The Assumption Presumption

We all remember what happens when we assume, right?

Self-styled Hitlerite Frazier Glenn Miller, Jr. didn’t, when he shot up a couple of suburban-Kansas City parking lots last week, and got two Methodists and a Catholic. (Lest anyone think I’m making light of that outrage, I am not. I weep bitter tears for those families, their communities, and for my country. Sure, every nation has its nutballs, but what other arms its as well as we do?)

Yet assuming is just what you and I do every time we cue up a “hi-rez” audio file from the growing collections stored on our music diskdrives. Sure it sounds better: It’s got bits up the yinyang and is oversampled to a rate approaching microwaves.

But does it? Science tells us no, that 16-bit quantization yields dynamic range of 128 dB (a LOT), and that Dr. Nyquist long ago proved that 44.1 kHz sampling will reconstruct waveforms up to (almost) 22,500 Hz. (Not approximate: reconstruct.) Few who made it much beyond high school calculus will argue either point.

When Sony first introduced the world to the Direct Stream Digital coding format—DSD, the massively oversampled one-bit basis for SACD audio, and now widely used in mastering digital in many professional environments—they invited groups of audio journalists (there were more of us back then) into various Sony Music control rooms to compare Redbook CD masters to SACD ones, under (ahem!) scientific conditions. Audible differences there indeed seemed to be, but they were very subtle indeed and frankly hard to hear, facts to which all parties, including Sony, agreed. And these were the guys selling the thing.

And yet hi-rez audio (“HRA”) certainly does seem to sound better, at least in many instances, than the CD. What’s going on? I don’t know, but that’s never stopped me from analogizing.

Twenty-five years ago us audio buffs snapped up thousands of “audiophile” LPs at prices several times those of the original record-store vinyl, because “they sounded better.” And they really did, and do. Anybody who’s ever A-B’d a Mobile Fidelity copy of Dark Side of the Moon to an original will concur: more bass, more highs, greater dynamics, quieter backgrounds, etc. And you don't have to be Aura d’Oro to hear it.

Those records were re-mastered: the original two-channel mix tapes were retrieved, and re-engineered without concerns for the many commercial strictures imposed upon the LP format. These included playability (put too much bass in a groove, particularly an inner groove, and a mass-market stylus/tone-arm will distort mightily or even physically skip), playing time (more signal, especially bass, meant shorter records, though less signal meant less dynamic range and higher noise), and they were pressed on higher-grade (quieter) vinyl.

The mastering process in our digital age is much less intrusive, especially today, but there is still potential for improvement in many cases, with different (less) equalization and dynamic-range control, especially in the pop world. In the jazz and classical spheres, many times the CD mastering processing is pretty minimal, and at least some HRA files are simply up-sampled from the two-track master, while DSD files, at least some of them, may simply be the master. I’m a big fan of DSD hi-rez, though I make no claims of personal immunity to confirmation bias or suggestibility. Just the same, we probably can all agree that less processing is better than more.

But here’s my point (always supposing I have one: in the early days of any new audio format, the programming self-selects for sound quality. E. Power Biggs and his massively overblown pipe organ launched the LP format; Telarc’s 1812 canons the compact disc. Now, HRA download sites overwhelmingly still feature good-sounding recordings created from the outset by people who cared more about audio quality than about making their track the loudest on the FM dial. Not too many people are going to line up for hi-rez purchase of, say, the first Kinks album (though I believe you can!), which might as well have been cut to a wire-recorder. In short, it’s the recording, not the format, and when we buy an HRA file we’re voting as much for the original's sound-quality-first recording practices as for bits’n’kilohertz.

I can’t go back and revert an HRA file to a master tape for comparison. I can, however, report an analogous experiment, one I’ve performed many times. Ripping my best-sounding LPs to CD-standard digital audio yields files that, under direct A-B, I believe I can just bleedin’ barely distinguish from simultaneous vinyl playback. And that’s using my audiophile-average turntable, cartridge, and phono preamp, and the encoding engine in my four-year-old civilian iMac. Plug in five-figure vinyl gear and workstation-class encoding, and I very well might get results that weren’t distinguishable at all.

