- Chapter 3: Digital Audio, Continued
- The Ear Proves (Once Again) It Has a Lot to Offer
- Dither: Compensating for a Digital Problem
This chapter is from the book
This chapter is from the book
This chapter is from the book
The Ear Proves (Once Again) It Has a Lot to Offer
It's at this point that we encounter one of the great tricks of digital recording. Take another look at Figure 3.6. Notice that the sound wave is smooth, but the quantization stair steps form a jagged pattern.
){kind=link}
How, you might ask, can we record a smoothly changing volume with something that uses stair steps?
A digital recorder, like a great dancer who can move up stair steps with an effortless smoothness, replicates a subtly changing volume level by assigning each small change to a given stair step.
Hot Link: Amplitude (http://www.sfu.ca/sonic-studio/Amplitude.html)
Amplitude, shown in Figure 3.8, lays out a lot of information about audio theory with very little words. We especially like the example in which they offer downloads of different-shaped sound waves (click the Waveform link).
){kind=link}
Figure 3.8 Amplitude offers downloads of various types of sound waves.
The key factor that gives a digital recorder the ability to accurately capture subtly changing volumes is how many steps it uses. In the case of an 8-bit audio file the number of volume gradations is 256. So an 8-bit recording quantizes all the volumes we hear, from whispery soft to jet engine loud, to 256 possible stair steps.
That seems like a lot, right? You'd think so, but once again the ear—that amazing organ—demonstrates what a superlative sound receptor it is. Because 8-bit sound does not fool the ear. Oh, it's adequate. You can distinguish a dog bark from a car engine. But to create audio that satisfies the ear as high fidelity, you need the 65,536 stair steps of 16-bit audio (16-bit audio is used in CDs, among other many other uses).
So this is digital recording's sneaky trick. It reproduces gradually changing volumes by using stair steps. But each step is so tiny that it fools the ear into thinking it's a smooth unbroken line. Pretty tricky, huh?
While 16-bit is a popular standard, as sound technology moves forward more applications offer 20-bit and 24-bit sound. Audio at 24-bit depth has a mind-boggling 16,777,216 volume gradations. That's probably enough.
Hot Link: Electronic Musician (http://www.electronicmusician.com)
Back before the desktop studio was the phenomenon it is today, this mag was doing a good job of covering it, and it still provides a lot of well-written commentary and advice (see Figure 3.9).
){kind=link}
Figure 3.9 Electronic Musician does a great job of covering desktop recording.
Dynamic Range
The term dynamic range refers to the difference between the softest possible signal and the loudest possible signal a system can reproduce. The better the bit depth of an audio file, the more dynamic range it can reproduce. Here are the dynamic ranges of various bit depths:
16-bit audio has 96dB of dynamic range.
18-bit audio has 108dB of dynamic range.
20-bit audio has 120dB of dynamic range.
24-bit audio has 144dB of dynamic range.
These figures are theoretical. In the real world, factors like speaker quality influence the dynamic range of an audio playback system.