If you’d never heard of digital audio, but were told that it was now possible to store and play back audio signals on a computer, I don’t think you would raise an eyebrow. After all, how difficult can it be? An audio signal is no different from any other ‘wiggly line’ that computers seem to manipulate with ease: graphics, high quality fonts, CAD drawings, maps etc. for all intents and purposes at infinite resolution.
But somehow, digital audio is seen as a special case, where no one quite believes that it works. Looking at various forum discussions it is apparent that, in fact, it wouldn’t matter how many bits, how many MHz of sample rate, how few femtoseconds of jitter was specified, audiophiles would still be convinced they could hear the ‘1’s and ‘0’s, jitter, quantisation distortion and so on. The noise and distortion inherent in tape and vinyl that is many orders of magnitude greater gets a free pass; the noise in digital audio no matter how minuscule must always be portrayed as ear-bleedingly offensive. Why?
I think there are several reasons:
- Digital audio is mathematically-based. Long after real world signals have become buried in noise and distortion due to unavoidable physics, the theoretical numbers associated with the maths remain pristine and, quite unambiguously, show errors! Clearly we need better numbers. And so it goes on. In other words, no matter how high the resolution, you can always zoom in and see a theoretical error that looks just as big and clear on the screen or page.
- From the outset, the theory behind digital audio was discussed openly, but very few people actually understood it fully (including me). Thirty-odd years later, the misunderstandings persist. These vary from assuming that digital audio cannot know, or fill in, what is in “the gaps”, to failing to understand the significance of dither.
- Digital audio provided a complete mathematical solution, in many ways superior to other computer-based wiggly lines. The system is so elegant and simple that people just don’t believe it can work the way it does. [03.03.16 just saw an article that says exactly that: “The intriguing aspect is that those who do understand refuse to believe”]
- Digital audio must always be chasing its tail, because as soon as a new performance level is achieved, it becomes possible for every Tom, Dick and Harry to buy the hardware for a small number of pounds, and even to start measuring signals at that level. Suddenly we’re all experts for whom -110dB is an average spec and must therefore be highly audible – although no one has ever heard a signal that quiet. No matter how good, digital audio will always seem mundane.
- Digital audio hardware is too complex to build using discrete circuitry. Integrated circuits are cheap. Audiophiles need to know they are buying better stuff than the hoi polloi, but digital audio doesn’t play the game. It remains persistently cheap enough for the masses to buy exactly the same measured performance as the most expensive fancily-boxed version of the same chip. (We are talking £30 versus £30,000, say). In the audiophile mind, this proves that measurements mean nothing and that “bits are not bits”, whereas in reality it shows the opposite.