Or so I assume.

veggieboy2001's picture

I've been struggling & researching this issue for a little while. It's very relevant considering Pono et all. More reviewers whose opinion I value seem to be saying the same thing. The higher bit-rate may make a difference, but the better produced/mastered recording definitely makes a difference!

again, thank you!

brewinmike's picture

my uncle, an engineer at one of our radio stations, described it to me this way: its not that you need a higher bitrate (or an analog source) to reproduce a higher frequency, but instead to keep the timing of each soundwave from left and right as close to original as possible.

if two waves (each representing one sound) leave each speaker exactly the same time and you are seated in the "sweet spot", your brain will tell you that the sound is dead-ahead, right in the middle of your two speakers. if the waves from the left channel lag those of the right, you're going to perceive the instrument is being played to your right... so on and so forth.

where a higher sampling rate comes into play is it's ability to have each wave lag or lead it's other-channel mate with a greater degree of accuracy. for instance, at 44.1, a wave can come from both speakers at exactly the same time, or if the sound should appear to be on your right, the wave from the left channel would be produced in multiples of 1/44100th of a second later. say, if the wave was lagging by 1/44100th of a second and you had amazing audiophile ears, you could point out that the snare drum is exactly 1 foot to the right of center between your speakers. at 44.1, that is the best imagining you could have.

however, if you had an 88.2 file, you would now be able to hear sound coming from half a foot from center, and at 1 foot, 18 inches and so on. 176khz would give even better imaging accuracy, and well created analog would give infinite accuracy.

I hope I didn't botch this explanation too bad, my uncle is the engineer, not I!

oh, and ΜΟΛΩΝ ΛΑΒΕ!

Rich67's picture

You're saying that the arrival difference in the leading edge of the reconstructed waveform will allow you to accurately place an instrument 9" closer to where it actually was when it was recorded because of the difference between a 44khz and 88khz signal. Also, that even if it was remotely true that that would make a difference that you could actually hear. Please!

brewinmike's picture

read this:

thanks for your interest.

dommyluc's picture

A bad recording badly mastered will only sound worse in hi-res, and a good one will only sound better, whether on MSFL vinyl, CD, DSD, or HR download. Junk wrapped in gold foil is still junk after the unwrapping. One slightly OT point: I am still amazed at the difference that Dolby Pro-Logic IIx Music mode can make on even a regular (but still well recorded and mastered) CD.

rhirschey's picture

For reasons explained in the article, as well as deeply explored in the professional and academic audio worlds....more bits do not necessarily equate to any discernable difference in sound quality. While I certainly won't turn down a 96kHz file versus a 44.1kHz, all this debate about more bits and what kinds of bits could be better put to use in SURROUND mixes. The sad part is, the genre of music that would benefit from this the most (ahem, classical) seems to have recording labels that want to argue about audible differences between 386kHz PCM versus 128 DSD (c'mon, how self serving when you're trying to sell both new and "remastered" recordings). Their efforts would be better served embracing surround encoding at, yes, hi-res...trying to open themselves up to a newer/larger audience. Have you ever watched a well-recorded classical concert on BluRay, where you get both hi-res picture and hi-res surround sound? That is the closest you'll ever get to the real thing when played on any decent system...and isn't that kind of the point? (no one I know closes their eyes when they go to a classical both hear and see the performance...rant over)

Dcbingaman's picture

I agree 100% with this post. My best recordings of classical music, by far, are Blu-Ray concerts with video content. Check out the 100th anniversary concert of the wonderful San Francisco Orchestra with Itzak Perlman playing the Mendelsohn Violin Concerto - pure bliss. This gaga reaction over high bit stereo recordings is silly. If you use an apodizing reconstruction filter (like Meridian does) you will have a very hard time telling the difference between a well-mastered Redbook CD and a high-rate digital download. A surround recording, on the other hand, makes a HUGE difference, and sounds to my ears much like what I hear when I go to Powell Hall to hear my beloved St. Louis Symphony Orchestra.

livengood1's picture

Mr. Kumin states "Just the same, we probably can all agree that less processing is better than more." Agreed. But how many people realize that to do any editing or mastering at all in the recording studio a DSD file must be converted to PCM and then back again. How can that make it better than straight PCM? Also, DSD requires heavy noise shaping that is not required in 24/96 PCM (or upsampled 44.1/16 PCM). DSD is a clever way to have us all buy our music for a fourth time. The recording and mastering are what is most important, far more important than the package in which the recording is sold.

dommyluc's picture

Why is it when there are advances in video tech, the better tech takes over, but in audio the lesser stays the norm? DVD obliterated VHS (although S-VHS should have been the tape standard), and now Blu-ray is replacing DVD, as it should, with the explosion and cheap prices of high quality big screen HDTVs. But nearly all A/V receivers now can play and stream hi-res music, entry level and flagship models.