What defines a digital or analogue circuit?

Discussions about anything analog, digital, MIDI, synth technology, techniques, theories and more.
User avatar
polardark
Active Member
Active Member
Posts: 357
Joined: Sat Sep 18, 2004 7:03 pm
Location: Stockholm, Sweden

Re: The VCO vs DCO thread

Post by polardark » Wed Jun 20, 2012 10:32 am

GuyaGuy wrote:Apart from defining DCO vs VCO I think it would be interesting to discuss what scientific properties make VCOs sound the way they do in comparison with DCOs--again not subjectively in terms of preference it rather objectively in terms of measurable information.
The jitter of a phase accumulator oscillator generating pulses should be a real measurable difference. The extent of the jitter would vary depending on the specifications of the phase accumulator circuitry.

User avatar
Stab Frenzy
Moderator
Moderator
Posts: 9723
Joined: Tue Jun 06, 2006 5:41 pm
Gear: Eurorack, RYTM, Ultranova, many FX
Location: monster island*
Contact:

Re: The VCO vs DCO thread

Post by Stab Frenzy » Wed Jun 20, 2012 11:50 am

nathanscribe wrote:The Crumar Bits (and Elka Synthex) build waves out of layered multiple-octave square waves (the Synthex uses more, so greater resolution) to form stepped sawtooths, which are then shaped by the usual analogue wave-shapers.

[snip]

Regarding summed squares to make stepped sawtooths which are then shaped and filtered by analogue means, the Korg Lambda does the same. I think the Korg PS-series employ it too, but am not entirely sure.

The Korg Poly 800 is similar in that it provides the user with four octaved squares that can then be combined to give, in total, a rough stepped saw as above, or odd combinations that don't match any common wave.

At the point, you're not talking about making waves so much as tailoring harmonics I suppose.
You know what this reminds me of? The way DACs work. You basically have a bunch of digital output pins of hooked up to transistors switching a voltage of an amount representative of the significance of the bit in the signal, so if you had a 4 bit output you have bit one switching 0.5V in or out, bit two 0.25V, bit three 0.125V, bit four 0.0625V. It's a lot like the digital microcontroller sending data to the CMOS logic in the P800 to make the waveform.

User avatar
meatballfulton
Moderator
Moderator
Posts: 5908
Joined: Wed Apr 13, 2005 9:29 pm
Gear: Logic Pro X

Re: What defines a digital or analogue circuit?

Post by meatballfulton » Wed Jun 20, 2012 12:52 pm

Calling the Poly 800 oscillator as DCO is a misnomer and leads to confusion. It's just a digital oscillator. Many other hybrids of that era like the Korg DWs, Ensoniq ESQ-1 and SQ-80, etc. use the term DCO but the oscs are 10% digital. The Roland Juno and JX as well as the Evolver DCOs are true analog oscs under digital control.

The Juno and Evolver oscillators are analog but retriggered by a digital timer. It's a very similar concept to using hard sync to keep two VCOs locked in tune.

The only thing that happens with a true DCO is that the tuning is kept stable by a digital reference. No more warm up time, no more needing to retune in the middle of a song, no more out of tune chords, etc.

As far as a digital circuit, to me as an electrical engineer it's cut and dried. As mentioned before, if the inputs and outputs of the circuit are digital it's a digital circuit.
I listened to Hatfield and the North at Rainbow. They were very wonderful and they made my heart a prisoner.

User avatar
nathanscribe
VSE Review Contributor
VSE Review Contributor
Posts: 2889
Joined: Fri Aug 11, 2006 1:03 pm
Location: The right side of the Pennines
Contact:

Re: The VCO vs DCO thread

Post by nathanscribe » Wed Jun 20, 2012 1:06 pm

Stab Frenzy wrote:
nathanscribe wrote:The Crumar Bits (and Elka Synthex) build waves out of layered multiple-octave square waves (the Synthex uses more, so greater resolution) to form stepped sawtooths, which are then shaped by the usual analogue wave-shapers.

[snip]
You know what this reminds me of? The way DACs work. You basically have a bunch of digital output pins of hooked up to transistors switching a voltage of an amount representative of the significance of the bit in the signal, so if you had a 4 bit output you have bit one switching 0.5V in or out, bit two 0.25V, bit three 0.125V, bit four 0.0625V. It's a lot like the digital microcontroller sending data to the CMOS logic in the P800 to make the waveform.
I'd not thought of it like that, but now you mention it... I suppose the main difference, results-wise, is that the squares in the above synths are cumulative (that is, the base of the step has to be there to lay another on top), whereas a DAC can output any bit as zero.

You getting ideas?

