What is the difference between digital and analog audio?
In: Technology
OK, here’s a really ELI5:
Sound travels in waves. Tie a jump rope to a fence and wave it up and down; the shape of the rope will resemble a sound wave. Now imagine you could freeze time, and you wanted to build a copy of the rope’s shape, but you only had bricks.
So, you take your bricks, and start to stack them up under the rope. Some times you’ll only need a couple of bricks; sometimes you may need to pile them up 10 or 12 high to touch the rope. After a while, if you step back a bit from your work, you can see how the piles of bricks look *very much*, but not exactly, the shape of the rope.
The rope is the “analog” wave form, while the bricks are the “digital representation”. The analog wave is *continuous* – the rope’s height above the ground can have any value between, say 2 inches and 4 feet. The digital representation is *discrete* – it can only be 1, 2, 3, 4, etc. number of bricks. It can’t be 3.867 bricks.
Analog systems capture the continuous wave. The groove in a record – do 5 year olds even know what those are anymore? – is a long continuous wiggle that copies the original sound wave. This is actually fairly simple to do – the first records were made of wax, with the platter rotating while a needle, driven by a microphone, made the groove on the surface. This is an analog to analog process.
Digital systems try to recreate the original wave by using standard sized pieces to fill in the space beneath the wave, just as we did with the rope. But how wide, and how tall, should each of these pieces be?
This is beyond ELI5, but there was a smart guy named Nyquist who figured out that to completely capture all the information in the original wave, it needs to be sampled at twice its highest frequency. This tells us how “wide” the bricks need to be. For example, if the highest frequency in the wave was 4000 cycles per second, then we would need 8000 samples, so our ‘bricks’ have to be 1/8000 of a second wide.
The height of the bricks are is a function of how many digital bits in each brick. If you use 8 bits, you can get 2^8 = 256 levels. If you use 16, you get 2^16 = 65,336 levels. If you use more bits, it makes the bricks less high, so you can squeeze the brick piles closer to the actual wave, and so sound more like the original.
Note the digital process requires an analog-to-digital conversion at the input, and then a digital-to-~~audio~~analog conversion at the output. There are some – Neil Young comes to mind – who believe that this distorts and ruins the original recording; others don’t notice it.
finally, and this is way beyond ELI5, digital techniques like Adaptive-predictive Pulse Code Modulation (ADPCM), use clever math and engineering tricks to get the sound even closer to the original, while using less bandwidth.
EDIT: Thanks for all the kind comments and awards. Thanks also to those who corrected the minor errors, and expanded on some of the stuff I left out.
EDIT EDIT: To all the longitudinal wave fans. yes, you’re right. So am I. A sound wave can be represented as a two-dimensional signal on an oscilloscope, and it was that representation I was referring to. I elided the silly scope reference because it’s ELI5.
Latest Answers