How music streaming is righting a wrong of the Compact Disc

Cyrus CDi player on wooden rack with magazines in background
(Image credit: What Hi-Fi?)

So the Compact Disc is in terminal decline, it seems. Everyone is streaming, so those charity shops once stacked high with boxes of discarded vinyl records now have stacks of old CDs instead at a buck or two a pop. It’s currently one of the cheapest ways to acquire music. 

I fear that such bargains will not last. Charity-shop CDs are thinning out already: there are only so many old CD collections to be donated, and besides, the very downward trajectory of CD has already started secondhand disc prices rising on online auction sites. The format is now possibly approaching the point where it may be unsuccessful enough to become cool again, like vinyl. Since cassettes have made a comeback I’ve given up trying to predict anything in this crazy world of audio reproduction.

I’ve always considered CD playback to be a pretty bizarre way of moving a digital music file from one place to another. Today we’d put the file on a stick or in Dropbox or even attach it to an email, and off it goes. But for many years record companies moved files from their storage systems to our hi-fi systems by the faintly ridiculous method of punching tiny holes in a sheet of silver and then trying to read the holes back accurately with a laser while spinning the disc at very high and variable speeds. Pretty complicated when you think about it, and it’s why CD players don’t last as long as loudspeakers. But it solved the problem of music files being too big to send easily any other way. Back then. 

From discs to downloads

But the internet got faster, while the size of music has remained relatively static. When we were first able to download files and just play them as files, I remember it seemed rather liberating, and a lot more sensible and potentially reliable than the whole punching-holes-in-silver thing. I fully expected the world to start downloading files instead of playing discs, but could never see quite how they’d sell them commercially – how would you package them in shops, on a stick? 

USB sticks were more expensive than CDs back then. So internet downloads seemed to be the future.

But of course few people were actually ‘buying’ them; for a good few years around the millennium, the nascent internet was so alive with shared BitTorrent and Limewire and Napster music files that it almost threatened the available internet bandwidth available for porn. Legally it was chaos; the record companies were doing their corporate nuts, while the networks claimed to be ‘peer-to-peer’, so not really distribution. Was this stealing, or sharing?

Only recently a friend of mine, a hi-fi writer who is a former member of the constabulary, explained to me that one of the necessary elements for the common law definition of larceny is "to permanently deprive the owner (of the property) of the use therein." And copying their music file doesn’t do that: they still have it. If I walked off with their CDs or albums, it would be a different matter. 

Copying and/or sharing is not really stealing, then, as such, your honour, honestly it isn’t. Music could be defined as a ‘non-rivalrous good’ – a public good consumed by people but whose supply is not affected by consumption.

So it is copyright protection which theoretically stops us from legally copying music. As my legally-informed friend instructs me, the US Constitution included protection for “Writings and Discoveries”, and that has been taken to include music. So the economic value of music, such as it is, is a legal creation of government.

All in all, I imagine record companies are now probably preferring streaming to the buying and downloading of files. Streaming has taken a lot of power out of their hands and into those of Spotify and Tidal, and we often hear how paltry are the payments to artists, but at least we can’t (easily) make endless copies and throw them around the internet any more. Well, yes, I know people still do that, but it’s fairly clearly considered to be wrong these days.

Back to 48k

Oh, and something else has changed as well. The rise of streaming, and the concomitant decline of CD has had one small but interesting effect, which you may have noticed if you have a DAC which shows you the quality of your stream.

The sampling frequency for music on CD has always been 44.1kHz, or 44,100 samples of 16-bit data points every second. That’s CD quality.

So in the early days of file streaming, your DAC would probably have shown 44.1kHz as well, because many streaming files were derived from the CD masters. The high-resolution equivalent of 88.2kHz is less common, for a very good reason: very few studios ever switched to recording at either 44.1kHz or 88.2kHz. Recording invariably happens at the 96kHz or 192kHz which is the usual sampling rate of high-resolution audio files. Hence most files streaming today come through at 48kHz or 96kHz, coming separately from studio masters, not CD masters.

Through the whole run of CD dominance, then, studios were recording at 96kHz, but the CDs were made from a file that had been converted to 44.1kHz. That’s not a good multiple; it involves resampling nearly every data point, so that in reality, what you’re hearing on a CD is an estimate of the sampled music’s values, not the actual music at all. Those silver discs – they’re all guesswork.

So why the heck did they choose 44.1kHz instead of 48kHz for CDs? This is one of those hi-fi questions that can have a different answer depending on where you look. The rough figure of 40kHz has an obvious derivation – ye olde Nyquist's theorem, which defines the top reproducible frequency as being half the sampling frequency; humans can hear up to 20kHz (if you’re young and healthy and have never worn headphones, otherwise you’re probably down to 16kHz like the rest of us). So our hearing demands around a 40kHz sampling rate. Add a bit extra to take account of a filter slope, and there we are, around 44.1 or 48kHz. But why the difference between the two? 

There are some entertaining rumours that the difference was deliberately introduced in the early days of digital music by record companies (Sony) paranoid about people making perfect copies to media such as DAT (digital audio tape, which ran at 48kHz). This is entirely believable, though I never heard of it at the time.

More likely is British audio writer John Watkinson's explanation that the 44.1kHz comes from the days when digital audio was being recorded to video tape, because hard drives didn't have enough capacity for long recordings (the ‘writing’ was also originally too slow for real-time recording), which is why we needed CDs in the first place. I still have one of these 'pseudo-video' digital tapes from 1984, the second time I ever hired a recording studio; if you play that tape on a VCR, it looks like rippling snowy vertical stripes of visually stored audio bits, three bits to a line. The rate at which these could be recorded depended on such variables as frame rate (or field rate, in an interlaced system), and the number of lines per field. 

The sums can be found fairly easily online, but basically 44.1kHz was the ideal for both 50Hz and 60Hz video systems once blanking lines were taken into account. This was locked in by the famous Red Book standard for CD mastering. 

So the majority of your CDs have probably been downsampled perilously from studio to disc. And streaming has now righted that wrong; we can indeed enjoy ‘studio-quality’ music from the streams of the internet, given the right provider, the right source and the required bandwidth. So take that, all you CD diehards.

Bit-perfect?

There is one remaining question: are the streaming services compressing the music destructively in order to squeeze it down the pipes of the interwebs? 

Or are they really sending down the full-fat goodness of the original studio file direct to our waiting hi-fis? Is it bit-perfect, or is there data compression?

This is a horrendous question to answer: I wouldn’t know where to start. But my friend and hi-fi writer mentioned above does, and he’s just done the testing to produce a fascinating, in-depth read on whether or not Tidal is truly lossless.

MORE:

CD, streaming and vinyl: this versatile hi-fi system has it all

Hi-res music streaming services compared: which should you subscribe to?

Compact Disc's star has faded, but I like it anyway

Jez Ford
Editor, Sound+Image magazine

Jez is the Editor of Sound+Image magazine, having inhabited that role since 2006, more or less a lustrum after departing his UK homeland to adopt an additional nationality under the more favourable climes and skies of Australia. Prior to his desertion he was Editor of the UK's Stuff magazine, and before that Editor of What Hi-Fi? magazine, and before that of the erstwhile Audiophile magazine and of Electronics Today International. He makes music as well as enjoying it, is alarmingly wedded to the notion that Led Zeppelin remains the highest point of rock'n'roll yet attained, though remains willing to assess modern pretenders. He lives in a modest shack on Sydney's Northern Beaches with his Canadian wife Deanna, a rescue greyhound called Jewels, and an assortment of changing wildlife under care.

