MP3 Is 25 Years Old!

In the streaming era, music is accessed from a variety of online services, ephemeral in nature and never living on board the device. However, the online audio revolution really kicked off with the development of one very special format. The subject of bitter raps and groundbreaking lawsuits, this development from Germany transformed the music industry as we know it. Twenty-five years on from the date the famous “.mp3” filename was chosen, we take a look back at how it came to be, and why it took over the world.

Audio Big, Disks Small

1995 hard drive prices from an LA Trade ad in BYTE Magazine. The least expensive option rings in at $0.22 per megabyte, which means your 700 MB audio CD would cost $154 to store without compression (10x the cost of buying an album at the time).

The road to MP3 was a long one. The aim was to create a codec capable of encoding high-quality audio at low bitrates. Finding a method of compression that didn’t compromise audio quality was key. In an era where hard drives were measured in tens or hundreds of megabytes, storing uncompressed digital audio at CD quality — around 10MB per minute — wasn’t practical.

In the 1980s, researchers around the world were working on various encoding methods to solve this problem. Things began to pick up steam when, in 1988, the Moving Picture Experts Group called out for an audio encoding standard. The next year, 14 proposals were submitted. Four working groups were created, which began to work further on a variety of encoding methods.

Around the time the MP3’s name was decided upon, the Pentium was cutting-edge technology. Desktop computers at the time with clock speeds under 100MHz would struggle to play CD-quality files.

One of the main techniques to come out of the process was MUSICAM, which adopted a psychoacoustic model of human hearing to aid compression. This takes advantage of the effect of auditory masking, a perceptual limitation of human hearing where some sounds mask others from being heard at the same time. By eliminating data corresponding to these sounds that aren’t perceived anyway, it became possible to store more audio in less space without any perceived effect for the listener.

The MUSICAM technology became the basis for much of the original MPEG 1 Audio Layers I and II. A team of researchers at the Fraunhofer Institute took the psycoacoustic coding filter bank techniques, while mixing in some ideas gleaned from the competing ASPEC proposal to MPEG. The aim was to create the Layer III codec that could deliver the same quality at 128 kbps as Layer II could at 192 kbps. The final results were published in the MPEG 1 standard in 1993.

With the development of the Internet happening at a rapid pace, the Fraunhofer team realised their standard had the possibility of becoming a defacto standard for audio on the platform. With its small file size and high quality, it was perfect for sharing over the slow connections of the time period. In a fateful email on July 14, 1995, the team decided that their files should bear the now-famous .MP3 extension.

No Business Model Survives First Contact With The Enemy

MusicMatch Jukebox was a popular CD ripper and MP3 player. MusicMatch are notable for having actually paid Fraunhofer for their MP3 license.

The original business plan was to monetise the technology through sales of encoders. These would be sold at a high price to companies that wished to create software or hardware capable of encoding MP3 files. To drive acceptance of the standard, the decoders used to play the MP3 files would be cheap or free, encouraging consumer uptake.

Winamp was one of the most popular audio players of the MP3 era. Teenagers of the time like yours truly loved it, because it looked like a cool vintage stereo.

While this initially seemed feasible, things quickly fell apart, thanks to the very Internet that Fraunhofer had pinned their fortunes on. In 1997, an Australian student purchased MP3 encoding software with a stolen credit card, before quickly sharing it on an FTP server online. Suddenly it was readily possible for anyone to create their own MP3 files. With the files out in the wild, calls to stop the spread of the software fell on deaf ears.

Within a short time, it was readily possible to download free programs to rip audio from CDs and store it in nearly the same quality at a tenth of the size as an MP3. Websites quickly sprung up, allowing users to freely download the music of their choice. While FTP servers were the defacto file sharing standard of the day, 1999 then saw the launch of Napster, a platform that allowed users with minimal technical knowledge to directly share their digital music collections with others. The music industry had just been changed forever.

Cats Don’t Go Back In Bags

Napster was the progenitor of the file streaming movement. While it lived a short life, it inspired many services to come.

Suddenly the idea of paying $16.98 for a CD seemed ludicrous, when it was readily possible to get the same music for free online. Record labels and artists scrambled to file lawsuits and sue music fans huge sums to discourage downloading. Despite some high profile legal fights, attitudes towards music had already been irrevocably altered. MP3 players had also hit the market, allowing users to carry huge numbers of songs around without having to juggle fragile CDs. These were similarly met with legal challenges, but the juggernaut that was MP3 could not be overcome.

Even in the wake of Napster’s bankruptcy, other services bloomed in the vacuum left by its closure. Pirates learned from the case, and decentralization became key to avoiding legal troubles. This put the onus of criminality on those sharing the files, rather than those running a peer-to-peer service which merely facilitated file transfers.

The Diamond Rio PMP3000 was one of the earliest MP3 players, attracting the ire of the RIAA on launch.

Services to sell digital audio would take many more years to flourish. Initial offerings lost out due to high prices and restrictive DRM that simply gave customers a worse experience than a clean, unencumbered MP3 available for free.

MP3s dominance only began to wane in the 2010s, when a transition to streaming technology and smartphones began to offer a better user experience. Rather than having to manage a multi-gigabyte collection of songs, and shuffle them from device to device, instead users could simply call up virtually any music they wanted at the click of a button. In the same way Facebook defeated Myspace, the ease of streaming quickly relegated MP3 players and the format itself to the past.

The Format Broke the Business of Recorded Music

While few of us still trawl file sharing networks looking for the latest albums, the MP3 was key in forever altering how people expected music to be delivered, and the price people were willing to pay for it.

The pay structure for artists and labels changed monumentally throughout this turbulent time. While post-MP3 services like iTunes once sold tracks at 99 cents a song, artists now receive fractions of a cent per stream. However, the lower importance of physical media has also, at least in theory, made it possible for artists to break out without needing a record label to shift product internationally. Genres like Soundcloud rap and Vaporwave sprung up organically from services that allowed budding musicians to share their music online. It’s easy to draw a direct link between such subcultures and the dawn of music sharing online spawned by MP3.

While Fraunhofer may not have gotten the business win they desired from the technology, the MP3 undoubtedly changed the face of music forever. Artists likely still weep at the diminishing returns from stingy streaming services versus album royalties of years past, and record labels will still grate at unlicenced copying as they have since the cassette era. However, MP3 remains a technology that democratized the access to and creation of music, and for that, it should be lauded. Happy birthday MP3, and here’s to another 25 years of quality compressed music!

127 thoughts on “MP3 Is 25 Years Old!

    1. Likewise, I still have my Rio Karma, with a CF swapped in for the 1.8″ platter drive. The Karma was impressive in that , way back in 2003, it had support for ogg and FLAC files, came with Sennheiser earbuds, and had a dock with dual RCA out, pulsing lights, and network file transfer support. I guess it’s pretty outdated now, but back then it was the Cadillac of media players. Good times. I don’t really use mine anymore, but still I can’t imagine getting rid of it.

      1. Still have a bunch of Rio Forges (refurbished from rio) that are still in bags. Got a few Chiba’s as well. I was so sad when I got a phone that was easier and stopped using them. I even made a AAA – AA adapter (AAA sozed straw to go in the player wires out the cover to a radioshack AA holder)

        Man I feel old. I still remember when my parents bought me a IBM XT computer to play with. It has a huge 10 megabyte drive! (it was old at he time LOL)

    2. I still have (at my parents house) my late 90s MP3 CD player… Forget your crummy 64 MB MP3 flash players, 650 MB MP3 CDs was where it (was) at. Only downside it gobbled batteries, but did have a DC jack too. Skip protection was like a standard CD in buffer terms, i.e. a few seconds of CDDA == a minute or so of MP3, so you really had to try to get it to skip.

      1. Maybe ten years ago I found some portable CD players that happened to handle Mp3s. Since they were under five dollars (CDs were on decline) I grabbed two, precisely because they represented a small window. A bit later MP3 players took.over, offering much more storage space, and smaller size.

        I bumped into a librarian right after I bought the second one, and mentioned it. She immediately grasped the idea of that brief window.

        MP3 players have fallen out of favor, people prefering their cellphones and even streaming. But they still xist, just not from Big Name companies

        My Sandisk Fuze is wearing out, so I got a $30 player through Amazon. And “splurged” on a $20 microSD card to expand it. All of 128gigs.

        I got my first MP3 plaher late, all of 512K. Cheap memory made MP3s really viable.

        1. Personally, I have ended up using my 3 upgrades ago android 2.3 cellphone as my default portable media device. Not too large, fairly rugged, battery still works. I was actually intending to use a WinCE handheld for that, but it’s charging or battery is a bit wacky. Do you got 15 minutes or 4 hours of runtime today? It’s a surprise!

          1. How about a Palm Tungsten E2? When it was new, one reviewer gave up on waiting for the battery to die after 12 hours of playing MP3 with the screen backlight off. Replace the battery with one of slightly higher capacity and it’ll go even longer.

        1. I’ve not tried the CD MP3 out in either of mine, it’s easier to just use the aux input. Well the newer one I can BT or USB it in also. The older one though I keep thinking I should stick my top 100 or so on a disk and leave it in there when I get a round tuit.

      2. ” so you really had to try to get it to skip”

        Sadly, no. All you had to do was running with them. After a while you had stuttering, which is why I actually prefered a 128 MB mp3 player …

  1. It didn’t help the recording industry that consumers had a major feeling of being scammed, when they were told CDs cost much less to produce than cassette tapes, so everything would be cheaper…. then they were charging 50% as much again for them as cassette albums.

      1. That is if they didn’t skip… something else that was promised to be impossible.
        And partially true, because as long as the scratch is perpendicular to the track it can be handled. But somehow scratches don’t seem to care about that and are happily applied in any direction, for instance by cleaning them in a rotary motion. In a car you needed a special CD player for stable playback… though tapes never skipped.

        Sure the quality of a CD can be perfect… I write “can” because it was in the early days of CD’s until they started to record/mix them in the wrong way, where in the early days of CD recording clipping was prevented, later on nobody seemed to really care anymore and now clipping is allowed and therefore audio is being distorted on purpose, with the only purpose of making it appear to be louder. And to be honest, who will notice… and regarding “notice” why again did we needed a better medium?

        But have no fear… a new format will soon solve all our problems, so we can buy the albums we had on 78, then on 33, then on tape then on CD, then on super CD. The documentary behind the album we first seen on VHS, then on DVD and then on blueray… ahhh the horror, when will the hurting stop.

        1. 1) tapes did waARrrbBble

          2) it’s not clipping but compression. It’s normalizing all the sounds at the same volume, so the overall power out of the speakers is like one long egg.

          1. Listen to the Metallica album “Death Magnetic” and tell me that all the distortions are just compression.

            Jan and I are not the only ones thinking of this as clipping:

            And this was a great example where it was perfectly fine to download the Guitar Hero version of the Album through the P2P network of your choice (if you already had the crappy one).
            Not from a legal standpoint but morally of course.

          2. If it’s clipping, it’s a sign of an incompetent studio crew. That’s not what the loudness war was about because clipping doesn’t make the music sound louder.

          3. Clipping doesn’t make the music sound louder – that’s just an incompetent studio crew doing something stupid. The loudness war was all about compression, which is increasing the overall power of the signal by equalizing the amplitudes between loud and quiet sounds and then cranking it all up to the maximum level.

            Clipping limits the power because the clipped portion is just a flat line and that effectively mutes all other sound for that instant.

        2. I never really had 78, I borrowed 33s ins school to copy them on tape :-) Later CDs to MiniDisc or CDR or even later to the harddisk. Many of my CDs are in fact CDRs

          1. I have yet to observe that, I’ve got a pile of 20 dead CD drives and players, but all the CDs still work. I keep wondering if it’s people blaming the slow deterioration of their players performance on the CD.

