A U D I O... P A P E R


by Steve Deckert
Oct. 2000


Recently an article in *Stereophile written by it's editor, John Atkinson, caught my attention. It was largely about trying to understand the importance of extended frequency response in digital recording / playback. It also makes some wonderful suggestions regarding the politics behind the emerging technology that make this "higher resolution" possible.

Without getting into a discussion about SACD & DVD-A formats that can be credited for this higher resolution, I would like to talk about the point of the article which was this: Will listening to recordings with extended frequency response (ie. an octave higher than the current digital format allows) on playback systems capable of playing it make any of us happier? Will it "sound" better?

I think not. I won't dispute the high frequency harmonics of cymbals (I am a drummer) or the distortion and cabinet resonance harmonics of your average guitar amp both generate content greater than 22kHz. In fact I won't argue that there is information at 40 kHz, there often is.

The spectrograms shown on page 64 and 65 are pretty good evidence, and something I've also played with in the past year. Having set up a my own modest studio about the time we started selling the Zen Triodes, I have had the first hand opportunity to experiment with sampling rates, and word lengths. My findings agree with Tony Faulkner, who is noted in the article to have the opinion that the sampling rate of 176.4 kHz with a 16 bit word length sounds better than 96 kHz with a 20 bit word length. Even though a calculator will suggest (and it seems to be a popular belief) that music can be perfectly captured up to a frequency no higher that ½ the sampling rate, John himself admits to hearing better sound from 24 bit 96 kHz .wav files. I have had the same experience, and can confirm that recordings made at 24/96 sound better than recordings made at 16 bit 44.1kHz - the long time standard for CDs.

As I continued to read the article, it became obvious that something is being overlooked and unless it pops up at the end somewhere, I planned to write a response. Then the closing statement: "Something's happening here but we don't know what is, do we, Mister Jones?"

Actually I found that refreshing, because unlike myself who will try to offer an explanation for everything, this editor had the reserve to throw the question up in the air and see where it lands.

Nevertheless, I feel strongly that there is another side to all of this, and all of this is nothing more than a natural evolution to the digital vs. analogue debate that has gone on since digital was first used.

I don't think anyone with ears and who's heard a well done master tape played back from the machine would disagree that in the categories of musicality, dynamics, timber, and sheer balls it kicks Digital's ass in a big way. Much of that can be heard on a good LP that was direct mastered, and even the LP seems to just destroy current CD format from a fidelity standpoint.

The question is, and as the article dwells on, will adding an octave of frequency response to our current format improve fidelity. I'd hate for people to mistake the article's content and interpret that higher frequency response is the only real bonus of going to a higher format like 24/96. To me it is perfectly clear that the biggest problem with digital is it's relatively low resolution when compared to analog.

I went through a phase a few years back where I became almost addicted to watching music on a 2,048 band real time analyzer while I listened. Setting the lower end of the display window to 20 Hz and the higher end to between 20 kHz and 40 kHz, I'd spend hours at a time watching harmonics unfold before my eyes in real time. It was simply fascinating how with more and more practice you could tell which notes or sounds in a musical passage were responsible for the peaks you would see. It's a lot to watch in real time, but using a technique called "Dream Time" known by most good drummers, it is possible to either slow time down, or look ahead in time depending on which way you look at it. It's also possible to record the display and play it back in slow motion. Anyway it gave me a real feel for the beautiful way that harmonics unfold and the incredible complexity of information stored there.

The point to all of this is fairly simple; I think we all grossly under-estimate the power of music. Music is not sounds, but an organization of sounds brought on by a higher spirit. Most if not all of the mathematical TESTS you see regarding the "fidelity" of playback /recording formats have NO real preference to whether the scientist injects test tones, bursts, sounds, or music. (Actually most puke on music) Music can be observed in it's purest state by watching what happens when musicians get together and free style jamb. It can be rough at first, but then it clicks and everyone has the same dream-time, the music becomes a single complex entity with what would seem like it's own conciseness.

Consider the amount of information God placed in a single human gnome and how small it is. Music is a way that we communicate spiritually, even when we don't realize it. And anything that dips into a spiritual depth means looking past our 3 dimensional perception of the universe and understanding the 4th dimensions. In that dimension everything we know to be true is most likely false. Rotational Physics is a good example of this. I don't know of any university that will try to teach such a course because rotational physics is inter-dimensional physics, and all of the laws of nature bend or change entirely.

Surmise it to say, for as smart as we all are, as a human race we really don't know shit. In music from the whole issue of fidelity, it's the quality of what's there, not the quantity. And more specifically the quality comes from the accurate capture and transmission of the harmonic information that IS music. That is why I think even though a 20 bit word is better sounding than a 16 bit word, the 4 least significant bits that are dropped off resulting in some lost ambiance are overcome by the resolution of the higher sampling rate.

