usable so the same components to build other waveforms that match the behaviour of the raw version, can be used.

DCO’s can be implemented in many ways. Microprocessors are only one method. Using timer and divider integrated circuits, clocking signals can be produced that can provide the required input signal. The EDP Wasp is a great example. Other manufacturers like Roland, combined Intel Programmable interval timers and their own custom wave generator IC (MC5534).

Takeaway #1: DCOs are not digital VCOs, nor are the final waveforms digitally created. All of the waveforms used musically are analogue in nature. We have just replaced an analog circuit that exponentially integrates aa analogue voltage source, with a circuit that integrates a frequency. Both raw waveforms are generated though analogue components. The only difference sound-wise, would be an absence of drift in the source CV.

Differences between VCOs & DCOs

Is there a sonic difference between VCOs and DCOs? Yes and no. The only real difference between the DCO and VCO designs depicted above is the lack of drift in the raw wave. Both designs will still have similar drift occurring in the circuits that generate ancillary waveform (they are still discrete analogue circuits). The DCO however, will have much less drift in the raw waveform due to the accuracy of the clocking input.

I have heard some people state that “DCOs sound too weak or too thin”, which is utter nonsense. We don’t listen to a VCO as a final result musically. Much of what we hear is the function of the filter settings and amplifier. As a backgrounder, just like VCOs, many Voltage Controlled Filters (VCFs) require the same exponential integration scheme that converts the control voltages into a raw waveform. This is required so a VCF can accurately track the VCO as frequency increases or decreases. Furthermore just like the digital control in a DCO, associated VCFs can also utilize the same digital control mechanism to provide this tracking. So, there can be a sonic difference in the final result as the lack of drift can effect harmonics etc. But many newer designs allow “drift adjustment” or “Slop” for adding small elements of randomness to the clock signal. DCOs though are extremely stable, with far less calibration and temperature concerns, especially when multiple oscillators are incorporated.

Digital Oscillator.

In a digital oscillator, all of the waveforms are generated through the firmware of the microprocessor or Digital Signal Processor. Much like VCOs & DCOs, digital oscillators generally have an input to control the pitch or frequency of the waveform through a controller or CV/MIDI source. Digital oscillators can be still used with analogue VCFs, VCAs and envelopes.

Digital oscillators can output all of the standard waveforms including sine, square, triangle and sawtooth. They can also provide a wide range of radically different wave shapes with multiple levels of harmonics. Wave shapes can even be derived from audio sampling, cross modulation, linear FM or just about anything you could imagine. These oscillators are highly stable and maintain tuning across any number of octaves. They also provide the ability to incorporate microtones  on the fly. Being very popular to sound scape designers, there are several monophonic, and polyphonic synthesizers (and many Eurorack modules) that feature digital oscillators.

there are several use cases for using a VCO with its warm sounds filled with harmonics. Choosing a DCO is beneficial for for keeping multiple voices in tight harmony when creating lush pads and strings. VCOs are sharp and distinct with amazing bass, but can start to sound muddy and lack clarity when combining multiple voices. Digital oscillators have brought a wealth of new sound scapes, through the use of FM synthesis, sampling and vector synthesis.

I hope this has helped. In the end, it is about creating music not getting stuck in the weeds by focusing on the minutia or obscure. Becoming an  “analogue only” purest or  maintaining a “digital rules all” mindset will limit you as an artist. You want as many tools as you can use.