Advantages of digital data over analogue data

Comments

• To the unaided human eye, the world appears to be analog in nature, but when you peer very very closely, you find that it too is quantized. Light from a monitor is quantized, sound is quantized, matter is quantized, energy is quantized, etc.
  
  In a sense, the world is digital, but the size of the quanta are so small, and the numbers of quanta are so vast that it appears to be analog.

  • Report
  

  by yeroco on October 29th, 2009

• Yes, but the term analogue would but somewhat useless if we looked at things that low. Plus, the layman's digital has evolved to mean only 1 or 0 rather than quantized and not. So in a practical sense, light and sound are analogue, though vitually everything is digital if you look at it technically.

  • Report
  

  by POP Fan on October 30th, 2009

Comments

• And there are ways to check to see if the digital data is correct, and if not correct it. There is no degradation, except if it is compressed using a lossy compression technique, which is only used if a little loss is unimportant.+

  • Report
  

  by JimmyG on October 29th, 2009

Comments

• See my comment to Greekgod's answer.

  • Report
  

  by JimmyG on October 29th, 2009

• ok you can be right, but this is what im taught in school

  • Report
  

  by Always Ready - COAT of courage on October 30th, 2009

Comments

Comments

• More accurate, yes, because it's accuracy can be checked, whereas with analog it is what it is, no way to check. Faster, no. Analog actually is faster because there is no processing involved. The first computers were analog and were very fast. Multiply and divide circuits were almost instantaneous, but their accuracy was only so-so and would change over time with the aging of components and with changes in conditions. You'd get different answers on a hot day than on a cold one. But even with the advances in speed with digital circuits an analog multiply circuit will still win, because the answer appears as fast as the electrons can propagate through the circuit.

  • Report
  

  by JimmyG on October 29th, 2009

• There is still processing involved in analogue circuits. It takes op amp circuits to do those calculations, and they aren't infinitely fast. However, there is some error checking you can do... for example, you can have 3 circuits perform the same calculation, and then compare the results. If the 3 results don't agree within a certain tolerance, you can flag an error condition. I don't know if this was done in practice, however. In those days, the circuits were expensive to construct, let alone tripling the amount needed.

  • Report
  

  by yeroco on October 29th, 2009

• Op amps don't calculate, they propagate. Having three of everything just to tell you the calculation was incorrect wouldn't be of much help when your only recourse would be to rebuild the circuits. Even with high precision components (expensive) they were and are only useful where their accuracy is 'good enough.' And there are countless situations where this is the case.

  • Report
  

  by JimmyG on October 29th, 2009

• I suggest you look at how multiplying and adding is done using analog circuitry. It is done using active electronics, including circuits which include op amps. Op amps are not just used as buffers or for propagation. They use carefully designed feedback circuits to produce the functions needed.

  • Report
  

  by yeroco on October 30th, 2009

• Well in the industrial instrument world I can assure you digital signals from flow meters like frequency o/p is much faster and more accurate than a milli amp signal.

  • Report
  

  by Greekgod on October 30th, 2009

• Aging if components are why you do calibrations to a reference although I have meters installed in clients for ovver 15 years now with no component drift.

  • Report
  

  by Greekgod on October 30th, 2009

• Maybe it's semantics. The simplest amplifier circuit uses feedback, I'm not sure what that has to do with anything. Every electronic circuit has a propagation delay, that is the time from when a signal is applied to the input until a valid signal arrives at the output. These delays are caused by the nature of the circuit and the limitations of physics. Feedback will affect this delay because it alters the input and it takes time for the circuit to achieve equilibrium but these delays are extremely short. Delays in a computer are built in because it is driven by a clock. Nothing happens until the clock ticks. And in the early times we are talking about a multiply of two unsigned 16-bit integers involved a loop that was executed 16 times, each of which involved two shifts, a test and branch, and a possible add, plus the setup before and cleanup after, each step dependent on ticks of the clock. The speed of the clock itself is dependent on propagation delays in the computer circuits.

  • Report
  

  by JimmyG on October 30th, 2009

• Well, in the early analogue computers, there was no clock, or shift registers, etc. You have analogue signals as input representing some values,... you run them through an analogue circuit which performs operations like multiplying, adding, exponentiation, etc. and out the other end pops an analogue result (which its inherent noise and other sources of error).

  • Report
  

  by yeroco on October 30th, 2009

• I dimly remember one of my first projects in electronics was building a 'computer' with only a battery, several potentiometers, an ammeter and a voltmeter. It could add, subtract, multiply and divide. There were instructions on how to implement higher functions which none of us bothered with. The point of the whole exercise had nothing to do with computers, which at the time only existed as behemoths with vacuum tubes, but was to teach us Ohm's law.

  • Report
  

  by JimmyG on October 30th, 2009

• Kept ya off the streets ;-)

  • Report
  

  by yeroco on October 30th, 2009

Comments