User avatar
samuraipizzacat29
Supporting Member!
Supporting Member!
Posts: 819
Joined: Mon Jan 31, 2011 6:40 pm
Real name: Nate

Re: What defines a digital or analogue circuit?

Post by samuraipizzacat29 » Wed Jun 20, 2012 2:31 pm

Can you guys please link to something that agrees with you? Listening to some guy rant on a forum is not a valid authority. Simply saying "this is the way it is" does not qualify as a legitimate source of truth. So far, i have supported everything ive said with real life examples.

A circuit is analog at the point where the signal output is the signal that reaches our ears. I thought that was a rudimentary fact, but it seems clear to me some would disagree with that. A digital signal is considered so because it contains a "picture" of what youre trying to describe, but that picture is contained in the 1s and 0s (a pulse wave on the electrical level) that is then decoded into the analog wave.

You used the example of an arduino, but you failed to point out that all of the inputs and outputs of the atmega controllers they ggenerally use are coupled to a/d or d/a converters. Look up a datasheet for an atmega328 controller and then look at the code for what you described. Youll see that theres DEFINITELY d/a conversion.
You can't use a microprocessor to create an oscillator by itself because the outputs would be more like triggers at that frequency.

Cmos ics are widely regarded to provide analog output. These ics are used in a wide array of machines whose output is considered analog (the wavefreaker I just made, the arp axle vco, the cgs01 and other cgs01 projects). Yet, they can be used to process digital signals as well.

I think it can be best summed up by saying that if the core output needs to be decoded to be intelligible, then I'd consider it digital. If the raw output is what is intended to hit our ears, then I think we'd all agree that at least at that point, the signal is analog. So the only thing that matters to us is how that signal is created.

In the case of the poly800, as described in the datasheet, it is a microprocessor that feeds a frequency counter that is input into a logic device. This is a textbook DCO. Whether you think a Cmos Ic is inherently analog or inherently digital is up to you, but it's output can be decidedly analog or digital depending on the application of the circuit.

User avatar
nathanscribe
VSE Review Contributor
VSE Review Contributor
Posts: 2889
Joined: Fri Aug 11, 2006 1:03 pm
Location: The right side of the Pennines
Contact:

Re: What defines a digital or analogue circuit?

Post by nathanscribe » Wed Jun 20, 2012 2:43 pm

If you want books, try Horowitz & Hill, The Art Of Electronics; or Don Lancaster, The CMOS Cookbook.

Digital is about logic. Analogue is about representing continuously. Neither is 'true' in that neither is necessarily the actual thing you want; both are, as you put it, a 'picture', but one is comprised of logic states, the other of a continuously variable quantity. Logic states may then be used to determine variable quantities, but they will be defined by discrete steps (1, 2, 3) rather than indeterminately smooth fluctuation.

There's a reason CMOS logic ICs are called a "Digital Logic Family" and not "Analogue Logic Family". The clue is in the word "logic".

CMOS is a manufacturing process. It stands for Complementary Metal Oxide Semiconductor. Not all CMOS devices are logic ICs. Not all logic devices are CMOS.
Last edited by nathanscribe on Wed Jun 20, 2012 3:31 pm, edited 1 time in total.

User avatar
samuraipizzacat29
Supporting Member!
Supporting Member!
Posts: 819
Joined: Mon Jan 31, 2011 6:40 pm
Real name: Nate

Re: What defines a digital or analogue circuit?

Post by samuraipizzacat29 » Wed Jun 20, 2012 2:47 pm

And the way the dcos are made in the example on the first page of this compares the saw waves but ignores the fact that the pulse waves are made exactly the same way. The saw waves are made much "better" by the Juno in that they're waveshaped and not made by 4 octaves of pulses (lame). But the oscillator itself is produced in exactly the same way. Again, this was explained in both of the datasheets I linked to. This is what makes a DCO a DCO. I don't understand why were saying a poly 800 is different, when it's clearly not.

User avatar
samuraipizzacat29
Supporting Member!
Supporting Member!
Posts: 819
Joined: Mon Jan 31, 2011 6:40 pm
Real name: Nate

Re: What defines a digital or analogue circuit?

Post by samuraipizzacat29 » Wed Jun 20, 2012 2:57 pm

nathanscribe wrote:If you want books, try Horowitz & Hill, The Art Of Electronics; or Don Lancaster, The CMOS Cookbook.

Digital is about logic. Analogue is about representing continuously. Neither is 'true' in that neither is necessarily the actual thing you want; both are, as you put it, a 'picture', but one is comprised of logic states, the other of a continuously variable quantity. Logic states may then be used to determine variable quantities, but they will be defined by discrete steps (1, 2, 3) rather than indeterminately smooth fluctuation.

