[ TNT | Who we are | HiFi Shows | HiFi Topics | Listening tests | Tweakings | Inter.Views ]

Inter.View with Bob Katz, co-owner and Mastering engineer of Digital Domain, audiophile recording engineer and former Technical Director at Chesky Records

by Lucio Cadeddu

[Italian version]

LC >
Would you please tell our readers what Digital Domain is and when (and why) did you start this project?

BK >
Digital Domain is a CD (and DVD) mastering studio. I founded Digital Domain in 1990, but I'd been doing mastering for many of my clients for many years before that, since I became a professional recording engineer around 1971.
I used other people's studios with a lot of my own equipment brought in.

LC >
We have published a comparative article of various audio formats: new 24/96 DVD vs 16/44 CD vs 45 rpm LP, all from the same master and the results have been quite encouraging.
Have you done a similar test or, in general, how do you consider the new 24/96 digital format?

BK >
I've done many such tests, and more important than tests, have actual CD and DVD releases that originated at 96/24. My conclusions are that the wordlength increase is the most dramatic improvement, with the sample rate increase being a secondary factor.
I like the results, but after a number of careful tests, I feel that 44.1 kHz/24 bit can be considerably improved, never as good as 96/24, but a lot better than people think. I am displeased with the losses I have heard with the current generation of sample rate/wordlength converters, when my 96/24 recordings were reduced to 44.1K. These converters are constantly improving, and you have not yet heard these recordings at their best in the CD format.

In fact, with a little less lazy engineering, 44.1K/24 can be extremely good. (16 bit is out of the running; 20 and 24 bit at either sample rate is always superior, wider, warmer, deeper, more resolved). It's a lot easier (with the relaxed filter requirements) to make a 96/24 (or 88.2K/24) recording sound good.
In fact, my experiments and those of others firmly point out that the sonic improvements we hear at the higher sampling rates are probably not due to the increased bandwidth (i.e., hearing frequencies that we previously could not hear), but rather to the reduction in artifacts in the 20-20 kHz band because the digital and analog filters don't have to work as hard.

The greatest benefit of working at a higher sampling rate is reduction of midband distortion during gain and equalization and dynamics processing. The errors (distortions) are spread over a much larger bandwidth, and it can be shown that these errors are reduced by perhaps 6 dB.
The result is cleaner digital equalization and compression (more analog-like sound). The mud comes out of the sound. For that reason alone, even if the release medium is 44.1kHz, it's a good thing to work at 88.2 K during the production process. You can hear the difference.

LC >
The 16/44 Compact Disc format is 15 years old but many audiophiles still want it to be the standard for many years to come. Do you see any future or possible improvements for this format?

BK >
You will be hearing better CDs when the origination medium is at 88.2 or 96K. We can still squeeze more quality out of the CD format.
But the real change will be surround sound. As much as I'll probably be dragged kicking and screaming into it, I foresee a data rate reduced (data-compressed as opposed to linear PCM) multi-channel (e.g., surround sound) format such as DTS or Dolby AC-3 taking over on CDs as soon as DVD with AC-3 starts taking over. The consumers will already have the decoders, and will use them to play AC-3 encoded CDs as easily as DVDs.

The reason I'll be kicking and screaming is that the recording methods that I'm best known for (recording acoustic music direct to 2-track with no overdubbing) will have far less utility when everyone has a surround system.
My 2-track recordings have wonderful ambience and space and depth and on a good reproduction system are very hard to beat. Then, someone comes along with a little less skill at 2-track, but records with a pair of ambience mikes fed to surround speakers, and *presto*, it sounds better because the ambience is spread around in a more natural manner.

In addition, although a good surround recording can sound much more natural than a 2-channel recording, the techniques for working in surround will by necessity be multitrack and require a lot of overdubbing, except possibly for classical music and "naturally-recorded jazz". The naturalness and spontaneity of the musician's performance will give way to overdubbing techniques and so on to take advantage of the extra channels.
In pop music, it's also very hard to record a pop group live as I do, direct to 2 track, and a surround version may prove technically impossible. We'll have to get into specialized and necessarily artificial post-production techniques, I'm afraid, which I think will be the end of the Katz-Chesky style of doing pop music without overdubs.

LC >
I know very well your excellent article on jitter, it is a *reference* on this topic.
May you explain, in simple words, what jitter is and in which way it affects the digital reproduction and recording?
In which way could we minimize its effects?

BK >
Thank you! Jitter is, simply, an instability of the converter clock, which translates into harder, edgier, narrower, and less-resolved sound.
For consumers, the reproduction, or the D/A side is the key to fixing jitter. A perfect D/A converter will completely eliminate jitter, but such a beast has not yet been invented. This is why better-pressed CDs are currently important.
But with a perfect D/A converter, the differences will be inaudible. I'm speaking about differences currently heard between digital cables and between poorly mastered (high jitter) CDs and well-mastered CDs (gold CDs or CDs mastered at 1X speed under carefully-clocked conditions). A good D/A can reject all of that jitter, in theory.

The best way to deal with jitter on the consumer side will be to insist that future playback systems use the new I2WS interface instead of the flawed S/PDIF or AES/EBU interface. Some CD transports and DACs equipped with I2WS interface have begun to appear, and if the analog stages of the DAC are well designed, this system will show no differences between aluminum CDs and so-called highend Gold CDs and cable difference questions will become a thing of the past.
Jitter is completely controllable; it's just that very small magnitudes of jitter are audible and it's extremely difficult to design a system using an S/PDIF interface that is audibly immune to jitter.

Now on the recording side, improving jitter situations.... well that's another story better left to another article.

