All about the compact disc
In previous editions of this column, vinyl has been all the rage. But what about the CD? Stephen Dawson looks into how the humble compact disc is making a comeback in the audiophile space.
The vinyl revival brought records back to the forefront with a resurgence of sales beginning around 2006. In the past couple of years, CDs are making a comeback of their own with significant sales increases in 2021, 2022 and 2023.
ADVERTISEMENT
At around 40 million sold in the US in 2021, it has a long way to go before challenging the one billion unit record set in 2000 but this resurgence has only just begun. So, it’s a good time to refresh our understanding of the technology.
2022 was the fortieth anniversary of the CD, with the first consumer player released in Japan on 1 October 1982. My very own CD anniversary was last year, forty years after 25 March 1983, when I spent $1,199 on a new Sony CDP-101, the first player on the market.
They have come a long, long way since then, but the core technology remains the same.
Right from the start, the CD was controversial in the audiophile community. It had obvious advantages: robustness, size, easy track access and none of the noise that was the bugaboo of vinyl: clicks, pops, scratches and the like. The headline signal-to-noise ratio was better than 90dB, rather than the 40 or 50 available on the most pristine, perfect vinyl. And very few records were pristine, even on the first play.
(By the way, while the value 90dB looks like it’s twice as good as, say, 45dB, these are logarithmic values. It’s actually around 180 times better in terms of the voltage level of the noise! In practice, noise levels at -90dB are inaudible.)
Furthermore, CD decks for use with other equipment were standardised at 2V output (for 0dB full scale). With such a hot signal – turntable outputs were measured in millivolts, usually in single digits – problems like signal interference and difficult-to-solve earthing problems were eliminated. Even if CDs were to disappear, that 2V would live on, established as the standard output for DACs and streamers.
But many audiophiles considered the sound of the compact disc harsh, hard on the ear, even “grainy” and lacking a certain presence of the analogue sound to which they were accustomed.
People being people, explanations were sought and one of the most popular was that the compact disc was under-specified. That the digital audio it carried was too coarse to fully describe the true analogue sound.
CD-standard digital audio
The CD used digital technology close to what was the highest technically feasible in those days: Pulse Code Modulation (PCM) with 16 bits of resolution, sampled 44,100 times per second. The extremely expensive SoundStream system used for many important digital recordings of the day, which predated the CD by a few years, also used 16 bits, but with 50,000 samples per second.
PCM works by instantaneously measuring the level of a waveform at regular intervals and recording the sequence of numbers. They can then be decoded back into analogue audio with a DAC – Digital to Analogue Converter.
The ways in which PCM can go wrong – assuming the signal remains intact – is if the intervals aren’t really regular, or if there are too few samples per second to fully capture the complete range of audible sound or if the resolution of the numbers isn’t enough to capture the fine detail.
The first problem – irregular sound – is called jitter and for many years now it has been an insignificant problem. As to the second, Nyquist and Shannon’s work shows that all frequencies up to half the sampling frequency and be fully and completely captured by PCM. With 44.1kHz sampling, that means at least 20kHz.
But what about the resolution? 16 bits means that the level of the signal is expressed on an integer scale from -32,768 to +32,767. Note, that there can only be whole numbers in this system.
So what about something in the signal that happens between two samples? Or what if the actual level of a sample – analogue is infinitely variable, subject to quantum limits – is 1,266.93? Surely that detail is lost?
Indeed it is. That detail is sacrificed in the name of the robustness ensured by digital audio technology. But the important question is:
Does it even matter?
In reality, 16 bits with 44.1kHz sampling is quite enough to capture sound to a resolution beyond that of the human ear. We know that for several reasons. First, while many early CDs did sound rather poor, audiophiles have over the years since been able to experience many CDs that sound simply glorious. Some are newly recorded, while some are well-done transfers of older analogue recordings.
And all these superb CDs use exactly the same technology as the very first CD to come off the production line. Some listeners have compared CDs to SACD or other high-resolution versions of the same content and found the former wanting. But SACD and high-res releases are usually remastered prior to release. You’re not hearing the improvement in digital resolution, but the greater care taken in preparing the content. Indeed, some years ago one group tested many people with several high-resolution tracks. People could compare the high-res version of the music with exactly the same version but throttled down on the fly to 16/44.1 before being converted to analogue. No one could reliably tell the difference.
Why did people have problems with CD sound? Well, a bunch of early CDs were poorly mastered for the new technology and did sound quite poor. And, yes, even fine CDs really do sound different to vinyl in a hard-to-define way. For one thing, you can expect at least 1% of harmonic distortion in the vinyl track and at least another 1% on playback, if everything is set up perfectly. That distortion is typically low order, which can add a nice warmth to the sound. That is absent in the CD.
But, also, people tend to look at the stair-step patterns of graphics claiming to represent CD sound and instinctively think that they can hear those stair steps if they’re large enough. In fact, there are no stair steps. If there were, they would manifest as ultrasonic noise, but CDs filter that stuff out. Indeed, every bit of detail that is not captured by 44.1kHz sampling would be realised as ultrasonic sound. And every bit of detail that is not captured by the 16-bit resolution would actually be below the better-than-90dB noise floor.
CD player buzzwords
Now that CDs are starting to come back, we might expect to see a greater range of CD players appearing in the coming years. Which means understanding what some of the terms mean.
First, CD Transport is a CD player without a digital-to-analogue converter. Typically it will offer TOSLINK (optical) or coaxial (electrical) digital audio outputs or both. Some go for things like AES/EBU outputs, which is the pro version of the S/PDIF signal carried by the other two. Some claim that one or another of these might sound better than the others. I don’t find that myself. A CD Transport needs an external DAC. These days, many quality stereo amplifiers incorporate digital inputs. The main advantage of CD transport is that it gives you the choice of DAC to use.
I should note that this is a controversial view amongst some audiophiles, who would argue that different CD transports can somehow make the music sound different.
(Some high-end units include an I2S output, a rawer kind of digital audio connection. These use a HDMI socket. Some DACs include i2S inputs. The pin connection standards are not well defined, and compatibility between brands is not guaranteed, although some DACs allow their inputs to be pin configured.)
These days most CD players with built-in DACs, external DACs and DACs built into amplifiers and the like use Sigma Delta processing with oversampling. Digital stuff is necessarily highly mathematical, so we won’t get into the fine details of this, except to note that these usually convert 16/44.1 sound to much higher resolutions (of course, any supposedly missing detail isn’t restored in this process), which then allows many more options in the digital to analogue conversion stage.
Indeed, many standalone DACs now allow you to select from a range of filters to remove the ultrasonic components that are always otherwise present in digital audio. Some audiophiles have strong preferences for these filters. I am sceptical about these preferences, since some tout so-called NOS DACs. These are DACs that eschew oversampling, delta-sigma processing and go back to the resistor ladders of the early days of the CD player. NOS stands for non-oversampling. But they go a step further. They also eschew the filter used in all early CD players, and which continues to be used in all sensibly designed players and DACs. These purportedly “better-sounding” CD players and DACs allow significant levels of spurious high-frequency noise to go through to your speakers, while at the same time rolling off the audible high frequencies measurably (-2dB at a little over 16kHz).
If you have a choice, the most accurate setting is usually called something like fast roll-off or sharp-roll-off (rather than slow or soft) and linear phase (or no named phase) rather than minimal phase.
Some high-end CD players or DACs have balanced XLR analogue outputs rather than the traditional RCA sockets. I doubt that they produce any advantage, but if the equipment has it, it might as well be used. Well-designed balanced connections do offer improved noise rejection on the cabling, although given the lengths involved in most home settings and the relatively high voltage of the signals, noise pickup is almost never a problem.
Panic attack: CD rot!
Over the years there have of course been… well, not quite panics, but exaggerated concerns about things like the longevity of CDs. A lot of this was natural. Some of the early marketing promised near indestructibility, when it soon became apparent that plastic surfaces can indeed be scratched. And that while some scratches (radial ones) didn’t seem to cause many problems, others (circular ones along the tracks) could make a disc unplayable.
All of which means, take care. But there was one issue that was independent of reasonable care: CD rot. This was the failure of CDs after some years, usually due to the ingress of air or moisture into the CD, making its way between the layers and damaging the data.
Well, you can still take reasonable care to avoid that. Keep CDs in their boxes and try to keep them in places of relatively low humidity. I can’t offer much more than that since I’ve never experienced a “rotted” CD. As I write that, I am listening to my CD transport playing the CD of the Dire Straits debut album that I bought that CD in 1984. It’s playing perfectly.
Bargains
Even second-hand vinyl is starting to get rather pricey. But now’s the time to check the charity shops, fetes and the like for CDs. You generally won’t find much interesting modern music there – unless you count Susan Boyle’s I Dream a Dream, of which there are always several copies.
But over the past couple of years, I’ve managed to massively improve my classical music collection. I just buy CDs with interesting music on major label discs – Decca, Deutsche Grammophon and so on. Inspect the surface for obvious damage. Out of several hundred CDs I’ve purchased this way, I’ve had fewer than half a dozen not work properly.
The cost? Typically between 50 cents and two dollars. Talk about a bargain.
-
ADVERTISEMENT
-
ADVERTISEMENT
-
ADVERTISEMENT
-
ADVERTISEMENT