There's a reason CMOS logic ICs are called a "Digital Logic Family" and not "Analogue Logic Family". The clue is in the word "logic".

CMOS is a manufacturing process. It stands for Complimentary Metal Oxide Semiconductor. Not all CMOS devices are logic ICs. Not all logic devices are CMOS.
Yes, BUT, all pulse waves are also streams of logic. This is what I said straight away. The way a Cmos logic Ic makes it's output square wave is exactly the same way a discrete transistor would. Again, stated previously. So to say that the output is "digital" because the chip can be used in digital systems is highly confusing to me.

If you're saying that a Cmos output is digital because it's hard ons/offs, then how is any transistors output any different? That is how we've made pulses from days of yore, and they've always been considered analog. Whether they can be used for logic is besides the point. Unless every pulse wave is digital..... I'm just trying to compare like things here.

User avatar
nathanscribe
VSE Review Contributor
VSE Review Contributor
Posts: 2889
Joined: Fri Aug 11, 2006 1:03 pm
Location: The right side of the Pennines
Contact:

Re: What defines a digital or analogue circuit?

Post by nathanscribe » Wed Jun 20, 2012 2:58 pm

samuraipizzacat29 wrote:You can't use a microprocessor to create an oscillator by itself because the outputs would be more like triggers at that frequency.
Yes you can. It just won't have an analogue output. An Oscillator is just something that cycles in value of some kind - voltage, number, whatever. A 'virtual' oscillator is still an oscillator, no? Otherwise your softsynth would not be a softsynth at all if the audio wasn't hooked up. If you generate a cycling value in software and view the results on your computer screen, does that make it any less of an oscillator? Your eyes see a waveform on screen, but all it is is dots. Your mind may interpret it as a sawtooth, but it isn't.
Cmos ics are widely regarded to provide analog output. These ics are used in a wide array of machines whose output is considered analog (the wavefreaker I just made, the arp axle vco, the cgs01 and other cgs01 projects). Yet, they can be used to process digital signals as well.
Which ICs? Where's the CMOS in the ARP Axxe VCO (if that's what you meant to type)? More to the point, where's the LOGIC in the ARP VCO?

Don't confuse Logic with CMOS with Digital. Etc.

There are units that utilise properties of CMOS logic ICs that push them into service in regular analogue circuits - such as the MXR Envelope Filter, which uses (IIRC) a 4069 in the middle, narrow part of its operation where the logic state is undefined - but that's a property of the device, whereas digital electronics is not inherently a property of the device - digital electronics is a mode of employment of a device that it is intended to serve.
Last edited by nathanscribe on Wed Jun 20, 2012 3:34 pm, edited 1 time in total.

User avatar
nathanscribe
VSE Review Contributor
VSE Review Contributor
Posts: 2889
Joined: Fri Aug 11, 2006 1:03 pm
Location: The right side of the Pennines
Contact:

Re: What defines a digital or analogue circuit?

Post by nathanscribe » Wed Jun 20, 2012 3:04 pm

samuraipizzacat29 wrote:And the way the dcos are made in the example on the first page of this compares the saw waves but ignores the fact that the pulse waves are made exactly the same way. The saw waves are made much "better" by the Juno in that they're waveshaped and not made by 4 octaves of pulses (lame). But the oscillator itself is produced in exactly the same way.
No it isn't.

The Juno generates a CV to pitch the oscillator, by which I mean the CV feeds an analogue integrator which is reset - but in a VCO, the reset is done by a comparator (small circuit that senses it has reached a level and puts out a trigger) - on the Juno, the reset is done by a clock that counts how long the wavecycle should be for any given pitch and sends out a reset trigger at that interval.

The Crumar uses clocks to drive counters that put out simultaneous discrete on/off logic signals that appear at frequencies that double for each. They are summed at levels proportional to their significance to produce a staircase that is waveshaped in the analogue domain.

The two methods are quite different.

What happens to the sawtooth after generation is not relevant to this discussion. Both machines use analogue waveshaping.

As for the digital aspects of each of these prior to that point, both machines run some kind of program that determines operation based on input and the results of fixed processes, and that can only be digital; both machines use components in a hybrid fashion; the Crumar's generation of layered square waves is the analogue use of the end point of a digitally controlled set of devices; the Juno's timed integrator reset is the use of digitally controlled analogue circuit. Hence Digitally CONTROLLED Oscillator in both instances, I suppose.
Last edited by nathanscribe on Wed Jun 20, 2012 3:44 pm, edited 5 times in total.

