Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Analog vs digital

  1. Jan 2, 2006 #1
    one of my (former) profs mentioned why the sound quality of an analog recording is better than the sound on a digital recording but i can't remember what he said. can anyone explain to me the difference, or tell me where i can find out for myself?
  2. jcsd
  3. Jan 2, 2006 #2
    i think people will in the future go back to using analogue, or create analogue computers. then they wold be like..........absolutely exact, rther than digital incraments.
  4. Jan 2, 2006 #3


    User Avatar
    Gold Member

    Digital conversions take large numbers of samples of music in order to digitize them. This allows them to be stored in much more compact forms. The downside is that you aren't getting "the whole story" when it comes to the song. Every sample is a small moment in time whereas analog is just complete and continuous. It's kind of like taking a photograph. If you have an old fashion camera, it replicates the light and everythign on a continuous scale. If you have a digital camera that can say, only capture 10,000 different shades of color and only 50,000 pixels (obviously technology is vastly better then this), you won't see anything between say shade #5802 and #5803 (theres a different system of determining what color something is though of course). Also, you would have a 1 mm^2 area represented in a picture that is 1 pixel to the camera and it can only have 1 color whereas in reality, if you walk up to the object, that 1mm^2 area has many many many more colors and more details. Same with digital audio recordings.

    Of course, at some point you can create a digital replica with so much detail that even audiophiles can't tell the difference... but technology today pretty much has it down to where the general public can't tell the difference between a vinyl analog record and an mp3. Not the technologies fault, more of the publics fault for not having that good of an ear.
  5. Jan 3, 2006 #4
    Audiophiles seem to have a preference for analogue media like LPs, saying they have a "warmer" sound. I guess it's a pretty subjective matter really.
    But some would argue that no recording media can ever be truly analogue since LPs, for example, are limited in sample rate/resolution by the size of a PVC molecule (which is small, but certainly not infinitesimal). Assuming this is true, would anyone care to hazard a guess as to the sampling rate/resolution of an LP?
    Well, DVD audio allows for a sampling rate of 192KHz at 24-bit resolution. By the sampling theorem, this would allow for perfect reproduction (as is my understanding) of signals up to 96KHz, which seems like overkill given that the upper limit of human hearing is only about 20KHz. Perhaps it's so that bats, whales, etc. can listen too. :rofl: Actually, I think it's something to do with filtering (having less steep cutoffs, yielding less distortion, or something like that).

    In any case, I would submit that the sheer convenience of digital media outweighs any other consideration.
  6. Jan 3, 2006 #5
    Phantom Photon, having a sampling rate of 192,000 samples per second, does have a difference to humans than say, 48KHz. It has nothing to do with our ability to hear pitches of sound higher than 20KHz.

    And yes, it is possible with digital to reach the human maximum. Who's the guy who came up with the theorum for this? It is possible to reproduce analog sound to sound the same to us...
  7. Jan 3, 2006 #6
    i read up a bit about this & found a (very) informal explanation. the so-called audiophiles like analog better than digital (ie records vs cds) because rather than listening to an approximation of a wave you hear the actual wave itself. however, the nyqvist-shannon sampling theorem (aha! there it is) says that if the sampling rate of the digital thing (usually 44.1khz for cds) is at least twice the bandwidth of the analog thing the sound will be identical. i haven't looked at any signal processing books or seen the real statement of the theorem or its proof but i think that's the general idea.
  8. Jan 3, 2006 #7
    Sorry, I must have misunderstood the sampling theorem. I thought that it was possible to recreate a 20KHz signal (and others of lower frequency) perfectly by means of a weighted sum of normalised sinc ( = sin(pi*x)/(pi*x)) functions but I think I got the wrong idea.
    You might be thinking of the sampling theorem of Claude Shannon and Harry Nyquist (also attributed to some other guys) which is what I have been referring to.
  9. Jan 3, 2006 #8
    That's what I was thinking, but there's still the issue of sample resolution... the sampling theorem doesn't tell us anything about that... or does it? Could someone with DSP expertise settle this matter for us, please?
    Last edited: Jan 3, 2006
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook