pf logo

audiodiscourse.jpg (10290 bytes)


In Which PF Does MoFi, hears GAIN 2 & DSD, and returns to River City very impressed!
by David W. Robinson

(From the PF archives of significant articles, 1998)


mofioutside1.web.jpg (18660 bytes)

On Friday, September 25, Positive Feedback visited Sebastopol, California, the headquarters of Mobile Fidelity. At the invitation of MoFi, Scott Frankland and Ye Olde Editor attended a presentation of the new DSD-based GAIN 2 mastering system. MoFi will debut GAIN 2 with its reissue of the Tom Petty Full Moon Fever in the next few weeks, according to MoFi’s Director of Public Relations, Karen Thomas.

Anticipation and curiosity have been increasing in audiophile circles over the past year concerning higher resolution digital standards. Both DVD-based 96 kHz/24-bit PCM and the 2.8224 MHz/1-bit DSD have emerged as the alternatives for the next generation of digital sound. Both offer significant advances over the current 44.1 kHz/16-bit CD standard. The question has been whether one or the other system would become the market "standard," or whether both might co-exist—a question that has yet to be answered.

Proponents for both standards exist. Companies like Classic Records, Chesky Records, Cardas Audio, Theta Digital, Muse Electronics, Cal Audio, Ayre Acoustics, conrad-johnson, and Resolution Audio have committed to 96/24. On the other hand, DMP and Telarc have announced their preference for DSD.

To the DSD list you may now add Mobile Fidelity and its GAIN 2 mastering system.

Why DSD? "We listened carefully to the alternative designs in 96/24," explained John Wood, MoFi’s Director of New Technology, in a telephone interview with Positive Feedback. "Every 96/24 system was good—much better than we currently have�but we still felt that there was some coloration in current implementations. When we checked out DSD, we felt that we had found what we had been looking for."


timnishio.web.jpg (12831 bytes)
Tim de Paravicini and Ayataka Nishio chat at
the lobby of Mobile Fidelity just before the conference

But there was much more to GAIN 2. In order to assure a first-rate source for mastering, MoFi had commissioned the noted analog(ue) artist Tim de Paravicini to completely rebuild their Studer A-80 � inch reel-to-reel, and bring it up to his very high standards of performance. Digital received an equally royal treatment: Ed Meitner, the gifted digital designer, was chosen to do the A/D and D/A side of the system. Sony’s Ayataka Nishio, the chief designer of Direct Stream Digital (DSD), worked closely with the MoFi GAIN 2 group to integrate DSD smoothly into the new mastering chain.

With such an intriguing collection of talent and technology, PF immediately accepted an invitation to attend a round table discussion and listening session at MoFi on September 25. (Who could resist?!)


karen.web.jpg (13366 bytes)
Karen Thomas, MoFi's Director of Public Relations

Upon arrival on Friday morning, Scott and I met de Paravicini and Nishio. Over bagels and fixings, plus some superior coffee from Karen Thomas (credit where credit is due!), we chatted about Tim’s achievements with MoFi’s Studer, which were apparently quite spectacular in the analog(ue) domain. Tim’s enthusiasm and passion for the work there was obvious. Nishio was pleasant, urbane, and reserved; only later did we get a chance to discover the mind behind DSD.

The President of MoFi, Lori Beaudoin, came in to greet us all. John Wood and Shawn Britton, MoFi’s Mastering
Engineer, also arrived, together with Ed Meitner, bringing us up to complement.


loried.web.jpg (11222 bytes)
Lori Beaudoin and Ed Meitner just before conference time

Once everyone was ready, the group moved into MoFi’s conference room. There Lori Beaudoin opened the proceedings by formally introducing the members of the MoFi group to PF. She made it clear that MoFi was not in the business of establishing audio standards or "pushing a format—that’s not what we do. We look at them, we assess them, and then once they’re established we decide if there’s anything we can do to improve them to bring better quality sound to consumers."


johnednishio2.web.jpg (12355 bytes)
John Wood, Ed Meitner, and Ayataka Nishio
at the conference table

John Wood outlined the history of MoFi’s interest in improved digital audio over the past several years. "We knew that high density digital audio was on its way." Pioneer’s high resolution DAT machine made an impression as far back as three years ago.

So did "the buzz with DVD audio. So Shawn and I started thinking about what we could do to better the system. We initially wanted to start out with the analog portion of the system, and we wanted to see if there was more information that we could extract from the tapes. We ran into Tim at one of the shows and talked to him; he said he thought he could deliver something that would be pretty extraordinary. We weren’t really sure if he was fibbing or not! (general laughter) But we went ahead and commissioned him, and he went to work."

"At the same time, we met with Nishio-san and David Kawakami, and read some articles about Sony’s DSD, and we decided that we wanted to investigate this technology. Shawn and I went out and checked one of Sony’s show demos at the HiFi Show in San Francisco, and were impressed with what we saw and heard. So we decided to move forward with some experimentation with the Sony DSD system."

Shawn Britton noted that "we literally ran into Tim de Paravicini in the halls at the same show. We talked to him about his tape machines—we knew of his reputation, of course—and asked him if he would be willing to not only modify our tape machine but also build the repros with the incredible high frequency response that we would need. Given the fact that DSD would reproduce an astounding frequency response, we thought that we’d better talk to someone who could give us the frequency response that we would need."

"Right! And make damned sure that we extracted everything that we could," continued John Wood. "Another thing was that we weren’t really sure what the ‘outer limits’ were with 15 ips quarter-inch tape."


shawnscott.web.jpg (13132 bytes)
MoFi’s Mastering Engineer Shawn Britton
and PF’s Scott Frankland following the discussion

Shawn Britton made the point that "Most of the gear that we’ve had and had built for us were designed around a frequency response of 20-20K. That’s pretty much how people have been designing things for a long time—never pushing it. When you see the demo that Tim does for you, you’ll see the result. We have some rather astounding frequency response now!"

"As far as DSD goes, we met up with Nishio and David Kawakami and were really interested in pursuing it," John continued. "David was kind enough to send a system over so that we could demo it. We did—and we were really impressed! I saw David again at Winter CES, and he was going to see Ed Meitner … "


johnednishio3.web.jpg (14678 bytes)

"Oh, that guy!" said Shawn. (general laughter)

"… and see if we could woo him." John continued, "From what we knew, Ed had previous experience with the delta sigma stuff, so that was why we were interested in him."


johnednishio2.web.jpg (12355 bytes)

"As a sidebar to this, there was only one down-converter machine in the world for taking pure DSD signal and converting it to 16/44.1. Nishio had that in Japan. We sent him tapes right after CES. The DSD downloads to a tape called an AIT, which we sent to Nishio; he down-converted it and sent back a sample to analyze. Shawn and I analyzed it; it was very promising."

"Meanwhile, we played with the DSD, and kept running different kinds of program material through it. Tim was working on the analog repro system, and we kept thinking, ‘Jeez, we’d sure like to get more information out of these tapes!’ Shawn kept calling Tim and saying ‘Tim! Tim!’"

"We made the connection with Ed, and Ed’s working on an A/D and D/A—so we approached him and asked if he could do something for us. And that’s how this all started."

Shawn Britton interjected some notes about DSD as a mastering engineer. "We found that DSD captures more of the ambient detail, more of the room. That’s what we were struck by. And we found that the Super Bit-Mapped (SBM) Direct, the down-converted format that Sony has developed, and we were impressed. We had the DSD system here for comparison, and we have some of the best A/D converters for the PCM stuff that have ever been made."

"We were pests! When you get excited about a new format or new gear, you want it! That’s why we wanted to press ahead with this new technology; we knew that we could have 16-bit/44.1kHz with improved resolution right now!"

John Wood then passed the baton to Tim de Paravicini. He spoke of his background and experience beginning with studio work in the ‘60s with rock and roll bands, then going on to Japan with various projects for Luxman in the ‘70s.


timkaren.web.jpg (11301 bytes)
Tim de Paravicini waxes eloquent;
Karen Thomas checks his technique!

"Throughout this period, I looked at all the recordings being made, from the early ‘51-‘52 when tapes first started to be used for recording music. I came to the conclusion that there was an awful lot of mythology in what people had believed tape recorders can and cannot do, and where is the weak link, and so on. So I went sifting through virtually all the tape recorders made, whether it was Telefunken, Studer, Phillips, Ampex, and looking at their record and repro capabilities. And I very quickly came to the conclusion that people had been assuming that they couldn’t record very well but that they could play back this master tape, and you’re going to get perfection. And I said, ‘I’m sorry, but that is wrong.’"