  • Mrmason62
    So CD is in terminal decline.
    Currently CD out sells vinyl by 2 to 1.
    CD does not benefit from things such as record store day to boost sales.
    The industry wants to see the end of CD as vinyl makes much higher profits.
    In reality a digital vinyl record is no more than a cd pressed on to vinyl. They come from the same digital master.
    In the UK CD is still massively the largest physical format. Sales dwarf those of vinyl.
    Please stop doing the work of big business and spreading the lies that cd is not popular in the UK.
    If cd is unpopular in the UK then what does that make vinyl?
    I love analogue vinyl. But digital vinyl is often a waste of money.
    Prices for digital vinyl are rediculous.
    Reply
  • manicm
    I don't know what the hell the author of this piece is on about.
    Reply
  • sobercats
    Most of that is way over my head but I do know one thing to be factual. When I play the same song via my phone versus a CD, the sound quality of the CD blows the streaming version out of the water. Not even close. Like I said, I may be missing the point but for my listening pleasure, I'll take the CD all day
    Reply
  • Silsoe123
    CDs are still relatively cheap, they enable ownership, some recordings are only available on physical and personally have found them sounding consistently better with less faff than vinyl. It is no fluke they were so popular and remain so. I am looking forward to upgrading my cd player. Thanks
    Reply
  • b.roberts
    First of all, any word length (bit depth) and sampling frequency (sample rate) other than master native, if the master is digital, does absolutely nothing for fidelity. If it changes at all, it is distortion. When I see streams touting 24/96k (or even higher) when I have first hand knowledge of the actual master, it has been manipulated. Different is not necessarily better. DAT tops out at 16/48k.

    Many DATs were made at 16/44.1k. Some DATs for projects beyond the 63min maximum of the DAT master tape itself were done at 16/32k! It turned slower allowing more time. Rare, but it happened.


    If you record vinyl to CD, chances are you will not hear the difference. Try it.


    CD music quality suffered only because of-


    Analogue masters.


    Analogue master tape deterioration.

    Linage of converters.

    Poor machine alignment.

    Poor choices concerning levels during the transfer.

    Over eager "mastering engineers" attempting to make it "better." (sad, that)

    2nd, 3rd or 4th Gen tape being used.

    Tape mislabling.


    As per digital mastering-


    Going outside of the native bit rate/sampling frequency does NOTHING to improve fidelity from the original master. If it sounds "better," congratulate yourself. You enjoy euphonic distortion or perhaps prefer something outside of genuine.
    Reply
  • Silsoe123
    I reckon I have heard consistent better sound from cd because it does not suffer from fluff on the needle, wear and tear of the record or stylus. That said a box fresh record would I dare say trump a cd. Hiccups mastering are upstream of all physical and digital formats.
    Reply
  • szoze.ernany
    The intellectual level of What Hi-Fi has been a terminal decline since it started. It is simply incredible that a garbage magazine like this still sells and that anyone reads these "expert" articles.
    Reply
  • schlechtj
    Whoa, 96kbps being the standard for the entire run of cds since 1982? A decade later ADAT was released with only the option for 44.1 or 48k and most early adopters of pro tools were certainly recording at these lower bit rates due to lack of processing power on computers of that era.

    You mentioned the Nyquest theorem and that really means that as long as your sampling frequency is over twice the highest frequency (and as you mentioned a little give for the low pass filter slope), the resulting file will reproduce the original b waveform "perfectly" with an emphasis on PERFECT. Sessions may be recorded internally at higher but rates to accommodate for but loss in the mixing process but the final mixdown bit rate has zero effect on the 20 - 20khz spectrum the human ear can detect.

    I'm a home studio musician and record all the time to 44.1k. When producers ask me for a higher sampling rate, I just convert it and they cannot tell the difference, I never get any complaints.

    Cassettes really aren't making a comeback... In 2023 the total industry sales of pre-recorded tapes was under a half a million.

    I find that having cd backups, perfect copies of the songs, is a good way of protecting yourself against crashes and screw ups when moving from computer to computer.

    When watching movies at high definition, streaming seems to always leave banding and other compression artifacts that I wouldn't see if I used a Blu-ray. I suspiciously think that audio streaming is doing the same thing.
    Reply
  • Mrmason62
    schlechtj said:
    Whoa, 96kbps being the standard for the entire run of cds since 1982? A decade later ADAT was released with only the option for 44.1 or 48k and most early adopters of pro tools were certainly recording at these lower bit rates due to lack of processing power on computers of that era.

