analog/digital

[mantella]

hey, this question may be really obvious and may completely cement my super-noob status, but I am going to ask it anyways.

when is a circuit considered to be "completely analog"? What is the thing that makes it digital? The use of a transistor? The use of an IC?

[Cliff Schecht]

Analog signals are continuous, digital signals are discrete (i.e. individual points that represent a real signal). A circuit is considered analog if the signal remains continuous and doesn't chop the signal up into single points represented by 1's and 0's. Digital circuits will take an analog signal and chop it up into single points represented as discrete values that occur at a certain time. There are a lot of different circuits and tricks that are sort of a mix of analog and digital as well (see switched capacitor filters). I'm sure others will expand on this further but the first sentence I wrote is the most basic difference.

[JKowalski]

Analog means your signal is some value that varies over continuous time, for example a voltage oscillating as a sine wave over time. Analog circuits use components to change this signal in some way (still as a continuous time signal).

Digital means you convert your "information" (audio signal) into 1's and 0's (binary values) at some non-continuous rate (at a sample rate) and perform mathematical operations on it using some kind of processor.


Any circuit component is not necessarily "digital", but can be used in ways that deal with digital information. For example, you can use a bipolar transistor to transmit digital information or analog information. Technically, every component is "analog" in that it functions over continuous time... But you can use any continuous time component to transmit non-continuous time information.

A "digital" IC like a microprocessor is made up of a ton of transistors and semiconductors (that could otherwise be used in analog ways) arranged in a way that makes it suited to perform mathematical operations on binary numbers.


I consider a circuit completely "analog" if the signal is NEVER converted to digital information at any point. Many circuits may use digital components to do various tasks but still not convert the audio signal, so I do not consider that digital.

[chi_boy]

How does "Solid State" fit in?

Does it mean "no tubes"?  If so then can a solid state circuit be either analog or digital?

[petemoore]

  Analog is a 'pulse' message, taking a 1/2wave and looking at it, it has a shape, this shape usually varies compared to the next half wave [music, having dynamics means as the pulses go by the string decreases swing distance, this is reflected fairly exactly by electronic pulses of rising and falling voltage potentials.
 Digital means the analog pulse shapes [sizes] have be coded, we're talking large numbers.
 Larger numbers of samples [sampling rate] = smaller 'slices' are taken, hence they're taken more frequently, the numbers which represent the 'up's and 'overs', and 'down and overs' [what you do when trying to make a curve with straight lines]...if you get enough of them they will make a curve, here's a digital signal, make all lines equal length, start at the left, work right:
 Up up, up up over up up over, up Over Over Over Down over down down over, down down, down down.
 Do that with 2x the amount of coded directions, and you get better resolution, you can stand back farther and it begins to resemble a curve instead of steps.
 Should look like a 1/2 wave.
 The military did digital computing with tubes, that's one reason we have tubes, then transistors started computing, and...this and that happened, and here we are, online etc., the number of tubes would be a lot easier to calculate than the number of transistors, whether doing it in some analog fashion or digitally...hafta say digital using SS components would seem the logical option now.
 

[PRR]

Digital stages have exactly two states: ON and OFF.

These can be implemented with relays, tubes, transistors, even rocks. (About 900 years ago, the taxman used a "counting table".)

Analog stages can take "all" positions between ON and OFF. For audio we normally "bias" the stage to "halfway on", and then wiggle either way. Tubes and transistors are the obvious electronic analog devices, but there are also meters and carbon-amplifiers. A car's power steering is an analog hydraulic amplifier.

BTW: all "digital" stages are analog stages OVER-driven to the MAX. The CD4000 series CMOS is sold for digital logic applications, but can be finessed into its linear analog mode.

> when is a circuit considered to be "completely analog"?

When the Marketing Department says it will sell better if they put that on the box.

A 1956 Fender Champ or 1966 Peavey was, of necessity, pure-analog in the audio system. However the power switch is a "digital" function.

[BAARON]

How does "Solid State" fit in?

Does it mean "no tubes"?  If so then can a solid state circuit be either analog or digital?

Yes.  It literally means that you use solid components.  Tubes are hollow, and thus are not solid state.  http://en.wikipedia.org/wiki/Solid_state_(electronics)

[Lurco]

How does "Solid State" fit in?

Does it mean "no tubes"?  If so then can a solid state circuit be either analog or digital?

Yes.  It literally means that you use solid components.  Tubes are hollow, and thus are not solid state.  http://en.wikipedia.org/wiki/Solid_state_(electronics)

my glass-diodes are also hollow ?

[JKowalski]

How does "Solid State" fit in?

Does it mean "no tubes"?  If so then can a solid state circuit be either analog or digital?

Yes.  It literally means that you use solid components.  Tubes are hollow, and thus are not solid state.  http://en.wikipedia.org/wiki/Solid_state_(electronics)

my glass-diodes are also hollow ?

