Is There a Future for Hi-Rez?
These were among the questions addressed by a panel of audio industry luminaries at a discussion entitled High Resolution Audio in an Age of Universal Playback, held the first day of the 117th Audio Engineering Society convention, October 28–30 at Moscone Center in San Francisco. Chaired by audio signal processing consultant Vicki Melchior of San Anselmo, CA, the panel included Stereophile editor John Atkinson, Professor Malcolm Hawksford of the UK's University of Essex, engineer Hajime Kawai of Texas Instruments, Japan, and renowned record producer/engineer and equipment designer George Massenburg.
Ms. Melchior launched the well-attended two-hour event by noting that the industry's response to the recent high-resolution and multichannel format proliferation has been to develop universal players. As a result, this Pandora's box of formats has not only caused plenty of confusion among consumers, but has presented many challenges to equipment designers eager to deliver good-sounding universal players at affordable prices. "Is true high resolution even possible in affordable universal players?" Melchior asked. "Or should it be dedicated to the high end?"
These were not trivial questions. Hawksford reminded the crowd that "the existence of high-resolution at all is fortunate," as he outlined theoretical reasons why higher sampling rates offer "better, more faithful" reproduction of captured waveforms, including accurate playback of supra-audible high-frequency harmonics, a subject that continues to provoke disagreements among audio experts. (Hawksford included photos of add-on "super tweeters" for those who must extract the absolute last milligram of detail from their recordings.)
After outlining the basis of DVD-A, a linear pulse code modulation (PCM) scheme, and Super Audio CD, a sigma-delta modulation scheme whose "noise shaping transfer function" (NTSF) pushes noise in the audible band up into the supra-audible region, Hawksford opined that standard CD when "properly processed" may be the better playback format simply because its digital datastream is accessible.
Digital rights management (DRM), the technological implementation of copyright concerns, is crippling high-resolution audio in the consumer market, although many manufacturers were demonstrating 24bit/96kHz and 24/192 processors, DACs, and ADCs at AES, and Sonoma rolled out a 24-track DSD work station. Thanks to pressure on legislators by the entertainment industry, high-rez datastreams can be accessed and manipulated in the studio, but not by music lovers at home.
Regarding SACD vs DVD-A, Hawkford stated that SACD could be better in lower-priced equipment, but that "cost-no-object gear may favor DVD-A." In either case, "bass management is a major pain," he stated to sporadic applause. Hawksford also left little doubt as to his feelings about SACD releases sourced from PCM recordings: "They should be banned!" File under the heading "Weapons of Music Destruction."
In her introductory remarks, Melchior had mentioned that deriving the utmost from CD, SACD, and DVD-A requires a dedicated player for each format, a reality that was reiterated by John Atkinson. While dedicated (and very expensive) players may be professional necessities for engineers and audio magazine editors, they are unattainable extravagances for ordinary music lovers.
Hence the rush to market of affordable universal disc players, machines that can detect and play all formats: DVD-Video, DVD-Audio, SACD, CD, MP3, and even JPEG images. These players fulfill what Melchior described as "user transparency," but often fail to deliver performance sufficient to highlight the differences between high-resolution recordings and normal CD. "Inexpensive players lose resolution due to a higher inherent noise floor," explained John Atkinson, noting that one inexpensive machine he had examined "had a very noisy power supply and lots of spurious noise." On such machines, SACD or DVD-A "won't be any better than CD," he emphasized.
Compounding the problem is the fact that the dynamic capabilities of high-resolution formats exceed the acoustical limits of many real-life recording venues, Atkinson explained. Texas Instruments' PCM 1792 universal high-rez chip, as described by Hajime Kawai, is capable of a theoretical 146dB dynamic range when processing 24-bit data, yet distance-miked measurements of a large hall that Atkinson provided set the acoustical noise floor at a lower figure, particularly at low frequencies—see the graph below. For this and other reasons, CD's 16 bits "may be good enough for a playback delivery medium," he speculated, though for recording and signal processing the more bits the better.
It's not just ordinary music lovers who sometimes can't tell the difference between high-rez and medium-rez. JA mentioned one of the magazine's reviewers who swore he couldn't hear a difference between the CD and SACD versions of Norah Jones's "Cold Cold Heart," using very good dedicated players and excellent ancillary gear. A waveform analysis showed why: They were identical except for the usual SACD noise added to the top end. Further exploration uncovered the fact that the SACD had been made from the 16-bit/44.1kHz-sampled PCM master rather than from the original analog tape.
Atkinson followed this revelation with a chart of the dynamic range of a track from Los Lobos' recent The Ride CD, showing abysmally low dynamic range. The group's Kiko, by contrast, was everything a great recording should be, "with air and space and tremendous feeling," he stated.
Such musical mutilations are often the result of budget constraints and marketing executives meddling in the recording and mastering processes. As revealed by the recording industry itself, high-rez discs in both formats sold only about 600,000 copies combined in the US in the first half of 2004. Without consumer demand, both of them are likely to fade away.
To prevent that, Atkinson asserted, music lovers have to perceive that features (such as multichannel) are also benefits. But in the face of inexpensive universal players flooding the market, he notes that high-rez multichannel and stereo playback through cheap universal machines "may not be better than CD, but could be worse," a comment that provoked the greatest applause of the afternoon.
In fact, said Atkinson, the real benefits, in the form of DRM schemes, actually accrue to the record companies. This has led to the lack of "portability" with the formats, a primary benefit that consumers demand. "Record companies are digging their own graves with DRM," said Atkinson. "Why are they treating their best customers, like me, as potential thieves?"
SACD and DVD-Audio DRM also prevents Atkinson easy access to disc data for his format testing, a barrier that clearly irritates him.
George Massenburg offered the studio and professional point of view. He enthused about the ability of high-resolution recording to capture "small sounds and localization cues" that add to the realism of playback, but wryly noted that the neurological and psychological processes of hearing, hearing music, and interpreting what we hear are still not well understood. He speculated that we may have reached an impasse in our efforts to improve recording and playback technology until more research is done on the fundamentals of the internal human experience, one that equally involves "ears, brains, and hearts."
Massenburg also brought up what he called "the MP3 paradox"—the fact that while technology has advanced to the point of offering very high quality to millions of music lovers, most of them have embraced low-resolution instead. He compared the enormous success of DVD-Video—the film industry has shipped more than three billion DVDs to retailers since the format's inception a few years ago—to the underwhelming market penetration of high-rez audio.
Massenburg said that despite improvements in digital audio, and the fact that engineers and producers "care passionately about preserving the original sound," high resolution has costs that often can't be justified economically.
Consumer demand creates marketing justifications. A couple of strategies that would help tip the scale for high-resolution audio, and for good-sounding recordings in general, would be to get "artists to listen to their own recordings" and to get the record industry "to stop making recordings for radio"—in other words, recordings so tightly compressed, as in JA's Los Lobos example, that they sound "loud" over the background noise in a moving car, the typical radio listening environment. The new generation of ultra-quiet vehicles with good sound systems may help change that, but perhaps not in time to save SACD and DVD-A from the Grim Reaper of economic reality.
"So then why do we persist even when the market tells us it doesn't matter?" Massenburg asked. "Because once we hear something and we like it, we miss it when it's taken away."