    You mentioned the Nyquest theorem and that really means that as long as your sampling frequency is over twice the highest frequency (and as you mentioned a little give for the low pass filter slope), the resulting file will reproduce the original b waveform "perfectly" with an emphasis on PERFECT. Sessions may be recorded internally at higher but rates to accommodate for but loss in the mixing process but the final mixdown bit rate has zero effect on the 20 - 20khz spectrum the human ear can detect.

    I'm a home studio musician and record all the time to 44.1k. When producers ask me for a higher sampling rate, I just convert it and they cannot tell the difference, I never get any complaints.

    Cassettes really aren't making a comeback... In 2023 the total industry sales of pre-recorded tapes was under a half a million.

    I find that having cd backups, perfect copies of the songs, is a good way of protecting yourself against crashes and screw ups when moving from computer to computer.

    When watching movies at high definition, streaming seems to always leave banding and other compression artifacts that I wouldn't see if I used a Blu-ray. I suspiciously think that audio streaming is doing the same thing.
    Hires audio is basically inaudible. Its way beyond the human level of perception.
    Nobody but nobody can discern the difference between 16 and 24bit sound
    But you can clearly hear the difference between an analogue source and digital. They sound very different
    I will always take analogue sound
    Reply
  • digloo
    This is all much ado about nothing. First, human hearing (what happens inside and behind the eardrums) degrades as we age. Second, the overall quality of what reaches our eardrums is affected by the entire signal chain, and is no better than the worst link in YOUR specific chain -- it could be the speakers or headphones or earpods or whatever; or the wires; or the quality of the DAC; or the internet speed; or the quality of the ADC; or, or, or ...

    For two people with the exact same equipment listening to the exact same music, then what's going to differ the most is their ears and neurology. But the same two people will have much different listening experiences if one is in a very quiet room and the other is on a high-speed train or plane or car with the windows rolled down.

    The comparable quality between an MP3 encoding vs. a WAV file is easy for most people to hear -- the MP3 sucks. The difference between a WAV file and a high-quality vinyl playback might not be as obvious, but a skilled person can still hear a difference.

    Show me someone who can do an A/B test between a 44.1 and 48 kHz rendering and I'll be amazed.

    Streaming services incur two main costs: bandwidth and storage. Their costs drop by using compression before storage, and then decompression at the user's end. Whatever they use may be fairly good, or it might not. But I bet if you get a hundred teens who have been using earpods or headphones since they were 10, I seriously doubt they could tell the difference between a WAV rendering or ANY compressed version of the same tracks.

    My hearing is nowhere as good as when I was younger, and is fairly flat up to 4 kHz and then falls off to -50 dB at 8 kHz before it flattens out again. Each ear is slightly different. I got a pair of Nuraphones and they measure the response curves in each ear and then program an internal EQ to compensate, much like hearing aids do. Listening with them is like going from B&W to FULL COLOR! It's amazing!

    My preferred music is 70's rock, which was mostly mastered to be pressed to vinyl and played over AM radio. Newer masterings make it much better, but at the end of the day, stuff recorded to analog tape is always going to sound different than stuff recorded digitally.

    All of these arguments fall apart for things recorded to analog media.

    For the roughly one-third of the population weaned on digital-only recordings, they're only hearing pristine sounds until they hit 20 or so, and then their ears start doing what happens to most of us, and then all bets are off.

    Fortunately, I suspect most people listen to songs from our teens for most of our lives, and we're more interested in the songs than their sonic quality. Listening to a remastered digital version of the Doors or old Pink Floyd albums is a very different experience than listening to the same albums played on a cassette in your car while you're barreling down the highway with the windows open. Trust me when I say there's no way to tell the difference between an FM radio version recorded to a cassette and a 96 kHz remastered digital version in that scenario.

    But most importantly, NOBODY CARES! It's the MUSIC that matters!
    Reply