CD Quality Is Not High-Res Audio

It's official. The future of audio hardware and software now has an acronym. It's HRA, or high-resolution audio, trumpets a press release from the Consumer Electronics Association. HRA may well emerge as a key theme of CEA's 2014 International Consumer Electronics Show. So this would be a good time to discuss what is, and is not, high-resolution audio.

CEA's HRA press release is long on statistics but short on specs. Even so, the stats are enough to turn heads. Most crucially: "CEA research finds four in ten (39 percent) consumers with a moderate interest in audio indicate they are willing to pay more for high quality audio electronics devices. Nine in ten consumers say sound quality is the most important component of a quality audio experience." Consumers are primed for a high-res feast. If it consisted of rotten meat and vinegary wine, what a wasted opportunity that would be.

That's what may happen if consumers and other interested parties continue to equate "CD quality" with "high-resolution audio." What if the expectant consumer hears what purports to be a high-res audio demo, hears CD-quality content, thinks "is that all there is?," and walks away bored and disillusioned? True HRA is not a subtle improvement. With the best software and hardware, a good recording, and good listening conditions, it is about as subtle as being whacked with a mallet, and I mean that in a good way. It is an eye opener. In lieu of "is that all there is?" you think "wow, listen to what I've been missing!"

What follows is my attempt to classify digital audio as high-res, mid-res, and low-res. Please note that mid-res is my own coinage, but the concept is useful and deserves to be part of the discussion.

The Compact Disc format is many good things but high-res it is not. It has a bit depth of 16 and a sampling rate of 44.1 kHz. In other words, it processes a string of 16 zeroes and ones 44,100 times per second. Digitally speaking, this is a case of arrested development dating back to the early 1980s. We can do better now. But the CD is not the worst of all possible worlds; I speak as someone who still collects CD box sets at an alarming pace. Some audio is above CD quality, some is below CD quality. That's why I refer to the CD and its file-format equivalents as mid-res.

Life could be worse. How much worse? Convert the CD into a lossy file format such as MP3, the Apple-favored AAC, or Microsoft's WMA. They omit the vast majority of the data using perceptual coding. Theoretically they lose only the data that are masked by more dominant data. But you needn't be golden-eared to tell the difference between a CD and, say, MP3 at 128 kilobits per second. There is a difference and it is painful. At 192 to 256 kbps, the sound becomes marginally less crude, and 320 kbps is closer to the CD in a relative sense. But any of these lossy formats is still sub-CD-quality. So let's call MP3 and its lossy brothers low-res formats.

How did the digital audio landscape come to be dominated by low-res MP3 and its ilk? Two words: storage limits. The first-generation iPod circa 2001 had a 5GB hard drive which, Steve Jobs told us at the time, put "1000 songs in your pocket." Had he not calculated that figure using a lossy file format, in this case AAC, the original iPod would have held little more than 100 songs. Not nearly as impressive, hmmm? MP3 is, if anything, an even lossier format than AAC. It was originally designed as the soundtrack of the Video CD, a pre-DVD videodisc format. The licensors made a goldmine when it became the favored file format of the P2P music sharing juggernaut but an entire generation of listeners has suffered from the decimation of data. This decimation must stop before it screws up another generation of listeners.

The good news is that a lot has happened in digital storage technology since the original iPod dropped. The iPod classic holds 160GB, a decent desktop or laptop computer will have at least a 320GB hard drive or better, and you can buy an external 1TB drive (a full terabyte) for less than a hundred bucks. With more space, and no mid-res bottleneck, you can store more bits in fatter files, and suddenly the table is set for true HRA.

If MP3 is low-res, and CD is mid-res, what is high-res? I define high-res audio as having at least 24 bits and anything greater than the CD's 44.1 kHz sampling rate. So there. I prefer 24/192 downloads when I can get them but I'm still happy to collect in 24/96 and I'm willing to at least try other sampling rates—as long as they offer something better than CD quality.

Looking at the nascent HRA marketplace (HDtracks is a great example) it's rare to find true HRA encoded at anything less than 24 bits. Sampling rates are all over the map depending on the retailer and album. They can be as low as the CD's 44.1 kHz (even with 24-bit depth). I question whether that is HRA. But you may see multiples of 44.1 (88.2, 176.4). The 96 and 192 kHz sampling rates are more common among high-res retailers. Most asynchronous USB DACs, those magic boxes that turn PCs and Macs into high-res songbirds, go up to 192 kHz and a few go higher. The next wave of USB DACs adds Sony's DSD format, which some believe to be the best-sounding of all.

