Why Not Improve Analogue Instead of Expanding Digital?

[munqui]

Originally posted by jimmy p
Well the lecture i went to, the dude said that...well i suppose the idea of digital is old fashioned with Morse Code and whatever. What i meant was that they are simple (eg. square) whereas normal waves are...wave shaped.

Yes, normal waves *are* 'wave shaped'; but you've missed the point about 'digital' representations a little -- they're not 'square' per se; just slightly jagged compared to the analogue.

So what's the difference? Simply, an analogue signal varies continuously. A digital one does not.

To illustrate; let's imagine you've drawn a waveform with pencil and paper. You want to show it to a friend, but you don't want to send the paper itself (its your only copy), so you measure the amplitude of the waveform at regular intevals, write down the values you measure, and post these values to your friend. He can 'reassemble' the waveform by plotting the amplitude values and joining the dots. The representation you send to your friend is digital; whereas the waveform you drew on the paper in the first place is analogue.

Here is where the distinction comes in -- your waveform on the piece of paper *isn't* the same as the numbers you sent your friend; as you made approximations each time you measured the amplitude (you can't measure more than +/- 0.5 mm on most rulers, for example). This is why analogue might be considered 'better' than digital -- digital has inherent inaccuracies.

So -- why use digital? (the original question)

i) You don't *need* all the information in an image to perceive the important information it conveys. Hence, if we represent an image digitally, then strip out the unneccessary information, we save on transmission costs

ii) Analogue signals are much harder to transmit than digital ones

iii) Digital signals allow good error checking, analogue signals don't

iv) Digital signals can be heavily manipulated by computers (DSP) and as such we can implement sophisticated systems using them that wouldn't be possible with analogue

The list goes on...

 

[jimmy p]

ok cool, that is what i would like to hear. So digital, though it can be heavily editted is not as finer quality than analogue, not that we can detect the difference very easily

 

[munqui]

Thats pretty much it. Digital signals are only ever a representation of their analogue counterparts. It really depends what you're talking about when you come to questions of quality, though -- as has been mentioned, audiophiles prefer the sound of vinyl over the 'cold' sound of CDs. However, if you're discussing broadcasting, much better quality can be achieved with digital systems because of the error checking and noise-resiliance discussed earlier.

 

[Cheman]

On the topic of analogue and digital I have a question - we always talk about having copper cables for analogue signals and fibre optics containing light for digital signals - why not digital signals in copper cables and analogue in fibre optic cables? After all, surely it would be possible since voltage could go on and off (to produce digital) and we could alter the intensity of light rather than just turn it on and off (for analogue). And thus why do we still use analogue with existing copper cables - since we are trying to replace them with fibre optics anyway surely we could also change the existing ones to digital in order to reduce noise?

 

[chroot]

You can do digital communication over copper quite easily. It's done all the time. That's what ethernet, USB, serial, DSL, digital cable television, etc. are.

In theory, you can also do analog communication over fiber, but it's not done in practice, since the laser drivers for an analog link would be expensive. It's cheaper to build a digital laser driver.

 

[Cheman]

Thanks chroot. : wink :

 

[Cheman]

( NB - that should be a simble not words! lol! ie -

 

[Cheman]

On the topic of analogue and digital, I've been taught that "digital signals travel further before needing to be amplified" - why is this the case?

 

[chroot]

Let's say you're transmitting a signal (sound, perhaps) which requires a minimum of a thousand discernably different sound pressures for good reproduction. If you transmit this analog signal with a one-volt amplitude, each of the "discernable steps" is one millivolt. As this signal loses amplitude over distance, the size of the discernable steps becomes smaller and smaller. It's easy to build electronics that can reliably discern signals on a millivolt scale, but it's much more difficult to build electronics that can reliably discern signals on microvolt or nanovolt scale.

On the other hand, digital signals only require two discernable values: zero and one. The electronics to discern zero from one don't have to have good linearity. You don't have to use very precise ampllifiers with very clean power supplies. You no longer need to discern a thousand distinct levels, only two. Even if your digital signal's amplitude is in the microvolt or nanovolt range, it's still pretty easy to build reliable electronics to recover the signal.

- Warren