How Hi-Fi Magazines Write about Cables, Part 21: Misunderstanding Digital Data

I’ve been browsing hi-fi equipment lately, thinking about upgrading some of my stuff. As is often the case, I have encountered some egregious bullshit, both in what manufacturers say, and in the way publications review products.

Case in point, this CD transport. First, you need to understand the difference between a CD transport and a CD player. The former merely reads that data on an optical disc and sends it to the next element in a chain – either a DAC or an amplifier or receiver that has its own DAC – over digital outputs. These are generally either Toslink or coaxial connectors.

A CD player, on the other hand, contains its own DAC, so it outputs analog audio. Some CD players also have digital outputs, allowing them to be used as transports.

You would use a CD transport if you have an external DAC (or your amplifier or receiver has its own DAC) and want to benefit from its audio quality. If you do have an external DAC, there’s really no point in paying for an expensive CD player, because all it is is a standard CD drive – with some extra electronics – and a DAC.

In other words, a CD transport has no real effect on sound quality; it is just a passthrough device to send data from a disc to a DAC.

So. This CD transport, according to the manufacturer, actually:

Pulls more data from the disc than typical disc-reading systems

Because, you know, your cheap CD player can only read the audible part of the disc; this one gets the bits that you can’t hear; or something.

One of the key elements of a CD transport or player is the error correction it provides. This helps “fill in” missing bits, when the CD is not read correctly. It’s entirely possible that this CD transport has better error correction; the manufacturer says:

• Up to 5 times fewer data errors
• Reads more data ‘right first time’ than any other CD system

This said, CD errors that are corrected are miniscule. There are two types of errors, C1 and C2 errors. There are also CU errors, which are those that are present after error correction, and generally make a disc unplayable (making it skip, for example). The CD is a mature technology, and manufacturers know how to make CD players – and transports – that reduce error correction. Suggesting that there are fewer data errors doesn’t mean that it’s not detecting errors on disc, but rather errors in the device actually reading the data. A $20 CD-Rom drive can read data without errors quite well; you can see this whenever you copy data from a CD to a computer. If there are read errors, then the files are corrupted and cannot be copied. Unlike with music, data needs to maintain integrity; error correction works on CDs because music has extra data allowing it to be corrected, to adapt to discs that are damaged or smudged. So the above is most likely an exaggeration.

However, What Hi-Fi, that audio publication that is the gift that keeps on giving, goes much further. They claim that this CD transport:

never sounds clinical or forced. It’s a truly accomplished performer that you could listen to endlessly.

Because they somehow think that the CD transport and its bits somehow affect the quality of the sound that comes out of their speakers. Again, error correction does matter, but how would that equate to this description:

Florence + The Machine’s You’ve Got The Love sounds crystal clear, astonishingly nimble and with a staggering amount of detail. The CD t breezes through the upbeat rhythm, while Florence Welch’s soaring vocals sound refined and smooth with no hint of clipping or brightness.

Or this:

The Inception soundtrack shows off just how nuanced and subtly dynamic the CD t is. It’s not just about the deep bass, but also about revealing the varied textures and layers of that rumbling bass note.

Or this:

It’s a remarkably precise performer, ripping through the rapid-fire drum intro on Van Halen’s Hot For Teacher with effortless speed and accuracy. Each instrument is easily placed in the spacious soundstage, and we love how musical and fluid the transport sounds.

What Hi-Fi thinks that the bits that this device sends to a DAC are better than average:

Its level of transparency, clarity and insight is unrivalled [sic] at this price, and is an effortless and musical listen.

In other words, this CD transport has better bits.

Uh, okay.

8 thoughts on “How Hi-Fi Magazines Write about Cables, Part 21: Misunderstanding Digital Data

  1. Let’s go a step further. Most digital formats have some form of error correction. (For CDs, it’s a cyclic redundancy check, I think.) Most errors can be fully corrected. There isn’t, nor can there be, any audible difference between a correctly read sample, and a corrupt sample that has been properly corrected. If a sample can’t be corrected, an interpolation can be made between the preceding and following samples. If a “too many” samples can’t be corrected, the audio is muted.

    As for the samples’ waveforms… There was a time when poorly formed data (eg, spiky rather than flat-topped) caused some DACs to misbehave. This appears to have solved long ago.

  2. Let’s go a step further. Most digital formats have some form of error correction. (For CDs, it’s a cyclic redundancy check, I think.) Most errors can be fully corrected. There isn’t, nor can there be, any audible difference between a correctly read sample, and a corrupt sample that has been properly corrected. If a sample can’t be corrected, an interpolation can be made between the preceding and following samples. If a “too many” samples can’t be corrected, the audio is muted.

    As for the samples’ waveforms… There was a time when poorly formed data (eg, spiky rather than flat-topped) caused some DACs to misbehave. This appears to have solved long ago.

  3. I want to say how much I appreciate your columns and discussions of audiophile folly. I have stood there giving the Homer Simpson blink as I have listened to several salespersons over the years try to sell me ridiculous Monster cables or other junk to go with equipment I have purchased. You can’t shut ’em up; just keep the transaction moving as best as possible. And then there is the capper. Was it a US$4K USB cable you wrote about a few years ago? Big fun.