"I did these checks and found that most of these recorders could actually record quite well; what they couldn’t do was pull it back off the tape! Even tape machines as early as ‘51-’52—the early Ampexes—could record signal below 20Hz quite well… say better than 10 dB down at 10Hz… and go out and get information beyond 20K. It may not be ruler flat, but they could put information out to 30+kHz; whatever the microphones and all the rest of the system could provide it with. But what all these machines suffered with—and Studer was a good example of this—was cheapskating on the playback electronics. It’s a much harder task to get playback right."

"… Anyway, I’ve been building since the ‘70s tape recorders with very wide performance." Tim discussed the work he did at that time, rebuilding cassette recorders so that they would have better frequency response. "Even into the ‘80s people making tape recorders didn’t understand; were they trying to save ten cents to throw a ton of (good) shit out of the window? How much cost cutting have they got to do? Or is it just plain ignorance? And I came to the conclusion that most of it was just ignorance; they didn’t understand what they were trying to achieve. With very little extra cost they could make their machines do a hell of a lot better performance."


timscott.web.jpg (11674 bytes)

"In listening qualities, most tape recorders have what I call ‘severe phase distortion’ in the low frequencies. So the rhythmic properties of tape no longer sound correct." Tim analyzed the mistakes made in the mastering of the Sgt. Pepper’s CD—problems in the roll-off and phase inaccuracies in the lower frequencies causing problems in pace and rhythm. He especially emphasized errors in phase: "If the phase properties are badly skewed, the rhythmic properties of this music sounds wrong. So that’s one of the things that makes you think, ‘It doesn’t sound like a master tape!’"

Tim de Paravicini’s standard for analog(ue) reproduction is the mic feed, "… the line-in quality." He stated that his 1" tape machines were able to attain that standard; his � inch machines were almost as good.

"All I wanted to get electronics that—with new heads—could just pull these tapes. And when you listen to some of these old tapes you’re just awe-struck; you’d think they were done yesterday. The point is, all these re-issues of records that most companies put out are never going to solve the problems, because they haven’t addressed how to get this information out of what’s in that material (the tapes). Once you find it, you realize that is the line sound that these guys were hearing in the ‘50s and ‘60s and putting down on tape. The machine could engrave it quite happily."

The discussion then shifted to a consideration of Ed Meitner’s A/D and D/A designs for the GAIN 2 chain.


ednishio.web.jpg (12066 bytes)

"We did some delta sigma work way back with dbx," Ed remembered. (PF Associate Editor Mike Pappas recalls this system from his days with Otari in the early ‘80s, and considered it to be the best sounding of the early digital systems they tested at that time.) "Compared with what delta sigma is today, it was kind of primitive. But it showed the possibilities."

"The problem in those days was that there wasn’t any really any storage medium that could store the amount of information delta sigma would generate. But we played with it, and we modeled, and that’s when it started."

"When Tom Jung (of DMP) and David Kawakami (of Sony) phoned me and said, ‘Hey, you want to do some work on DSD?" I said ‘Hey! Great! Why not?!’ This is more interesting than carrying on with 44.1 or 96/24 or whatever."

Meitner briefly summarized the work that he had done in developing the A/D and D/A interface, then—with time being very limited at that point—the discussion shifted to Ayataka Nishio’s work with DSD.

Nishio responded to one question by confirming that Sony’s plans for the Super Audio Compact Disc (SACD) were for a two-layer disc. The top layer would be for a new form of Super Bit Mapping: "Super Bit Mapping Direct." It is a full 16-bit PCM digital word system compatible with standard CD players, produced by down-converting a DSD master to SBM Direct.

The lower layer would be the DSD high-density layer, with a capacity of 4.7 gigabytes. Nishio made it clear that not only would DSD support two channel stereo at 192kHz per channel, but that the DSD SACD specification would also included six channels of equally high resolution sound for home theatre/surround sound. Text, graphics and video would also be supported on the DSD layer.


sacd.web.jpg (16489 bytes)

The DSD bitstream sampling rate is 2.8224MHz, giving a sampling rate of DC-100kHz per channel on all channels, with a dynamic range of 120 dB.

