The machine itself is tiny: 2.25 x 1.42 x 0.58 inches. You could probably get six of them into a single shirt pocket. It holds, following some tweaks to system memory and to the operating system, more than five thousand songs. At three minutes per song, this works out to a week and a half of continuous playing time, if the battery lasted that long, which it doesn't.

None of this would matter if it sounded terrible. It does not. There are dozens of audio settings built into the replacement firmware, though to me it sounds better with most of them toggled off or set to their center positions, and not just because changing those settings is a rather clumsy task involving multiple menus. Still, I have trouble thinking of it as a piece of proper high-fidelity gear: it's a convenience, nothing more.

Perhaps this can be blamed on good old-fashioned audio snobbery. Forty years ago, I shelled out somewhere in the low four figures for what used to be called a stereo component system. (Actually, it was a quadraphonic component system, but we won't go there for now.) Total outlay for this little music box: $70. And if I connect the little music box to one of the system's auxiliary inputs, I'll get pretty decent sound though the compressed tracks on the box are maybe just a hair lower in quality than the same tracks whirling around the CD tree on the far side of the room, and some of those CDs, I think, don't sound as good as my original vinyl.

The key word here, of course, is "compression." Compression itself is not a Bad Thing: consider, if you will, Ray Dolby's so-called B system, a simplification of his original noise-reduction scheme, designed to bring some sort of high-fidelity credentials to the lowly cassette. Tape noise tends to be most noticeable in the upper frequency ranges; Dolby B recording dynamically boosts the highs according to a proprietary algorithm, and then Dolby B playback cuts them by the same amount, simultaneously reducing the noise. It works very well, when it works; but Dolby tracking has to be correct the amount of boost and the amount of cut must be perfectly balanced and if you forget to set the B button in playback, not only is the noise still there, but the music is unpleasantly shrill.

The advantage of MP3 compression, used in most downloadable music Apple's iTunes store uses AAC, which is similar in design if not in specific parameters is that you don't have to remember to set anything in playback: all the heavy lifting is done in the recording process. But maybe "lifting" isn't the right word. What's happening is that the algorithm is determining what can be heard and what can't be at every point in the track, and then squeezing down the bits to maximize what can be heard, at the expense of what can't. Ideally, this should be a perfect reproduction of the original, but of course it never is, because something's permanently changed in the recording process.

The question then becomes: Are the compromises inherent in compression acceptable? Were I an audio purist, I'd of course say no: you threw part of the audio signal away for the sake of storage space. (Were they stored as standard waveforms instead of MP3s, only about 600 songs would fit on my little music box.) And in a proper concert hall, the dynamic range the distance between loudest and softest is at its maximum; compression algorithms reason that you can't hear the softest, and toss out the bits that define it. But is this noticeable in most listening venues? You can barely hear it in my living room; you won't hear it wearing earbuds while jogging, or while driving, because the background noise level drowns it out.

After that, it gets complicated. A lot of stuff these days gets released with little or no dynamic range to speak of: it's all loud. A trace from a downloadable track I paid for last year suggests that they recorded it as loud as they possibly could: it's like the top and bottom of the curve were cut off just barely below the point where distortion becomes total. (This is another difference between analog and digital. You gradually turn up the gain on an analog file, and the distortion increases more or less linearly, up to the point where it doesn't. In digital, you have theoretically no distortion, and then suddenly you have nothing but, because you have too many bits for the container size.)

And this doesn't even get into the question of frequency response: how high can you hear, how low can you go? Peter seems to hear about like I do:

I accept that common MP3 files are low-resolution; that's why, when I download them or burn my own CD's onto my computer, I specify the maximum possible sampling rate to ensure the best audio fidelity. I can hear the difference between the files when I do that. However, when working in other formats promising much higher frequency response, I often can't hear much difference between them and a high-end MP3 file (particularly given my aging ears, which have lost much of the sensitivity they once had loud, repeated gunfire will do that to you, and there isn't always time or opportunity to insert hearing protection). I think mixing the sound is much more important than its frequency response getting the balance right between instruments, vocals, etc. and balancing bass, treble and other notes.

That's pretty much me, minus most of the gunfire, although headphone abuse as a kid comes close. And few of us can hear beyond the canonical 20 Hz to 20,000 Hz, with the high end disappearing first. (Then again, stereo FM cuts off at 15,000 Hz, but draws few complaints, at least as regards frequency response.)

Still, there's nothing quite like gently removing the vinyl biscuit from its container, placing it on the spindle, passing the mighty Discwasher brush over its grooves, setting the stylus into the lead-in groove, and relaxing in my Big Chair until THUNK oh hell there's a new scratch where did that come from got to get up and clean it again while the music plays.

Aw, screw it. Where's my damn MP3 player?

The Vent

  1 September 2014

 | Vent menu |

 Copyright © 2014 by Charles G. Hill