You are reading the older HTML site

Positive Feedback ISSUE 25


A Response to Steve Lefkowicz
by Lynn Olson


Regrettably, I have to agree with Steve—and I've been a big fan of high-res formats (all of them!) for the last 30 years, going back to my first purchase of a Sheffield Direct-to-Disk.

To step back, we need to look at the formats used in professional recording, not just the consumer format. I visited the Super Audio Center in Boulder (Colorado) a couple of weeks ago, where Gus, the resident engineer, demonstrated surround DSD on the mastering system. See pix at:

The sound, as you might expect, was nothing short of spectacular, true high-end, the real deal, and in flawlessly mixed surround with perfectly even image quality throughout the room. A memorable demonstration, and I hope the Super Audio Center sets up a demo at the upcoming Rocky Mountain Audio Festival. After the demo (three different recordings, all very impressive) we talked about various aspects of the recording industry—I'm an outsider, and I was curious about what's considered current best practice these days (if you have a strong stomach, there's an article about worst practice here:

Well, sorry to break anyone's heart, but analog recording in the studio is dying. The 2-inch analog multi-track tape machines are getting harder to find, tracking down experts who can do the finicky—and required—regular alignments is getting difficult, and now the tape itself is becoming scarce with the shutdown of Quantegy. There are a few—a very few—recordings made on purist 2 and 4-track machines from the dawn of tape recordings—museum-quality Ampex 350's and their successors—but these are for a extremely small specialist market. The vast majority of recordings made in the last ten years are made on 24 and 32-track computer-based digital systems, typically with 96/24 resolution, with 192/24 slowly displacing 96/24 in high-end studios. The $65,000 Sonoma system that I heard at Super Audio Center is apparently the first, and so far only, multi-track DSD system.

I've noticed that the reverb on SACD/DSD recordings has a noticeably different character than the main DSD sound, and Gus confirmed my suspicion—the only available reverb, even for DSD, is 96/24 or 192/24 PCM, which is then transcoded and mixed in to the DSD tracks. DSD reverb is still on the drawing boards. I asked about the EMT reverb plates that were so popular in the late Sixties and all through the Seventies, and unfortunately, they share the same fate as 2-inch multi-track recorders—long out of production, hard to maintain, and the people who maintain them are getting old. The pro world is basically 99% high-res PCM, and has been so for more than a decade, as the last analog equipment has been retired—this includes analog mix boards, as computers have gotten more powerful and the disadvantages of repeated conversions in and out of the digital world became apparent.

The consumer distribution formats are RedBook 44.1/16 PCM CDs, a niche market in LPs (mostly for dance music, not audiophiles), very small markets in DVD-A and SACD, and unfortunately, rapidly growing markets in MP3, AAC, and WMA compressed-digital formats. Whether we like it or not, that's the reality of the situation. Sales of CDs have been declining for a number of years, gross abuses of PCM such as intentional digital clipping is epidemic (to the point where engineers now consider music "noise with a beat"), and the phenomenal success of the iPods and its imitators has demonstrated in the most obvious way that most people really don't care about sound quality at all—millions of people consider 10:1 compression "digitally perfect".

The myth of digital perfection is so deeply embedded that the latest scheme for digital radio—at an appalling rate of 96k/bps—is actually marketed as HD Radio! See for yourself at: Visit the website and see for yourself the claim of "enhanced sound fidelity" and "upgraded audio quality". At 96k/bps! I don't care what miracles are claimed for the latest variant of AAC, that's a compression ratio of 14:1—from the already marginal Red Book rate of 44.1/16.

Looking at the big picture over a long time span, we have very high quality analog magnetic-tape recording from the early Fifties through the early Eighties. After that, there's a steep drop-off with really grim first-generation digital, recorded for all time with abominable 741 and 5532 op-amps, no dithering, plenty of jitter, brickwall filters in the ADC, and the marginal Philips/Sony 44.1/16 standard—which by the way, was substantially inferior to the already established Tom Stockham 50/16 system, which had been in use since the mid-Seventies (and released on Denon records). By the late Eighties, dithering came into wide use—too bad about the entire recording industry being used for beta test, eh?—and bit depths slowly crawled into professional-quality 20-bit, which allowed digital-domain signal processing without gross artifacts. By the early Nineties, Apogee and others developed much more sophisticated dithering and transcoding algorithms, which allowed high-quality professional recording at 88.2 and 96kHz rates, and 20 to 24-bit depth.