When asked about the availability of editing tools and a storage medium for DSD, Nishio confirmed that Sonic Solutions is developing a version of SonicStudio for DSD, and that storage via tape and hard disk is "not a problem." He outlined the many advantages of DSD, including its simplicity of concept and the avoidance of filtering and requantization. The unique dual layering of SBM Direct and DSD should provide consumers with maximum performance without the loss of 44.1kHz/16-bit format compatibility.

John Wood and Shawn Britton re-iterated the fact that despite the future possibilities for SACD, the combination of DSD mastering and down-converting to SBM Direct at Mobile Fidelity would happen now. Consumers wouldn’t have to wait to begin to benefit from the advantages of DSD; GAIN 2 will bring it to them now. (The first release in the new GAIN 2 series will be a particular favorite of Ye Olde Editor’s, Tom Petty’s Full Moon Fever. Bring it on!!

The Studio: Listening Notes

The MoFi/PF group then moved to the mastering studio to listen to GAIN 2/DSD/SBM Direct. (15823 bytes)

First, Tim de Paravicini demonstrated his newly re-built Studer A-80 � inch 15 ips/30 ips analog(ue) front end. The Studer was connected to both an oscilloscope and to a phase detection array, so that we could monitor output (via VU’s), and proper phase (via ‘scope and array). (18092 bytes)
The remarkable Tim deP Studer

Scott Frankland and I then witnessed an extraordinary display of open reel performance. Tim ran the unit through a series of frequency tests: first from 1kHz down to 10Hz (!), then from 1kHz out to 44kHz (!!) The VU’s indicated no less than 0 dB at any point, and no more than about 1 dB at any point (!!!) Phase remained proper at all points. As far as we could see, minor fluctuations seemed to remain within the limits of the open reel medium itself. (17934 bytes)

This represents a truly remarkable achievement: an open reel machine that is within a dB from 10Hz – 44kHz is utterly extraordinary. It holds the promise of being able to deliver every nuance from master tapes to the balance of the GAIN 2 mastering chain, without compromise. But—how did it sound? Did listening confirm the promise of the measurements? (15697 bytes)
In spades!

Tim and Shawn cued a Coltrane master tape. Scott and I were allowed to sit in the mastering engineer’s sweet spot. The mastering chain went from the Studer, through the custom panel, to outputs wired to a pair of Nelson Pass’s fine amplifiers. The cabling was by Sahuaro Audio; the power line conditioner was the Sound Application CF-2; the speakers were the top-o’-the-line Egglestonworks with custom stands.

The sound was nothing less than the finest that I’ve ever heard from an open reel machine! (And yes, I’ve listened to other Studer A-80’s before.) The noise floor was imperceptible; the signal-to-noise ration had to be excellent. Scott and I were both quite struck by the ease of the music as we listened—it simply flowed from the speakers like water.

Let the reader note: The danger of listening to music at this mountaintop level is that it can make it hard to go back to the valleys.

(On the other hand, one should calibrate one’s ears to the highest standard available—it keeps you from raving over the mediocre.)

Given the world-class quality of the GAIN 2 Studer machine, Scott and I were eager—and fearful—to hear what effect the Sony/Meitner DSD A/D & D/A, together with SBM Direct would have on such glorious sound. (11721 bytes)
A rare beast: The Sony DSD station in the studio at MoFi

Mobile Fidelity made the mechanics of the comparison easy by setting up their I/O controls in the studio so that Scott and I could soft-touch our way back and forth among "Studio" (the Studer), "DSD" (feed the Studer to the DSD electronics and Meitner A/D and D/A), and "D/A" (feed the Studer to the SBM Direct down-converter). All digital feeds would happen in real-time. The three selections were level-matched to within .1 dB for valid listening comparison. (16967 bytes)
Ed Meitner’s custom A/D and D/A for DSD
at Mobile Fidelity

Scott and I listened to the master tape for a few minutes… then I leaned forward and hit the "DSD" button.

Did I hear a change? Maybe I heard a change—maybe not—time to toggle back to the Studer. Listen again very carefully to the sound of the brush on snare drum… switch back to "DSD"… nope, sounds right. Go back to the Studer… listen to the saxophone, savor the feel of saliva on reed, place it in space, then go back to "DSD." That sounds right too. Go back to the Studer, listen for soundstage depth, imaging… go back to "DSD" and listen carefully for shifting placement of vocals, drum kit, back wall… no, that sounds right too!

