What Is It That We’re Hearing, Anyway?
For the last 2 months I’ve been trying to offend everybody by suggesting some things about the audibility of high-resolution audio, and I’ve taken the position that hi-res isn’t very audible, compared to 16-bit or even analog audio. I’ve tried to also explain why there’s such a fuss about hi-res (it’s because we can
hear it, a little, and we’d like to think that makes it important and us cool).
This month I’d like to discuss what it is we probably do hear when we listen to our favorite hi-res audio and are sure
we really do
hear a difference. Although most of the limits of hi-res are well-beyond the various thresholds that limit and/or mask our perceptions, we do hear stuff – and it is worth discussing what we do in fact hear.
First, let’s admit right upfront that some of the hi-res goodnesses we hear are imaginary. Now that doesn’t mean we don’t really hear them, or that they aren’t important to us. But you’ve gotta understand that we hear things, with that remarkably complex multifaceted hearing system of ours, things that sometimes aren’t there physically. We are really good at extracting meaning out of chaos, and sometimes we extract meaning from signals that really are … well, just chaos – signals that don’t have any meaning in them. AWESOMELY EVANESCENT
There’s nothing wrong with this. Such cognition is an essential part of survival by making quick sense of the world around us. So, when we feel like we’re really picking out the 24th bit down there in the noise, as a function of those superdetailed reverb trails and awesomely evanescent images floating there in our mind’s ears, we may very well be just making it up. In any case, we are almost always guessing when we are trying to pick out these really small differences.
But there’s more to this. Along with the imaginary, there’s the real stuff, too. We hear things remarkably well, and some of my golden-eared brothers and sisters can hear some amazing audio details. I could tell you stories … .
However, exactly what
it is that we’re hearing is a different question. Some of it is relevant to hi-res audio, some of it isn’t. As reader Dave Riddle pointed out to me a couple of months back, when we glory in the good ol’ analog sound of the ’50s and ’60s (when, believe it or not, there were some really great recordings were made) maybe what distinguishes the sound isn’t the clean, robust analog signal flow at all, but instead the 20-50 transformers hung (at every device’s input and output points) along the signal path. Maybe the sweet fatness of those sounds is all in the transformer windings! STUPID TRANSFORMERS
See what I mean? All this time, we’ve assumed it was the analog purity, when actually it might just as well have been the stupid transformers. In our brave new hi-res world, similar things play. When we compare a 16-bit box to a 24-bit box, we may hear a difference, and that difference may have nothing to do with wordlength. It may be that the 24-bit piece’s manufacturer used a "better" grade of capacitors, or maybe didn’t use capacitors at all, or maybe, even, used transformers
Meanwhile, we say, "Man, doesn’t 24-bit really
clean up the sound? Awesome!"
Sometimes, also, it is in fact the hi-res. I have a colleague who recently told me, "Boy, 24-bit really makes a difference. Mixing is so much easier!"
As we were both drinking at the time, this had the makings of a really interesting discussion. But when we got into it, it turned out he was talking about mixing in a DAW. And he wasn’t just knocking together a voice-track, a music bed and some FX. He was mixing a ton of tracks, all with heavy processing. And so his observation was quite reasonable that the extra wordlength really seemed to make a difference in terms of transparency when he started combining all these gazillions of tracks. He really noticed it when he got stuck with 16-bit resolution in this kind of production environment. DISTINCT ADVANTAGES
Now, my colleague may be imagining it, but he seems sure enough of his ground that I’m inclined to believe he really hears it. And it is reasonable that he should – the additional wordlength does provide some distinct advantages when dealing with the accumulating errors that build up when lots of tracks and processing are all happening at once.
So, it seems to me we can reasonably assume that in conditions where there are massive digital accumulations of rounding errors, processing artifacts and level changes, hi-res is good. I assume also that there may be other instances like this, where the presence of hi-res keeps us out of audio doo-doo that would turn out to be audible, perhaps quite disturbingly audible. LOUD AND CLEAR
There is another instance where I’m also sure we could hear it – loud and clear. And it’d sound great, if we could afford it. If we built up our analog gear and our loudspeakers so that the Least Significant Bit (the 24th one) equaled the analog noise floor (-100 dBV) and the threshold of hearing (0 dB SPL), why then we’d hear this hi-res benefit for sure. Personally, I think it would be fabulous.
Our normal calibrated listening level would be, of course, something like -60 dBFS for 84 dB SPL, and our loudspeakers would require water-cooling and our power amps would have to be capable of generating perhaps a megawatt of power – they’d probably need water-cooling as well.
The enhanced headroom of such a system would be absolutely glorious – and audible, I’m absolutely sure (particularly when we decided, after a couple of double tequila sunrises, to really crank it!). Too bad it’s not practical or affordable.
Next month we’ll wrap this up and talk about the meaning of it all. In the meantime, keep in mind that on Thanksgiving Day as on every day, we are what we eat.