        3. I doubt there’s going to be a new physical format that goes mainstream. Maybe Opus will replace MP3 eventually, assuming “buying music” continues to be a thing. But it’s hard to imagine myself choosing physical media over digital, unless the digital media is DRMed.

          Paper books are here to stay, digital is just too convenient for everything else, most people are moving that direction.

    1. The article discussed one reason for rise of .MP3’s – disintermediation and the internet. You bring up the second part of the perfect storm – cost. When CDs first came out they were expensive to produce as they needed to be lithographed and etched using similar equipment to microchip wafers. Then someone came up with the brilliant idea of glass-mastering and stamping the aluminium foil rather than etching it. Wholesale costs went from $5-8/disc in 1982 to around $0.8/disc (in runs > 10K) by mid-late ’80’s.

      From mid 90’s until around 2010 short runs (<1K) were still be run on Rimage and Micorboard equipment using CD-R.

      Good article – thanks for the share.

      1. But they didn’t drop the prices on CDs as production costs dropped, despite the much higher complexity and longer per tape production time of Compact Cassettes.

        A Compact Cassette may have a varying number of components.

        A basic tape with the bare minimum number of parts
        2 snap together shell halves with open slots or solid clear, with integral idler roller pins
        2 spools
        2 leader clips (or pins, holding leaders to spools)
        2 leaders directly bonded to the tape
        2 idler rollers
        1 pressure pad
        1 pressure pad leaf spring
        1 tape

        At least 13 parts that have to be assembled. ISTR encountering a few *really cheap* ones that left out the idler rollers for curved bits molded into one shell half, so the number could go as low as 11 parts. Many older tapes made after the advent of liners used the sheets of clear plastic in lieu of inserted clear plastic windows. An improvement offset by opening two huge holes for dust to get in. When solid clear shells came along I assume the overall cost was less than for opaque plastic, while also shaving cost by not needing inserted windows *and* better dust protection by not having two big holes in the shell.

        A high end tape with more parts
        2 screw together opaque shell halves
        2 tape windows -2 parts for no window inserts in solid clear shell halves
        2 anti-friction liners
        5 screws
        2 spools
        2 leader clips (or pins, holding leaders to spools)
        2 leaders
        2 pieces splicing to connect leaders to tape
        2 idler rollers
        2 metal idler roller pins
        1 pressure pad
        1 pressure pad leaf spring
        1 tape

        26 components, or 24 with solid clear shell halves. Get fancy with two color spools (rim and spokes two parts pressed together or possibly one overmolded onto the other) or the mini-reels with sides and the number goes even higher.

        Then add in that the tape had to be recorded in analog, though at high speed and in bulk on several large reels in parallel for large runs a lot of tapes per hour could be recorded then auto spliced to leaders and loaded into assembled tapes.

        The stamping, metallizing, lacquer coating, then printing of audio CDs could be faster, especially with several machines for the slower parts of the process running in parallel.

        Components for an audio CD
        1 plastic disc
        1 layer of vacuum vapor deposited aluminum
        1 layer of clear lacquer
        1 to 4 ink colors for labeling

        So between 1 and 7 components for a CD VS 13 to 26 or more for a Compact Cassette.

        Something I did to some pre-recorded tapes packed in cheaper housings without anti-friction liners was to open them up and install liners taken from a higher quality blank tape one of my machines had eaten. The reduced drag sure seemed to make the music sound better. At the least that eliminated the sound of the tape scraping against the inside of the shell, and it would wind up evenly rather than all ragged looking. Undoubtedly the tape tracked better from the cleaner winding.

        1. Yeah they didn’t drop the price, and they did drop the quality of the physical CDs, to the point, I can virtually predict how badly a CD will skip or if it will be unplayable entirely from start to finish based on the year it was stamped and it has nothing to do with how badly the CD was treated….