HRA-capable file formats fall into two classes. Neither of them is a lossy format as described above. An HRA file can be uncompressed or lossless.

Uncompressed formats, such as Microsoft's WAV and Apple's AIFF, are the least efficient kind because they make no attempt to economize on bits. They are just dumb bit buckets that hold every drop of data, discarding nothing. Then there are lossless formats such as FLAC and ALE (Apple Lossless). They do omit some data, but only temporarily. When encoding an uncompressed file, they pack the data into a lossless container. Then, during decoding, they unpack the data to create a bit-for-bit replica of the original, omitting nothing. Theoretically an uncompressed file converted to FLAC should sound exactly the same (though some quibble about the possibility of limited computing power inhibiting the unpacking). Lossless files take up more storage space than lossy files, but not as much as uncompressed files, making lossless the happy medium. You can fit more files onto a device with no loss of sound quality.

Does any uncompressed or lossless file qualify as HRA? Not necessarily. If, say, a FLAC is ripped from a CD, it can never achieve anything higher than CD quality, or as I define it, mid-res quality. Mid-res in, mid-res out. But if a true high-resolution digital recording (say, 24/96) is delivered in its original form, that's high-res. Analog recording technology at its best is also high-res, so if analog is mastered at 24/96, that is also HRA. In addition, most audiophiles classify pure analog—that is, vinyl—as HRA under ideal conditions (good recording, good pressing, good equipment).

In what form can you buy HRA? It was first encoded onto discs in the DVD-Audio and Super Audio CD formats. Unfortunately they annihilated one another in yet another absurd format war. That has left HRA floating in space, like some incorporeal alien in a Star Trek episode, consisting of pure bits, able to take physical form only when downloaded onto a hard drive or flash memory. So disc players as HRA hardware are increasingly passé (though I must confess I still depend heavily on my universal disc player and turntables).

The HRA hardware of the future may be your computer feeding a USB DAC feeding a high-end audio system, a computer plugged directly into an a/v receiver with USB DAC built in, a FLAC-capable Android phone, an NAS drive, a dedicated media server, a network music player (see below), or a FLAC-capable portable music player such as the Astell & Kern AK100 (review upcoming in our magazine and website).

I type these words a few hours after seeing Sony introduce three HRA DAC-amps at a press conference in New York, including the UDA-1 ($799), HAP-S1 ($999, pictured above), and HAP-Z1ES ($1999). The first is feed by computers or other digital source components while the latter two are more self-contained thanks to their built-in hard drives (500GB, 1TB) and large front-panel color displays. All have Gracenote metadata support (to keep track info and cover art tidy), support high-res file formats up to 24/192, and add Sony's world-beating DSD formats (DSF, DSDIFF), which have become an underground sensation in the computer audio community. The trickle of DSD downloads is about to become a flood now that Chad Kassem's Acoustic Sounds has signed a pact with Sony Music, following a similar one with Universal Music Group, to offer their back catalogues.

So welcome to the brave new world of high-resolution audio. There will be spot quizzes in my future reviews, so if I drop HRA into the conversation, I expect you to know what that means. I've already begun reviewing HRA hardware in the form of USB DACs, receivers with USB DACs built in, network audio players, and the occasional mobile device. Watch your bits and hear your listening life get better.

COMMENTS
jagxjr15's picture

"The Compact Disc format is many good things but high-res it is not. It has a bit depth of 16 and a sampling rate of 44.1 kHz. In other words, it processes a string of 16 zeroes and ones 44,100 times per second...I refer to the CD and its file-format equivalents as mid-res"

This is the assumption of all too many audio dilettantes who do not understand the engineering and physics supporting digital audio in general and the CD Redbook specification in particular. In fact, the evidence does not support the contention that there is an audible difference between CD and the so-called 'high resolution' audio formats, which differ from CD by using longer word sizes and sampling rates than CD. For a reality check, see "Proven: Good Old Redbook CD Sounds the Same as the Hi-Rez Formats", http://www.theaudiocritic.com/plog/(posted Oct 17, 2007)

rossgs's picture

It's a question of mathematics. Once the reproduction of music in terms of frequency response and distortion exceeds what the human ear can detect, further "advances" cannot improve the sound. It's really that simple. You can't tell the difference reliably so why bother with new gear? When I read articles like this I keep thinking about all the people who don't have a lot of money who you are tricking into buying into what is effectively a placebo with a metal case.

david harper's picture

The difference between CD and hires is NOT "like being hit with a mallet".
That is laughable nonsense. The difference is, in fact, nonexistent. No one has ever been able to distinguish between the two in a properly controlled double blind test.

X