I’m a member of the Consumer Electronics Association Audio Board thanks to my friendship with Ray Kimber of Kimber cables. Over the years I’ve found the regular conference calls and occasional meetings very interesting and informative. During the meeting that we had in New York back in June, the group heard the presentation “Lost In Translation” by Andrew Scheps. I wrote a post about that session but lamented the fact that it didn’t go far enough. So I pitched idea of a CEA backed investigation into the realities behind High-Resolution/High-Definition audio. The CEA head suggested that I write up a proposal and bring it to the board. Every year the CEA does some sort of market survey on behalf of the group and she thought my idea might have some interest. The reaction was very positive.
A few days ago, I pulled out an academic paper that I wrote and presented at the 31st London AES meeting back in 2007. I used the format for the new proposal entitled, “A PROPOSAL TO INVESTIGATE HIGH-DEFINITION AND SURROUND AUDIO WITH REGARDS TO PERCEPTABILITY, PREFERENCE AND MARKETABILITY”. There was an abstract, illustrations and sections on the methodologies to be used.
The major thrust of the paper is to do a rigorous investigation into the anecdotal and unproven claims made by advocates of high-definition audio, which includes me. The AES paper published in 2007 called, “Audibility of a CD-Standard A/DA/A Loop Inserted into High-Resolution Audio Playback” by David Moran and Brad Meyer has been taken by many as proof positive that HD-Audio or high-resolution music is completely unneeded. They concluded, “The test results show that the CD-quality A/D/A loop was undetectable at normal-to-loud listening levels, by any of the subjects, on any of the playback systems.” However, their study was so fundamentally flawed as to make their results completely unreliable and meaningless. Here’s why.
I decided to go back a take another look at the BAS study in order to bolster my case for opening a new investigation. My assertion was that their study did nothing to answer the question that we all want answered. My reasoning is that the list of recordings brought to the study by the membership of the group included nothing that actually contained fidelity greater than that of a compact disc. So naturally, if a listener was submitted to a double blind test where they were expected to be able to detect a “so-called” high-resolution recording vs. the downconverted CD spec version, they couldn’t. They both contain the same frequency contour and dynamics.
So I went to the BAS site where they discuss additional details about the investigation. One of the things that the original study failed to mention was the content that was auditioned during their study. This situation was rectified to some degree with the follow up information on the web site. The vast majority of recordings on the list are SACDs, which immediately rules them out because of the LPF required and the fact that they were created from older standard definition recordings. Of the couple of DVD-Audio discs on the list, I decided I would acquire the 96 kHz/24-bit version of “Two Against Nature” by Steely Dan from HDTracks.com (I’m turning out to be a regular customer!) and do some quick analysis on a track. I did that yesterday morning.
I chose the tune “What A Shame About Me” but all of the tracks have similar spectra. The master tracks were recorded on an analog 2-inch tape machine (which limits the dynamic range to around 10-12 bits according to the 6 dB per bit relationship and the fact that analog tape does about 65 dB of signal to noise ratio). Then they were mixed to another analog tape followed by a mastering session with yet another copy. The music that made it onto the distribution CD format has only 6 dB worth of dynamic range…something that a CD can more than handle. This number is typical in commercially available recordings.
Figure 1 – A spectragraph a “What A Shame About Me” in CD resolution AND “high-resolution”…they are virtually identical. (Click to enlarge)
There is a clear roll off from 22-24 kHz as seen in the right hand plot. The spikes to the right are from the bias of the analog machines involved.
David Moran and Brad Meyer of the BAS Society expected the participants in their study to reliably distinguish between the “high-resolution” version at 96 kHz/24-bits (which in reality has 6 dB of SNR and tops out at 23 kHz) and the CD version that has the same dynamic range and virtually the same frequency response. This is obviously an impossible task. Yet many forum visitors proclaimed this study “irrefutable proof” that high-resolution audio is a waste of effort.
The music industry desperately needs to revisit this question with material and playback systems that actually deliver real high-definition audio. It needs to be a double blind test of as many participants as possible…I’m proposing 1000 participants.
I’m revising the proposal per the directions of the group and preparing a budget. If the study gets funded, perhaps we’ll know once and for all the truth about high-definition audio recording and formats.
We’ll see where this leads but it’s a least being discussed.