User avatar
nathanscribe
VSE Review Contributor
VSE Review Contributor
Posts: 2889
Joined: Fri Aug 11, 2006 1:03 pm
Location: The right side of the Pennines
Contact:

Re: What defines a digital or analogue circuit?

Post by nathanscribe » Wed Jun 20, 2012 3:12 pm

samuraipizzacat29 wrote:Yes, BUT, all pulse waves are also streams of logic.

{cut}

If you're saying that a Cmos output is digital because it's hard ons/offs, then how is any transistors output any different?
No they're not, and no I'm not.

A pulse train is a pulse train. Whether it is digital or analogue depends on its implementation. Digital is a means of representing information encoded in sets of states of 1 or 0, but the definition of those states is arbitrary. Different logic families work differently - check out the differences between TTL and CMOS, for example. What defines a 1? What defines a 0? What lies between?

Generating a square wave is not the same thing as representing a square wave digitally. For example, take an analogue square wave. Low might be 0V, or it might be -10V. High might be 5V, or +10V. But the key thing is that any other value between or outside those might just as easily be part of the natural signal. That square wave could be derived from a sawtooth, for example, using the same voltage rails and no digital circuitry at all.

A digital square wave might be represented using the values 00000000 for low, and 11111111 for high in an 8-bit system. Or in decimal, 0 and 255. Duration would also have to be represented digitally, which may be done either as a string of values that appear at regular intervals (maybe 44.1kHz) or you might encode the duration as a digital value on itself - with each discrete bit step representing a millisecond, or something. That would depend on your chosen implementation.

The two are quite different.

I think you might be confusing the shape of a train of logic bits as they're represented by a voltage with the representation of analogue signals that may or may not be square/pulse in shape.

User avatar
samuraipizzacat29
Supporting Member!
Supporting Member!
Posts: 819
Joined: Mon Jan 31, 2011 6:40 pm
Real name: Nate

Re: What defines a digital or analogue circuit?

Post by samuraipizzacat29 » Wed Jun 20, 2012 3:44 pm

I'm not confused at all. To recap:

1) all ons and offs are made by transistors, valves, etc. Some big , some small, but always those. A Cmos Ic is still transistors.

2)all digital signals are represented in a pulse train.

3) a true pulse (square) wave is represented by an alternation of a low voltage and a high voltage. There are no points in between those values on a TRUE square wave.

All of the above is what we'd agree to be fact, correct?

User avatar
nathanscribe
VSE Review Contributor
VSE Review Contributor
Posts: 2889
Joined: Fri Aug 11, 2006 1:03 pm
Location: The right side of the Pennines
Contact:

Re: What defines a digital or analogue circuit?

Post by nathanscribe » Wed Jun 20, 2012 3:45 pm

Go on.

User avatar
samuraipizzacat29
Supporting Member!
Supporting Member!
Posts: 819
Joined: Mon Jan 31, 2011 6:40 pm
Real name: Nate

Re: What defines a digital or analogue circuit?

Post by samuraipizzacat29 » Wed Jun 20, 2012 3:51 pm

So then the only thing that separates (electrically) a pulse we'd consider to be digital from a pulse we'd consider to be analog is frequency and amplitude.

User avatar
nathanscribe
VSE Review Contributor
VSE Review Contributor
Posts: 2889
Joined: Fri Aug 11, 2006 1:03 pm
Location: The right side of the Pennines
Contact:

Re: What defines a digital or analogue circuit?

Post by nathanscribe » Wed Jun 20, 2012 3:58 pm

Not at all.

The difference is in its implementation.

A digital signal can be of any frequency, and of any amplitude.

An analogue signal can also be of any frequency and any amplitude.

Actually I'd consider your point no.2 above to be slightly off, as not all digital information is a pulse train, exactly. I'd say that digital signals are only concerned with being either a 1 or 0, yes or no, and the definition of those states in any system is arbitrary.

As Stabs pointed out right at the start of this thread, and I repeated later, digital is about using two states to represent information encoded in patterns of those two states, and those two states are all the possibilities. Analogue is about a continuously variable quantity; whether or not that analogue signal looks like a pulse is irrelevant, the point is that any quantity can be directly represented by any amount within the range permitted, and not encoded as multiple instances of 1 and 0.

The fact that digital signals have to be generated and transmitted and received and acted upon by physical components, which by nature is continuous and variable in its properties, means that there is going to be crossover when you talk about certain kinds of circuits. But there's a reason digital is called digital and analogue is called analogue. People have been doing this for half a century and more, I tend to accept they're right - after all, they invented the tech we're using to waste our time on such trivialities.
Last edited by nathanscribe on Wed Jun 20, 2012 4:10 pm, edited 2 times in total.

Locked