Today, we’re at an interesting point in audio history, a point where we are seriously moving toward both increasing the resolution of our recordings and increasing the number of playback channels. About 15 years ago, digital audio came into wide use. Hyped as “perfect,” it offered us some nice benefits, including low cost, easy editing, and high-quality storage. In the interim, we’ve noticed that it isn’t perfect at all, and numerous proposals and schemes have been touted to correct that deficiency, mostly by increased word length and sampling frequency.

Meanwhile, about 40 years ago, stereo came on the scene. Hyped as “bringing the concert hall into the living room,” it really enhanced the listening-to-records experience. Stereo is now a mature, well established medium that is at least moderately well understood. However, we’ve noticed that multichannel surround (as used in movies) can effectively improve on stereophony when it’s done well.

Well, these two facets of audio history are colliding today. While we are getting all lathered up about how crummy our 16-bit 44.1 kHz. really is and seriously moving toward more high-resolution production and release formats, we’re also getting the hots for multichannel surround recordings. In a perfect world, this wouldn’t be a problem. Hi-Rez Surround Would Rule! But in our less than perfect world, we’ve got a conflict between these two diverging paths. We’ve got, sadly, data limits.

Let’s consider hi-rez stereo first. Because 16-bit sucks (as everybody knows), and because maybe we can hear way higher than bats and dolphins and there’s all this really cool ultrasonic music up there, we, like, definitely need longer words and greater bandwidth.

Or do we? I just got back from AES Amsterdam, where I attended a meeting of the Psychoacoustics and Subjective Evaluation Committee (I’m a member, unlikely as that may seem). Well, the Psychoacoustics Committee received a proposal for what its authors call Advanced Digital Audio (ADA). This is to be a format for the next 20 years, and its specs are pretty wild: 100 kHz. bandwidth (at least 200 kHz. sampling rate) with 24-bit resolution in the digital realm and the equivalent of 20-bit resolution (120 dB) in the analog realm. They mention multichannel as a remote possibility, but their default position is that stereo is to be preferred.

Right now, 16 bit/44k stereo audio requires 10 megabytes per minute (70 minutes on a CD). Hi-rez stereo will require 70 megabytes of data per minute, therefore yielding 10 minutes of music on a conventional CD, or 57 minutes on an audio-only DVD.

Meanwhile, many surround recordings to be played through one’s home theater system are in release or production. If we were to load 6 channels of discrete hi-rez 24-bit/200 kHz. audio onto a CD, we’d get 24 seconds of music! On a 4 gigabyte DVD, we’d get about 18.5 minutes. Not enough! The resolution might be phenomenal, but not for long! Something has got to give.

My feeling is that it is resolution. We’ve worked ourselves into a frenzy about resolution without really carefully considering what it means to the end user. When we say 16-bit sucks, and 44.1 kHz. really screws up the 10th octave, we are really talking loosely. In the Psychoacoustics Committee meeting I mentioned above, somebody politely asked, “Er, does anybody know of evidence that we really need a bandwidth of 100 kHz.?” We all knew of one Japanese study from 1991 (most of us have some doubts about it). None of us (including a couple of world authorities) knew of any other scientifically robust research that suggests such a need. Think of it! A collection of experienced audio psycho-nerds sitting around and none of them can think of a valid study justifying 100 kHz. And we want to make this a standard?

Think some more. 24-bit audio yields, in theory, a dynamic range of 144 dB. Such a range is considerably greater than we can listen to safely. If we anchor the bottom of it at 0 dB SPL, the threshold of hearing, the maximum level (144 dB SPL) is seriously dangerous and illegal. Further, if we anchor the bottom of that range electrically at 10 microvolts (-100 dBV), which is about as low a noise floor as we can get in studio practice, then the maximum level will be around 160 Volts (44 dBV), which presents serious safety issues for people like Underwriters Laboratories (patching will become very exciting!).

At the same time, there IS clear evidence, drawn from Floyd Toole’s loudspeaker research, that stereo enhances the quality of poor loudspeakers, in comparison with mono. There is further, if less definitive, evidence that multichannel systems have a similar enhancing effect relative to stereo. I have found that, from a listening enjoyment standpoint, more channels/speakers sounds better! More importantly, I have found such enhancement to be huge, compared to the differences I hear comparing different bit-rates, A/D converters and so on.

So, if I were king, I’d throw our data resources toward more audio channels. I suggest that, for all that we like to claim that it sucks, the 16-bit 44.1 kHz. digital audio standard we now use is really pretty decent as a release medium, and only a tiny fraction of the recordings we release challenge its limits at all. And five or six 16/44 channels, at their best, sound totally awesome.

It’s worth thinking about. Thanks for listening.