We have good recordings made from the Fifties through the early Eighties, a lost decade of permanently degraded recordings (thank you, Philps and Sony, it wouldn't have possible without you), and the present era—where PCM-in-the-studio is at least reasonably good, and a distribution system that is driving the least common denominator—clipped and distorted "radio-ready" CDs using the 20-year-old Red Book format, and even worse compressed digital for Internet distribution.

This is not a happy picture for audiophiles and the high-end industry as a whole. If the vast majority of buyers have shown with their dollars they don't give a damn about sound quality, where does that leave us? High-end audio has always been a niche market, going back more than fifty years, but we still need good recordings to survive. If highly compressed 44.1/16 PCM is the dominant format ten years from now, will there be any point in high-end equipment at all?

OK, OK, I'll lighten up a bit. I know the final death of hi-fi, and for that matter music worth listening to, is an appalling thought, one we should back away from before it sinks in too deeply. But I think we should keep the big picture in mind—how music is recorded, and how it is sold to the public. Sales in Wal-Mart don't matter to us, but the state of record stores does. If record stores don't carry SACD or DVD-A, that matters. If the only LP's they carry are intended for DJ's, that matters too. We may despise Red Book, but think of the alternative—what if record stores go away completely? What then?

I think in these days of Open Source we need to start thinking about possible niches for high-quality music distribution. The high-res content is bottled up in the studio, existing distribution systems are becoming dsyfunctional even for the major companies, and the domination of Red Book has convinced most people that Digital Anything is "perfectly" good enough—certainly good enough to sell zillions of iPods. Actually, if a 96/24 capable iPod came out, I'd probably be the first in line at the Apple Store to buy one—but I'm not holding my breath.

Since we are a niche market—the comparison to Linux comes to mind—maybe we should start acting like one.e.

The user-hostile DVD-A interface—requiring a TV set to even select the correct high-res layer—has to go. If that means making limited- production-run DVD-V's, so be it. They play on *all* DVD players, which are ubiquitous these days, and don't require an annoying TV interface. The SACD folks—yes, I'm talking to you—should forget about limiting DSD recording systems to large studios.

News Flash: the big boys don't care about quality, and never will. If DSD recording isn't available on a $1000 Echo Audio or MOTU 12- channel Firewire interface (, along with a $500 software package, forget it, because I can buy a complete 192/24 multichannel solution that works just fine on either a PC or Mac workstation. Dual-processor computers and RAID disk arrays aren't that expensive anymore, you know.

There's no reason we shouldn't be seeing plenty of audiophile recordings mastered in either DVD or SACD format, and created and duplicated on a PC. People are downloading movies off the Internet, editing them on their computers, and burning them on the DVD burner. Don't tell me 96/24 or DSD is harder to do than video—and any modern computer does HD video, not just SD. The whole mystique of computer audio collapses when generic white-box PCs process HD video with ease. If the SACD people don't offer a competitively priced SACD burner that works with any modern computer, well, that's the end of SACD.

Remember: Niche Market. That means small. That means limited distribution. That means using Open Source methods to bypass Corporate America, which is Not Interested in a handful of audiophiles and a non-youth-friendly demographic. The average $200 million movie is aimed at a 14-year-old male teenager, which is why most of the scripts are based on comic books and video games. Take a good look at the demographic next time you go to the St. Tropez or T.H.E. Show at the CES. See any teen-agers there?

During the transition to a niche market—and I do think it will develop—audiophiles need to be broad-minded about their playback sources. We'll need tweaked ultra-low-jitter Red Book players that extract what can be gotten from 44.1/16, Universal players that do justice to DVD-V, DVD-A, and SACD formats, and a decent-quality phonograph to play the hundreds of thousands of wonderful LPs that will never be made into decent-sounding digital. We need to remember when we wrangle about SACD vs. DVD vs. LPs, the iPod-toting masses just laugh at us, and download another hundred MP3 songs off the Internet.

And we need to keep an eye on PC-based playback as well—hard-disk playback has a lot of advantages over a spinning optical disk.

Lynn Olson is a Senior Assistant Editor for Positive Feedback Online and maintains an excellent fine audio site at