No, they have a "solid" chunk of semiconductor in the glass that is the diode.

Unless these are tube "glass diodes"

[Processaurus]

Analog delay is half digital, as it has a sample rate, and really crummy one at that!  A PT2399 based delay on the other hand, has a sample rate in the mHz as it uses sigma delta 1 bit ADA.  So there's more to understand than analog or digital... 

[Mark Hammer]

Until our instruments generate digital signals and speakers are stepper-motor-driven pistons, there will always be an element of analog at the input and output of any effect that is classified as "digital".  Indeed, a great many high-end digitalunits are perceived as high end or marketed as such, largely because of the quality of the conversion from analog to digital (and back again) within the unit.

So, there is no clearcut distinction between "analog" and "digital".  No, let me rephrase that.  If you were to draw a nice thick line between analog and digital it would be a crisp clear line on the analog side and smudged on the digital side.  It is obviously the case that a circuit which lacks any digital conversion is truly and purely analog, but anything which has at least some element of digital conversion can have varying amounts of analog circuitry.

I think the poster child for this blurring is the case of low-end delay pedals.  People say they like the "warmth" of analog delays that use bucket brigade chips or tape.  A big share of that "warmth" actually results from the filtering that has to accompany such circuits in order to manage noise levels and clock-related artifacts within the audio range.  When one applies such filtering to inexpensive digital delay chips, they can sound pretty good and have many, if not all, of the qualities that people used to think stemmed purely from using bucket brigade chips (formally classifiable as "analog shift registers").

[mantella]

yeah! Thanks for all the great responses. These last two posts are kind of what I have been trying to wrap my head around with this whole thing. I get the whole binary on or off thing, but really that's only saying "yes, the voltage reached 5 volts here", or "no it didn't reach 5 volts, so I'm going to pretend it didn't exist at all".
I think my hang up here revolves around my understanding (or lack thereof) of the way transistors operate. Can transistors pass a continuous (analog) signal?

[earthtonesaudio]

Can transistors pass a continuous (analog) signal?

Absolutely.  The overwhelming majority of stompboxes use transistors in this way.



I would add to the general discussion that binary digital represents only one type of digital (even if it is by far the most common).  To me digital simply means that the analog information has been encoded as numbers (i.e. "digits") at some point.

[PRR]

> "no it didn't reach 5 volts, so I'm going to pretend it didn't exist at all".

Misunderstanding.

"Zero" is a perfectly good number. And significant: if your wallet has zero dollars in it, you can't pretend the situation does not exist. If your heater uses zero fuel, that's fabulously significant.

You speak of one-bit logic, but most interesting stuff uses 8 or more bits. With 8 bits you can indicate any number from zero to 255.

Take a slowly varying (and analog) signal. Measure the voltage. Measure it again. Keep notes of time and voltage. Say the readings are 0, 97, 123, 84, etc. You can store or transmit those numbers. Then you can re-create the original signal by making voltages of those numbers at those times.

For good audio you must take readings faster than your highest frequency and use more than 8 bits. 8,000 samples per second with a modified 8-bit code gives telephone quality. CD quality is 48,000 16-bit samples per second. With this many samples and that many possible values, you can reproduce audio very well, objectively better than most pure-analog systems.

The pure analog system does not "sample" the signal and slice it into time and voltage. It is a lever. In theory it can move continuously with infinitely small motion. We just bias a transistor, not on or off, but half-way in the middle. Not 0mA or 10mA, but 5mA. Then we can swing 4mA to 6mA, or 1mA to 9mA, or 0.499mA to 5.001mA.

Simple analog systems are far cheaper than digital audio systems.

But analog is prone to small errors. We can't really pass infinitely small signals, they run into universal hiss. We can't pass huge signals without some errors. Real systems have many stages, each adding a little more error.

In digital audio, once you have captured the original analog air-wave or voltage into a bunch of hard numbers, these may be stored or transmitted "perfectly". A computer won't take "222" and output "221". (For other reasons, practical digital audio DOES have errors, but of different sorts than analog, and potentially fixable.)

Digital audio would be grossly expensive except there has been SO much development and production in digital systems for gunnery tables, email, and pedal-porn websites. Massive digital power is now frikkin cheap. If a sound system needs more than a volume control, it is getting cheaper to digitize it and use software to re-shape it. 99 effects pedals would cost you $5,000; you can buy a digital pedal with 99 sounds for $59. Partly from not needing the other 98 boxes and jacks, but also because once you blow for the digital chip, you can program-in more sounds without extra parts or construction.

[Lurco]

In digital audio, once you have captured the original analog air-wave or voltage into a bunch of hard numbers, these may be stored or transmitted "perfectly". A computer won't take "222" and output "221". (For other reasons, practical digital audio DOES have errors, but of different sorts than analog, and potentially fixable.)

yes: it will output "222" after a while. If you chain 5 pedals, the last one will put it out after 5 whiles : perfect for delay purposes.