The Audio Engineering Society Convention: An Industry In A High-Resolution Funk

I haven’t yet seen any solid attendance statistics from this year’s U.S. AES (the spring version of the show always takes place in Europe, with the fall iteration alternating on a yearly basis between the Right and Left Coasts), but my eyes and the words I’ve heard from a number of vendors I’ve met with all tell me the same thing…the conference is having a ‘down’ year.

Why? Large music labels are being rocked (pun intended) by declining sales; in response, EMI and Universal (along with, briefly, Warner in a now-shuttered AnywhereCD experiment) have already gambled on discarding the digital rights management which was previously deemed essential to preventing revenue-sapping piracy. And ironically, I’m being told by a number of sources that the healthiest segments (from a unit shipment standpoint, but not necessarily with respect to revenue or, more importantly, profit) of the audio hardware and software markets are the entry-level ends of these businesses.

High-quality hardware is increasingly cost-effective, thanks in no small part to the feature integration and cost reduction efforts of the building block semiconductor vendors exhibiting or otherwise present here (apologies to anyone I’ve forgotten):

• AKM Semiconductor
• Analog Devices
• Cirrus Logic
• Freescale Semiconductor
• National Semiconductor
• Texas Instruments
• THAT
• Wolfson Microelectronics

Software vendors are also ratcheting up their products’ features, in lockstep with the core-multiplication strides of CPU suppliers. Musicians with home-based studios, increasingly bypassing the labels in favour of direct promotion and distribution on sites like MySpace, are the end destinations of all this equipment and code. And the vast majority of it supports large sample sizes and high sample rates, since there’s tangible value to retaining as much meaningful audio data precision as possible during interim mixing and mastering steps.

But high-resolution audio support still hasn’t trickled down into the real volume market, consumer playback equipment. And frankly, with the demise of DVD-Audio and SACD, it may never do so (with the exception of the strident but miniscule audiophile niche). As I’ve suggested any number of times both online and in print, given the way that most folks listen (or, I’d argue, don’t really listen) to music, their desire to squeeze as much content as possible into a given-sized and –cost amount of storage substantially overrides any aspiration for optimum quality of that content.

Look, for example, at today’s tutorial entitled ‘High Definition Perception—Psychoacoustics, Physiology, and the Numbers Race‘, moderated by Poppy Crum. Don’t get me wrong; Crum is a well-known, widely published and technically solid researcher from Johns Hopkins University’s Laboratory of Auditory Neurophysiology in the Department of Biomedical Engineering. And it’s certainly possible to find data suggesting that high-resolution audio has an incremental effect on the human auditory system beyond that of conventional 16-bit, 44.1 kHz material; I pointed out one such study in my 2003 writeup on the subject. However, this is the Audio Engineering Society we’re talking about here. The organization’s members have a vested interest in ensuring that the audio improvement treadmill which fuels regular upgrades in customers’ hardware and software not stall. And in an environment such as this, one must be very careful to not allow predetermined and desired conclusions to unduly influence the testing methodology and the analysis of the amassed results.

Even if the data and the means used to collect it are flawless, the conclusions from the numerous studies I’ve examined over the years are so subtle as to only be discernable with stringent A-B comparisons using high-end playback equipment installed in anechoic chambers and other pristine listening environments. None of this even remotely approaches real life for the masses. Which tells me that it has absolutely no chance of ever going mainstream. Your thoughts, folks?