I say this because if you take a complex wave form and expand time to a point where you can follow it like a winding highway, you start to realize just how fragile the relationship is between the delicate little bends in the wave form. Analogue is parallel with nature, digital is not… at least not at this current speed.

In the article, Mr. Atkinson makes the comment about Michael Story of dCS who feels that the time smearing that occurs as a result of low-pass and anti-aliasing processes in the digital format are shortened or reduced at the higher sampling rates. I wouldn't be a bit surprised to find that to be true, but time smearing is something I'm pretty good at identifying when listening to reproduced music. After all we are all exposed to it in some degree 100% of the time by our playback gear. Time smearing, regardless of where it occurs in the system, can be identified by a subjective veil over the music, not unlike a slightly fogged lens on a camera. I am fairly certain that this is NOT the reason why CD's typically sound bad. It is possible to have a playback experience rich with body and weight and be wonderfully musical, if not live sounding with some time smearing or veil over the presentation. It's also possible to get the same results with a high noise floor, less than perfect acoustics, and no content above 12K. Yes some will be able to tell there is a slight roll of in the very top end, but nevertheless become emotionally engaged with the performance.

I for one would rather listen to an analogue performance that is lopped off at 12 kHz than a digital recording that goes up to 40 kHz because the human hearing and mind has no problem with bandwidth issues. As long as what is there is right, it works and works well. The fact that people are and have been infatuated with the subliminal effects on the brain of high frequency's (25K on up) in it's relationship to music, along with products that are appearing to accomplish this are a fine example of the respectable ignorance noted in the closing statement: "Something's happening here but we don't know what it is, do we, Mister Jones?"

Musicality is the goal, and I've heard plenty of CD's done at 44.1 kHz in the RIGHT systems that were musical and enjoyable to listen to even with the knowledge that higher resolution (analogue and digital) formats exist. The reason for this has to do with the correct balance of speed throughout the playback system. It's certainly better to not hear what's wrong at the expense of some detail than to hear everything that's wrong with ultra fast gear. Ironically the opposite is true of LP based systems where speed is your friend, since there is nothing bad to hear. I expect this to be less true with the better digital formats in some ways but can just as easily anticipate the opposite. 24/96 formats could also reveal twice as much information that you didn't want to hear. It's still digital. By that I simply mean that I wouldn't go out and buy tweeters that kill roaches just to experience the effects of 40 kHz artifacts in hopes that your subconscious will be positively stimulated in some way. At that speed (40kHz) or frequency there is nothing from digital you need to hear or absorb anyway. In nature, when God creates those high frequency harmonics they are with perfect resolution, and in that case could as far as I'm concerned do all kinds of things.

Anyway, with the format wars and politics aside, I look forward to a consumer based format for digital that takes the next step. It will be good enough on the right systems to be convincing for most of us, most of the time. But it will still remain a convenience vs. fidelity issue for many years to come.

Consider this: I have a friend who is also a drummer, but doing it 30 years longer than myself. He is almost completely deaf. If you rub your fingers together in front of your ears right now (try it) … he hears silence. (I btw., am far from it and take lengthy precautions to prevent hearing damage). He has to look right at you to hear what you're saying because he reads your lips when you talk. A very interesting thing happened the other day…. His band went to an all-digital format with regards to the studio gear and now has all the latest bells and whistles. That's been an amusing story, but I'll save the details for later. He owns a decent real-to-real tape deck that he often used to record practices with (live two-track) and it finally took a dump. He showed up at the shop literally desperate shortly afterwards because he wanted it fixed. He has been using the digital recorder (at 20 bit, 48K) in place of the real-to-real, and can't hear any detail on playback. He can't pick out the leading edges of notes making it hard to hear what he did. On the analogue recorder - which we know has limited performance on paper compared to the digital - he has no problem hearing clearly. That would indicate to me that the secret is in the harmonic integrity of mid band frequencies - say out to 10 kHz or so.

I wouldn't be at all surprised if some day in the future when we can scientifically look deeper into things we discover that harmonics in music are not two dimensional accelerations occurring around a direction change, but more holographic in nature. Information contained in the harmonics would be holographic or complete pictures of the message repeated throughout the decays. This would explain why a deaf man can perceive detail on analogue equipment and not on digital… at least to me.

Let's not worry about what's going on up there, and worry about what's going on in the audible band until we can make that "digital sound" go away! When this new higher res format comes around it will be a certain improvement, unless people decide to reproduce the 20~40 kHz octave. It's so specy.. ya know? In a time when the audio masses are hopelessly confused and cling to specifications and the "more is better" solution to everything, this would be a step in the wrong direction.


*Stereophile magazine, October 2000 issue, page 63.





Decware is a trademark of High Fidelity Engineering Co.
Copyright © 1996 1997 1998 1999 2000 2001 2002 2003 2004  2005 2006 2007 2008 by Steve Decker