Righteous bitsPublication date: 20 November 2013
Originally published 2013, in PC & Tech Authority
(in which Atomic magazine is now a section)
Last modified 20-Nov-2013.
I've been waiting for the "Pono" story to be revealed as a hoax, but I'm starting to fear that'll never happen.
Pono, Hawaiian for "righteousness", is a new digital audio format that Neil Young, the famous musician, has poured a lot of money into.
(Young is also an inventor. He's a big name in model trains!)
So far, so good. Wouldn't it be great if wealthy rockers and actors and authors all threw some of their money at crazy technology projects? A special note to any wildly wealthy readers: Don't buy that third Aston Martin - bankroll nutty new technology instead!
Pono is 24-bit, 192,000-sample-per-second digital audio, as often used in recording studios. The only really new thing about it is that there'll be a portable Pono player, which can play back that high-data-rate audio as well as more common formats.
Pono seems to be uncompressed, too - and specifically not lossily compressed. Uncompressed 24-bit, 192kHz audio - which I'll call "24/192" from now on - takes up 65.9 megabytes per minute, and that seems to be the kind of data rate early reports on Pono promise.
It'll be a little perverse if Pono doesn't use even lossless compression. A compression algorithm like FLAC will with little CPU load give you bit-identical audio that takes up only 60 to 70% as much space as the uncompressed data did. And lossy compression, of course, gives you much, much more music per megabyte. Uncompressed CD audio is about ten megabytes per minute; even super-high-quality 320-kilobit compression is only 2.3 megs a minute.
The big deal about Pono is, of course, that 24/192 audio is meant to sound better even than CD, let alone lossily-compressed MP3s or AACs. According to Neil Young, digital-music listeners today, who are almost all listening to music data-reduced via MP3 or some other lossy codec, are as a result enduring sound worse than that from a 78-RPM shellac record.
Seriously, that's what he says. He's been on the Late Show and the Daily Show and various other Shows saying it for more than a year now. And he may be exaggerating a bit, but he is not joking.
And actually, I think that from his own point of view he may be right, in a way. But the only way for him to be right is a terrible one, that leaves me wondering if everybody else is just humouring this old guy with a large wallet.
Problem one, which is a bit of a biggie, is that 24/192 doesn't actually sound better than CD audio.
High bit rate and sample depth are used in recording studios for the same reason photo editors use 48-bit colour and very high resolution instead of 24-bit colour and image dimensions the same as the final target. The larger format doesn't actually sound or look any better than the final - actually, works in the editing progress generally look and sound a lot worse than the final product, which is kind of the idea. But extra resolution and depth gives a bigger range for processing and correction without running out of resolution and creating odd artifacts.
Once you've finished mixing and retouching, you drop the result to the normal "consumer" format for output. Because you know that even you, the professional photo editor or sound producer, can't actually perceive any benefit from keeping the product in the extra-large editing format.
The argument put up against this by 24/192 enthusiasts is that the much higher sample rate and rather smoother waveform capture esoterica like ultrasonic instrument resonances which, on playback, combine to give a noticeably better sound.
Most instruments do not output such frequencies, and almost no microphones, speakers or headphones work significantly above the normal human audio range either. So, unsurprisingly, these opinions are shot down by blinded testing. And, equally unsurprisingly, if Neil's done any blinded tests of Pono, he's keeping them a secret.
In a blinded test - where the listeners don't know what system they're listening to, and preferably the testers don't know either (that makes a test "double blind"), careful listening can just about barely, for certain kinds of music and for certain listeners and with good enough playback equipment, reveal differences between lossless CD-quality audio and a 256-or-320-kilobit MP3 version of the same thing. Most listeners can't tell the difference most of the time, but it is possible, sometimes.
CD audio and 24/192, though? Nope.
Actually, high-bit-rate lossy audio consistently tests very well against every lossless format. This fact has caused the fringe audiophile community to simply conclude that scientific investigation of their claims, from lunatic speaker cables through green CD pens and outrageous quantum flapdoodle supposedly applicable to modern digital formats, is a denial of their holy religious faith. The more scientific you are in testing their claims, the less justified they seem, so science must be utterly abjured, lest it sully your audiophilic purity.
The observation that you can consistently get golden-eared audiophiles to think some hard-to-quantify improvement has been made in the sound when all you've actually done is turn up the volume a bit is not one they wish, somewhat ironically, to hear.
And then there's problem two, which is what makes me think the rest of the Pono people may just be humouring Neil.
As I write this, Neil Young is sixty-seven years old. He has been playing loud rock music for longer than I have been alive.
He hasn't been playing loud music non-stop over that time. He did so much damage to his ears before he was even fifty that he gave up electric rock entirely for a couple of years, to try to get the tinnitus to fade a bit.
Even without the loud music, a person in their sixties has usually lost so much high-frequency hearing that they have trouble telling words like "cot", "sot" and "tot" apart. The difference between those words is all in the high frequencies, and high-frequency response naturally fades with age.
A schoolteacher with fifty years of experience may, if she's lucky, be able to detect kids talking in the back row every bit as well as she could in 1965. But she'll have a lot more trouble figuring out what they're saying, especially if she can't see their lips moving.
So when Neil Young says that what comes out of an iPod sounds no better than a 78-RPM record, he may, as I said, be right. But only from his own point of view, and that of anybody else with severely and chronically mistreated ears. 78s can manage a frequency range of only about a hundred to five thousand Hertz, and are of course also monophonic, exceedingly noisy and have appalling dynamic range.
Good 78s aren't as bad as you might think, though; almost every 78 that exists today is several decades old and in dreadful condition, so it really does sound terrible, but a brand new 78 is a lot better. This freshly-pressed modern 78 sounds fantastic, but I'm pretty sure it goes too far the other way, being pressed with in modern smooth low-noise vinyl rather than the harsh needle-wearing shellac of "proper" 78s.
On a good phonograph, an unworn 78 should sound significantly better than the phone speaker of that kid who's decided the whole bus needs to hear the new Kesha single. But you can't say a great deal more in their favour, and they certainly don't sound as good, through proper speakers or headphones, than a decently-encoded MP3, even one at only 128 kilobits per second.
(If the MP3's monophonic, like a 78, then a 128-kilobit-per-second data rate is more than enough for superb quality. 128-kilobit mono is, however, only as good as either channel of a 256-kilobit stereo file if the stereo file is encoded with each channel a separate 128-kilobit stream. Most stereo MP3s use joint stereo for higher quality than this. And just to confuse things further, most mono MP3s are encoded as stereo but with identical channels.)
If you're eight-tenths deaf already, and can't even remember what it was like to not be, the whole world already sounds like a phone conversation to you. Add a dash of redirected curmudgeonality with regard to the stuff the kids listen to today on their WalkPods and MP-Phones, and you'll be fully primed to spend money on a new idea, like Pono, whose biggest selling points miss the point entirely.
Part of the Pono idea is that content for the system will need to be carefully transferred from high-fidelity studio masters. That actually may make some of the music sound better; popular music mastering has been a casualty of the "loudness war" for some time now. But better-mastered music will sound better on cassette than badly-mastered music would at a zillion bits per second.
I hope he's not going to fade away into senility instead.
Some readers were rather upset by this column.
I respond to them in this one!