  4. I want to say how much I appreciate your columns and discussions of audiophile folly. I have stood there giving the Homer Simpson blink as I have listened to several salespersons over the years try to sell me ridiculous Monster cables or other junk to go with equipment I have purchased. You can’t shut ’em up; just keep the transaction moving as best as possible. And then there is the capper. Was it a US$4K USB cable you wrote about a few years ago? Big fun.

  5. You’re calling out BS marketing but that’s a feature of marketing, not just audiophile marketing.
    There are two points you’ve missed.
    When a computer reads an optical disc if necessary multiple re-reads—to help correct data from a physically dodgy disc—are okay because the time it takes is not important. But when a music player reads an optical disc multiple re-reads of a dodgy disc can’t be allowed because the music would stop. So the latter may have to fill in the missing data with a ‘best guess’. This would explain why some transports could be better than others: if they read better they’ll produce fewer guesses. The best transports work like a computer: they read fast and re-read if necessary. They store the data in solid-state memory and read it out at the speed expected by a DAC. Therefore there IS a mechanism by which transports might produce more or less accurate bits.
    The second point is that electrically conducted interference can affect the analogue output of a DAC. That interference can come from a variety of sources, one of which is a transport’s digital, electrical connection to the DAC. So a DAC that sends less interference could affect the sound quality produced by the DAC. This is a second mechanism by which transports might affect sound quality.
    By the way, some electrical, digital interconnects may screen out interference better than others, thus providing a mechanism by which they, too, might affect sound quality.

    None of this is to say that I don’t think that…
    • audiophile marketing-speak is often preposterous
    • the prices of many audiophile wires and tweaks is an insult to the buyers
    • some of the claimed benefits have no basis in known physics
    • some of the claimed benefits are nonsense because of limitations elsewhere in audio systems
    • audiophiles appear to exhibit a dismaying lack of understanding of the placebo effect

    Regards

    • Yes, but.

      While I’ve had CDs churn when ripping them on a computer, as the computer re-reads sections, I’ve never heard that when playing a CD on an optical drive connected to a computer. I think the technology is good enough so that it almost never happens; the data is read well ahead of when it is needed, and it’s buffered. Also, ripping a CD is anywhere from 4 to 24 times as fast as the normal read speed, and errors are probably more likely. Nevertheless, even a $20 optical drive can read data CDs with no problem; CDs where the integrity of each bit is important.

      And I’m not sure about the interference. If you are using a Toslink (digital optical) connector, there is no electrical energy sent with the data; it’s just light. Perhaps interference is possible with a coaxial connection, but even then, the circuitry should be filtering out everything that is not digital data.

  6. You’re calling out BS marketing but that’s a feature of marketing, not just audiophile marketing.
    There are two points you’ve missed.
    When a computer reads an optical disc if necessary multiple re-reads—to help correct data from a physically dodgy disc—are okay because the time it takes is not important. But when a music player reads an optical disc multiple re-reads of a dodgy disc can’t be allowed because the music would stop. So the latter may have to fill in the missing data with a ‘best guess’. This would explain why some transports could be better than others: if they read better they’ll produce fewer guesses. The best transports work like a computer: they read fast and re-read if necessary. They store the data in solid-state memory and read it out at the speed expected by a DAC. Therefore there IS a mechanism by which transports might produce more or less accurate bits.
    The second point is that electrically conducted interference can affect the analogue output of a DAC. That interference can come from a variety of sources, one of which is a transport’s digital, electrical connection to the DAC. So a DAC that sends less interference could affect the sound quality produced by the DAC. This is a second mechanism by which transports might affect sound quality.
    By the way, some electrical, digital interconnects may screen out interference better than others, thus providing a mechanism by which they, too, might affect sound quality.

    None of this is to say that I don’t think that…
    • audiophile marketing-speak is often preposterous
    • the prices of many audiophile wires and tweaks is an insult to the buyers
    • some of the claimed benefits have no basis in known physics
    • some of the claimed benefits are nonsense because of limitations elsewhere in audio systems
    • audiophiles appear to exhibit a dismaying lack of understanding of the placebo effect

    Regards

    • Yes, but.

      While I’ve had CDs churn when ripping them on a computer, as the computer re-reads sections, I’ve never heard that when playing a CD on an optical drive connected to a computer. I think the technology is good enough so that it almost never happens; the data is read well ahead of when it is needed, and it’s buffered. Also, ripping a CD is anywhere from 4 to 24 times as fast as the normal read speed, and errors are probably more likely. Nevertheless, even a $20 optical drive can read data CDs with no problem; CDs where the integrity of each bit is important.

      And I’m not sure about the interference. If you are using a Toslink (digital optical) connector, there is no electrical energy sent with the data; it’s just light. Perhaps interference is possible with a coaxial connection, but even then, the circuitry should be filtering out everything that is not digital data.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.