It’s interesting, and something I’ve been wondering about for a while. It’s generally true that people will opt for convenience and cheap/free over quality. People will put up with a pretty fuzzy picture if it’s something they want to watch, and MP3 is an acceptable standard.
Human hearing ranges from 20Hz (cycles per second) to 20kHz. There’s usually very little musical information below 40Hz, and your fundamental bass notes are around 80-200Hz. The lowest E on a guitar is 82Hz, but you’d rarely play an open E string. Music compression can work in a number of ways. The Sony Minidisc method was to strip away information that was masked or hidden by louder sounds in front of it. The MP3 method is to remove information from both ends of the human hearing spectrum.
I’ve analysed MP3 files in the past, and noted how anything about 15kHz tends to get stripped off. That can mean all the brightness and “air” on an instrument can be missing. It seems to be replaced with an artificial “sizzle” at around 15kHz, which is wearing on the ears – unless it’s all you know, of course.
Early listeners to CDs complained of “digital fatigue”, because the early implementations of the 16-bit 44.1kHz digital standard were low quality. The problem there was the 16bits. “Bits” refers to numbers, and it’s literally the number of digits used to describe a given sample of sound. You can have a decimal point that floats around, but you only have 16 slots for numbers: anything beyond that gets stripped away. There are two ways this can become a problem. The first is with samples that go beyond the 16bit level – they’ll just be stripped back, leaving a truncated number, no longer exactly describing what it was sampling.
The other problem comes when you “do maths” on a sample of sound. Say it’s not loud enough, or you want to boost it at 5kHz for some “presence.” You’re essentially performing a multiplication on a portion of the sample. Multiply a decimal by another decimal, and you can soon end up with more than 16 digits. What happens to the extra digits after the decimal point? They get stripped away, leaving another inexact sample. This kind of thing happens every time you use a calculator with just 8 digits on its display – all the answers are limited to 8 digits, even if they’re really not (like Pi, which is 22/7 and goes on forever).
Music software these days has internal 32-bit or 48-bit processing, and recordings are made using 24 bits. When the stereo file is “bounced” to a 16-bit file, they use something called “dither” to add random numbers to samples, which sounds more pleasing to the ear than the previously truncated samples.
In other words, all systems for recording music are inexact, and noise can be a virtuous companion to the musical signal. No system for recording music has ever been entirely free of noise, whether we’re talking about the wow and flutter, pops and scratches of a vinyl record, the hiss of a tape, or the dither on a CD. What you prefer tends to be what you’re used to, which is why people of my generation sometimes look back with longing at the great sound of vinyl.
But I go back to my original point. People will put up with sub-optimal listening conditions. I used to listen to stereo records on a mono record player. Fact! I didn’t hear the guitar solo on “While My Guitar Gently Weeps” until I’d owned it about a year and a half. I used to have a fetish for mono, and when my band made a single, I insisted it was in mono. People thought we did it because it was cheap, so that kind of fell a bit flat.
Anyway, this idea, that no system for recording music is without some form of noise, is a perfect illustration of Michel Serres’ theory of the parasite: the rat in the foundations. Noise isn’t just part of the signal; the signal depends on the noise for its very existence. Noise is all the stuff that science, mathematics, and theory used to ignore:
Leopards break into the temple and drink the sacrificial chalices dry. This happens again and again, repeatedly. Finally it can be counted on beforehand and becomes part of the ceremony.
–Franz Kafka, Parables