Scott spent some time going back and forth; each time we listened for different elements of the recording. Each time, careful listening showed that DSD was putting us so close to the master tape that I was finding it virtually impossible to fault any aspect of what I was hearing.

I must emphasize the extraordinary nature of this result. Scott and I were free to switch back and forth. We are both very experienced listeners. We were not comparing "vinyl with digital"; we were listening to a master tape on (arguably) the finest open reel machine of its kind in the world, a Studer whose remarkable performance characteristics had just been demonstrated to us. The source Studer was fully capable of playing back 10 Hz – 44 kHz to within a dB, and was phase correct across that range. The room was extremely quiet; there were only six of us present (no terrible "show conditions" to intrude on the assessment of the sound).

I would not be willing to stake a nickel—much less my life—on being able to exegete the difference between the Studer source tape and DSD. While high-end audio, like football, is often a "game of inches," I would have to say that Scott and I found ourselves so close to the analog(ue) source that we were unable to characterize the minute differences, if any, that we might have been (?) hearing.

Which puts the whole experience into the realm of the truly outstanding…

When we switched to the SBM Direct loop, we did notice some losses in the listening experience. I would characterize them as a foreshortening of the depth of the soundfield, a loss of the sense of presence and atmosphere, a slight shift in imaging towards left and right, and an increase in harshness of sibilants. The sense of "rounded sound" was no longer there, and the remarkable ease of the master tape, held astonishingly well by the DSD feed, was, to a great degree, surrendered.

Not that the SBM Direct was "poor"—it was just lesser. It is by far the best SBM that I’ve ever heard, which is something notable; I have to confess that I had not been impressed with SBM at all until now. The DSD-based SBM Direct moves SBM to an entirely different level, and makes it quite acceptable as the basis for the GAIN 2 system.

We had a further chance to listen to SBM Direct when we compared a "standard issue" of Tom Petty’s Full Moon Fever with a brand new copy of the MoFi SBM Direct version.

No contest. The SBM Direct won hands down, and going away. More depth, more spaciousness, better timbre on the luscious sound of Petty’s Rickenbacker, less "tizziness" on the sound of cymbals in the drum kit, more fully developed bass… superior in every way.

In Retrospect…

The listening experience at Mobile Fidelity was a fine opportunity to meet a unique and gifted group of people, and to hear the results of the work that they’ve done. Many thanks to Lori, Karen, John and Shawn for inviting Positive Feedback to this unique demonstration. (13984 bytes)
The guys: Shawn Britton, Ayataka Nishio,
Ye Olde Editor, Ed Meitner, John Wood,
Tim de Paravicini, Scott Frankland,
after a helluva good morning

GAIN 2 will clearly be an improvement over the current GAIN system; MoFi fans have nothing to worry about. The combined work of de Paravicini, Meitner, Nishio/Sony, and the management group at MoFi has produced a definite step forward in the quality that we can expect in the future from their new GAIN 2 system.

It is very important that the reader should note that MoFi has also made it clear that they have not ruled out parallel work with 96/24 PCM reissues; they simply want it to be better than it currently is. To that end, MoFi is working with Theta Digital to see if 96/24 can be brought up to the same level of transparency as the DSD standard.

In other words, the next high-density, high-resolution digital standard may not be a zero-sum game.

And what of the future of digital?

The listening that Scott and I did at MoFi confirmed the suspicion that I’ve had for over four years now: the problems that audiophiles have with digital have nothing to do with "digital." The fault lies with the 44.1 kHz/16-bit standard—often very poorly implemented, with mediocre audio engineering/mastering aiding and abetting the crime.

After the listening session with DSD at MoFi, as well as spending a great deal of time with my current evaluation of the first of the 96 kHz/24-bit PCM DVD units at PF, it is clear to me that digital audio has far greater possibilities than we have realized so far. 

We must move to the highest possible density and resolution before we will begin to see the strengths of digital match its mere convenience. Positive Feedback therefore supports the movement to 96 kHz (or higher!)/24-bit PCM and SACD at the earliest possible moment, and calls for re-mastering and reissues of all of our great stereo masterpieces before the analog(ue) masters decay and are lost forever. Our audio archives MUST be preserved for the next century. High density digital, properly implemented, has the capability of doing just that—and with a transparency and faithfulness to the master tape that is breathtaking.

It’s time for something better.