LC >
Suppose you're an absolute beginner in the field (OK, it's hard...but try :-) ) and you want to record some Music in a simple yet audiophile way.
Suppose, for example, you want to record a single instrument, a voice, a guitar etc. Which are the basic needs to start?
Are a good pair of mics, a mic preamp and a DAT recorder a good starting point? May you give some advice for an absolute beginner?

BK >
If you are going for a purist (simple mike recording), the most important piece of equipment for you is your ears and talent.
Learn to discern differences in sound and be sure to have the sound you are going for in your head, for if you don't know what you're going for, then you'll never know it even if you have it. The second most important piece of "equipment" is the performer.
Without a good performer(s), even with the best equipment will sound bad. The third piece of equipment is a good room. If you have a bad recording environment, no amount of equipment will work. Seek out these three factors and you'll be in much better shape, even if you're using a cassette recorder and a pair of cheap microphones.

The next piece of advice I can give to a beginner is that mike placement is the key. The difference between a poor and a great recording can be one inch in mike movement. You have to know where to put the microphones.
One error often made by amateur recordists working with simple (e.g., 2-mike) techniques is to treat the recording like a snapshot with a 50 mm lens. You'll often end up with a boring, uninteresting presentation. The best simple mike recordings manage to have a great detail of intimacy and inner detail as well as ambience and space.
You can't get that without a lot more effort than simply finding the "sweet spot" and sticking with it. You can only get that magic spark by thinking creatively and using placement to "paint" a foreground, middleground and background in your works, like the best of the Renaissance masters.

Listen to my recording of Johnny Frigo (Chesky JD 119). There is a distinct foreground, middleground, and background, and a palpable space that surrounds that.
Most people can't believe that jazz recording was made with a single stereo microphone. Remember: Intimacy and clarity and warmth and space and depth. A good recording needs all of those, and that can only be achieved by very careful mike placement and placement of the performers.

LC >
There are many audiophile recordings, mainly of jazz and classical Music.
Good pop/rock recordings (of well know artists) are quite rare. I remember the Joe Jackson's attempt, in the album Body and Soul, where he tried (and partially succedeed) to capture pop Music with just two mics and a strictly audiophile recording chain.
Is Pop Music so difficult to capture? Aren't the famous artists interested in good recordings?

BK >
As you may know, current pop recordings of entire groups are made by putting up many close microphones,and mixing them later. In 20 minutes, the recording is ready to go. Or, the recording is put together artificially in little pieces (bass, drums, vocal, percussion, effects) over a period of weeks or months. You gain great control and flexibility, but lose depth, space, and naturalness.
The problems of getting a good performance and simultaneously managing the requirements of a simple-miked recording are very taxing on a performer. It's extremely difficult to capture pop music with a simple-mike technique. It's an art that requires tremendous dedication. Often the best performances are lost due to improper mike placement.
It takes many hours to find just the right mike placement, and by that time, the performers can be worn out. The engineer has to work very fast and efficiently and also have the human resources of a Mother Teresa to keep the performers happy.

Thus, I fear that there will not be many more recordings of this sort. Most performers, even those who appreciate great sound, do not have the patience or ability to do this. Especially pop music with singers. Sara K (on Chesky) is a rare exception, and that's largely because we keep the arrangements simple.
The most complex "pop-vocal" recording I did live to 2-track was Rebekka Pigeon's first album (Chesky JD 115). I still don't believe I did that recording all at once, everyone playing. Instrumental music (jazz) is about the only "semi-pop" music that still lends itself to this specialized recording technique, and I love doing those recordings. I hope there will be more of those.
Write to Chesky and tell them you like my jazz recordings, the company seems more inclined to try other things lately.

LC >
Now suppose you want to copy a CD you own to play it into your car stereo system. CD copies done my means of a PC and a CD recorder aren't always identical to the original.
Since this is a very Frequently Asked Question would you please tell us something about this topic?

BK >
The key is the writer. Most of these software packages can produce perfect clones, unless you buy a package that allows you to change the levels or EQ. Those packages will always sound inferior. Don't get into that game if you're an audiophile. It's impossible to copy a CD, add EQ or adjust levels without making a sound inferior to the original. Only if you have access to the 24-bit unequalized master can you produce a CD that's better or equal to the original.

But if you are just interested in making exact copies (clones), here are the issues....
If you have a good "copying" package that produces perfect clones, then the writer must be of extremely good quality. The computer industry has made commodities out of cheap writers that can produce good data, but are terrible with audio.
The CD writer I use is the best in the world. It cost over $9000 (U.S.) and is no longer made. It has a very heavy duty base, strong laser, stable clock, stable motor, etc. Current writers can now be bought for $400-$500. Don't you think something was sacrificed in the process?
But as I said above, if you have a perfect D/A converter, then the quality of the writer is less important. Forget about the writers... you won't find a great one any more; it's economically impossible to get one made that pros will buy at that price. Instead, buy a D/A Converter and CD transport combination that makes the differences inaudible.

LC >
What's on your desk right now? Secret plans?

BK >
I don't know about plans, but I'd like your audience to know that I master progressive rock, heavy metal, jazz, folk music, and classical for a lot of independent labels, and am not just an "audiophile" recording and mastering engineer. Even the most processed material deserves an audiophile's touch.

As for dreams, I'm dreaming of better equipment but 96/24 done right costs 10 times what my current investment is. That's big money... and it's what makes it hardest for me right now. Especially if I have to invest in surround.
So, although we expect great improvements right away, don't hold your breath. It's very expensive for the audio industry to do this upgrade.
Anyone want to sponsor the Digital Domain equipment upgrade? (g)

Courtesy Bob Katz for TNT.

Copyright © 1998 Lucio Cadeddu

[ TNT | Who we are | HiFi Shows | HiFi Topics | Listening tests | Tweakings | Inter.Views ]