12
$\begingroup$

I have just started studying Digital Signal Processing. Can someone explain what is the difference between a Discrete Signal and a Digital Signal in simple words?

Thanks in advance !

$\endgroup$

4 Answers 4

19
$\begingroup$

Suppose you have a continuous time analog signal. It is continuous in both time and amplitude. Now when you sample it ,you get discrete samples every Ts seconds. Now you have discrete samples(discrete in time) each of which can take continuous value( in amplitude). This is normally referred to as discrete signal(discrete in time but continuous in amplitude).

Now further when you take this discrete signal and quantise it i.e. say assign each sample which takes continuous amplitude to one of the N discrete quantisation levels of a quantiser, Then your total signal now is a digital signal. So a digital signal is discrete in time and discrete in amplitude.

$\endgroup$
7
  • $\begingroup$ @jojek: ADC: Analog to digital converter. Output of an ADC is the digital signal. What you mean is the sample and hold part an ADC block, but your definition gives the wrong impression that an ADC can provide you continuous valued outputs. This may cause confusion to the novice. Besides the quantizer is already inside the ADC block not some other block later to apply to the output of an ADC. $\endgroup$
    – Fat32
    Jan 25, 2015 at 20:43
  • $\begingroup$ @BulentS. Please post a comment next time (as you did now) pointing out the mistake and do not change the someone's answer. $\endgroup$
    – jojeck
    Jan 25, 2015 at 20:48
  • $\begingroup$ @jojek: I have not changed it. I have corrected it. besides why then there is the "edit" button ? $\endgroup$
    – Fat32
    Jan 25, 2015 at 20:52
  • $\begingroup$ @BulentS.: This is a pointless thing to argue about, but there is a thin line between editing (i.e. correcting equations, spelling, typos, formatting) and correcting, that is alternating the meaning of the answer (although that it is not the case with Community Wiki answers). I am sure you will agree with that. Usual thing to do is place a comment. $\endgroup$
    – jojeck
    Jan 25, 2015 at 20:56
  • $\begingroup$ @jojek: It was a conceptual edit. The pointless thing is your retainig the wrong answer. An answer with 6 upvotes must be a correct one. Some of my answers were also edited to make them better. This is for the sake of community. $\endgroup$
    – Fat32
    Jan 25, 2015 at 21:00
1
$\begingroup$

In electrical engineering, the terms "discrete signal" and "discrete-time signal" are used intermittently to denote the same thing, the latter term is more common and more precise. The wording "discrete-time signal" is self-explained to a degree.

Many engineers and theorists have equated and are equating digital signal processing to signal processing of quantized discrete time signals. There are subtleties, however. The information-theory concept of signal being too hard to squeeze into "simple-word" explanation, I suggest that you learn first what is a "digital waveform". A waveform is not a signal but may represent a signal in your designs of processing units. For a beginner in DSP studies, knowing what is a digital waveform is a practically useful knowledge with a promise to help you advance through the subtleties of the information-theoretical concept in due time.

Digital waveform is a voltage or a current that varies with time between values called "logic levels", these levels are defined by "digital abstraction" discipline ( http://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-002-circuits-and-electronics-spring-2007/video-lectures/6002_l4.pdf ). The varying voltage or current values not always fall into limits prescribed by digital abstraction rules for signals values at logic levels: when switching between levels, a voltage or current, being a continuous function, passes "forbidden" ranges. The same digital abstraction discipline prescribes to consider restrictions not only for values of logic levels, but also for timing parameters, of which the very important are the setup and hold times -- guard time intervals of signal stability surrounding the assertion edge of the clock. See a good tutorial on digital waveforms in http://www.ni.com/white-paper/3299/en/ . Quite remarkably, digital waveforms are called digital signals in this tutorial.

In (relatively simple) words, digital waveform is a waveform with logical signal levels and timing parameters conforming to digital abstraction rules for a given design. What is a legitimate digital waveform for one design may happen to be inappropriate for another.

$\endgroup$
0
$\begingroup$

@Talasila was right.

In short,

Analog(continuous) >> sampling >> discrete >> quantization >> digital

$\endgroup$
-3
$\begingroup$

Digital signal is a continuous signal.Discrete signal is a non continuous signal. i.e the digital signal is present at all time(continuous signal). the discrete signal is present some time values only(non continuous signal)

example for digital signal

example for discrete signal

$\endgroup$
6
  • $\begingroup$ Your answer is wrong. As most (if not all) the other answers state, a digital signal is NOT continuous. Both discrete and digital signals are discrete in time so neither of them is continuous and "present" at all time! Please consider answering questions that you only know the answer for sure and if possible provide some "evidence" (such as references and links) $\endgroup$
    – ZaellixA
    Aug 26, 2020 at 9:35
  • $\begingroup$ @ZaellixA: A digital signal may certainly be continuous-time. Consider the output of a comparator. If the input signals are continuous analog, then the output is continuous (asynchronous) digital. That fact does not save this answer however. $\endgroup$
    – Ben Voigt
    Jan 11, 2023 at 19:40
  • $\begingroup$ @BenVoigt, part of the definition of a digital signal is that is first of all discrete in time. This in turn means that the signal is defined only at specific time instances. $\endgroup$
    – ZaellixA
    Jan 11, 2023 at 20:02
  • $\begingroup$ @ZaellixA: No, it isn't. Sampled digital signals are a very very useful category, to the point where you can collect an entire library of books who only discuss digital signals which are discrete-time. But that doesn't mean that asynchronous digital signals don't exist, or that all digital logic is clock-synchronous. $\endgroup$
    – Ben Voigt
    Jan 11, 2023 at 20:05
  • $\begingroup$ @BenVoigt, thanks for the comments. I believe I do not completely understand your suggestions. If I understand correctly though, you suggest that when treating digital signals that are not "synchronised" in some way, they are assumed to acquire values between their respective sample intervals. I have to admit that I have never encountered such a case (I am not suggesting it may not be important though), but to me this sounds like in such cases those signals are treated as continuous-time, discrete-amplitude signals. If this is the case, these are not digital signals according to (cont.) $\endgroup$
    – ZaellixA
    Jan 14, 2023 at 12:21

Not the answer you're looking for? Browse other questions tagged or ask your own question.