Digital audio, whuzthat ?
A sound is a physical freak. A vibration movin’ in the air. An example ? When you pluck a guitar string, when a hammer hits a piano string, this string’s gonna vibrate and put it’s surrounding air out of shape. Your ears among others, will catch this vibration, transmit it to your brain and translate it in a sound. That’s in out line what happens.
Up to there, it’s a natural freak ! But it’s gonna be a bit harder when we write the sound and then store. We know two way of doing : an analogic one and a digital. Everybody knows the first, that’s for example the magnetic tapes in their different formats. The second one consists in coding the signal in numbers. Let’s go !
It begins by an incoherence ! A sound is a continuous freak. A vibrating guitar string does it without stopping. When numerizing, the sound is divided in slices. When obtain a discontinuous information ! The trick to get round this peculiarity is to create such closes slices that your ears will make neither head nor tails of it.
I play A on a guitar. I’ll say nothing about the transmission of the vibration of the string through the microphone, then through the pre-amp, the loud speaker and the capture of this sound for the loud speaker to the microphone. Phew !!! My A is just at the entrance of the analogic / digital converter. Up to there, I had a vibration, it became electricity (incoming through the guitar microphone) and once again a vibration (out coming through speakers). The microphone (connected to the converter) caught this new vibration and transformed it in electricity anew, this latest coming up to the converter : “hello, here I am, I’m the A” !!! This is analogical. We had an electric flow that the loud speaker was about to give back providing an analogous(nearly) wave. Incidentally : regarding to the trip of our A, from the string to the converters, it would be surprising and in a word unexpected that it would be similar between the beginning and the arrival. Just a thought in order to complicate things.
Ok then, we’re at the entrance of the converters. Those contraptions will take very close snapshots of our “A”. The number of pictures per second may determine the Sample Rate. The scale is given in Hertz (Hz). Each measurement’s gonna be calculated. This measurement have to be very precise. Everything depends on the number of values you wanna use. As you know, our stupid computers only understand 1 and 0 ! We call this couple a Bit (an English word meaning Binary digIT). The eight BITS packet name is Octet. When 8 bits, the computer will use values from 0 to 255. Not much ! With 16 bits, the CPU got 65535 values to it’s disposal. That’s better isn't ? That’s the quantification.
For example, in the famous CD quality, 44100 Hertz (44.1KHz) for 16 Bits, The A will be sliced 44100 times per second, each slice got it’s number from 0 to 65535. Why 44100 Hertz? Well, Mister Nyquist, a little guy who passed all his exams decided ( demonstrated ?) that for a good digitalisation, the sample frequency had to be twice higher than the highest audible frequency (for the human Kind nft). Hunk isn’t he. Most of us are able to receive frequencies from 20 to 20 000 Hertz, which explains our 44100 Hertz. This is the sampling process.
OK then but why 16 Bits ? Well, we’re entering in the great controversy of the digital sampling. Some think that we stopped at 16 Bits just because of marketing ! 44100 Hertz, 16 Bits gives a 74 minutes audio-CD. Perfect for the albums distribution. It only remained to persuade the basics consumers that this quality was the top of the top ! When you know that 99 % of the music runs on machines that only get the name of Hi-Fi, it wasn’t too hard !
Then the professionals and the purists are all up in arms because of that sting. AAAAARRRRRRGGGGG !!!! They want 20 or 24 Bits. They must be quite right because you can find sound cards for CPU offering such resolutions. Your ears only can judge : do you think the sound of your CD’s so crap ? Just make your own opinion !
By the way, why digital instead of analogic ? Well, With numbers, you can do a lot of things ! Calculations if you realise ! A lot ! And also store datas without any loss, which happens always one day on a magnetic tape. We’ll reach a 100 dB (decibels : measurement unity of a sound level. For instance, the dynamic corresponds to the difference between the lowest and the highest noise an audio device can produce) versus 70 maximum with a tape. We can also dink around (patch up with the sound), reverse the sound, editing, correct the pitch, (cheat in fact), the tempo, copy an paste, without cutting or sticking the tape. Much better, without losing the original file. Who can do better ?
Digital recording is perfect isn’t it ? Well hum, not really. You may no use the various treatment above-mentioned to excess. It come’s from all those calculations are made with whole numbers, with corrected results so. For your computer, 3 divided by 2 gives 1. If you multiply this 1 by 2, you’ll get 2. Do you imagine the mess. So if you misuse the digital recording potentialities, you may obtain a..., how should I say mmmm, an quite... exotic signal !
This explanation is a bit simplist. It’s a first approach. You can find various books and web sites talking about this subject. You’ve got to search on your own depending on the way you come along. See you and good luck. And keep connected.
PS : decency requires that I specify that the special issue of Home Studio dedicated to the digital sound was a precious source for me. Specials thanks then to Jean Poncet.