          And they are just small enough to make them very hard to keep track of, so good riddance to that awful format. For all intents and purposes my CD collection is basically worthless except for ripping, whenever I find the disc that is definitely not going to ever be in its case, and when I can wait around all day for the drive to read some sectors and/or clean it repeatedly until it can, because they just won’t play all the way through. I have many more vinyl records up to 60 years old and except for a few bad pops, with adjustment of the cartridge weight, they all still play through from start to finish, you have to change the disc and flip them over more, but at least they play unlike half of my CDs.

          Making money by selling copies of recordings was just a short blip in the history of music; now back to how it was, make a living from performances… Which pretty much means it’s impossible to make a living as a musician right now during this pandemic unless you’re already rich and famous…

    1. Haha! Me to. I also made a “time capsule” out of an old 120MB IDE drive at that time. Probably full of Pamela Anderson nudes and terrible techno music. I really don’t remember. I’ll give it another 10 years or so, and the I’ll try and read out the data.

      1. The Prodigy seemed to bootleg themselves back in the day, ’93 or so I was at a gig and they had the sound guy with cases full of white label stuff on sale. Though I think it was somewhat water testing for new material and such.

  2. too bad the mp3 format sucks. Introduces all sorts of nasty sounding audio, especially when it comes to cymbals. Sounds like a bunch of muffled garbage, even at ‘high bit rates’. FLAC is what I rip CD’s into…. no more garbled garbage sounds.

      1. ya, and that was also 25 years ago. and less than 1 year later, processing speeds doubled or better and so did storage capacity. so realistically speaking we should have ditched mp3 (because it wasn’t worth the copyright bullshit) and made something better less than 2 years after it was born. Unfortunately we had to wait until 2001 when flac hit the mainstream. Mini disc had been around since 92, and offered a more portable full quality experience. which later adopted flac.

        1. I still can’t play FLAC files in my car audio player. But to be honest: Why should I need high quality “Free Lossless Audio Codec” files do hear music in a car? If you listen to high quality classical music in your high prized audiophil home environment this may be OK. Most of the users today listen cheap bumm-bumm-hooo-hooo-bumm-bumm music on cheap smartphones over cheap earphones. Even 128kpbs mp3 is sufficient for that purpose.

      2. “And if the rest of your HiFi system is not good it does not really matter.”
        I wholeheartedly agree.
        My setup back when MP3 gained traction was mostly dumpster dived. I was in my early teens,, just wanted a lot of bass, however distorted it was. Think standard HiFi set, with like 16 random speakers tacked onto it.

        Now, I couldn’t anymore. I like my music the way it was intended to be

    1. Obviously you have only listed to MP3s that weren’t recorded using equipment connected via solid platinum Monster cables and that has been re-capped using only the lowest ESR capacitors. Also, don’t even think about judging the MP3 format if you aren’t listening to your MP3 on a computer whose sound card has at least 2 vacuum tubes.

      1. actually, i use the same kenwood unit I have used for over 20 years. I also have it on a full band eq to adjust for my speakers (which are admittedly not great atm) it’s highly sought after due to it’s discreet amp and it’s excellent sound reproduction. Yes, some of us actually have proper equipment. I know what a guitar is supposed to sound like, I know what a properly set up tube amp sounds like, i know what cymbals are supposed to sound like. And yes, most sound cards do suck a big pile of severed donkey wangs. But that deosn’t mean that mp3 doesn’t introduce a bunch of crap into the audio stream. I definitely know that that warble garbage sound is not what a drumset sounds like IRL, and it definitely doesn’t sound like that when i rip it as a FLAC and replay it. if you can’t hear what i’m talking about, don’t be disappointed, most people can’t. I can hear computer parts squealing when most people hear nothing at all.

        ie: get as close to the original recording as possible, and if you must compress it, try to be as lossless as possible, because music is meant to be enjoyed, not suffered. I cannot stand to listen to my favorite songs at low quality, because I know exactly what is missing. I would MUCH rather hear music live in a proper venue with proper sound checks, such as an orchestra, than listen to even the best of songs compressed down to a crap mp3 or over the radio, or over some crap bluetooth connection.

        I have literally never bought an mp3 player. I had minidisc when it came out because of it’s compatibility with flac and the superior audio quality for my go-to portable music. most phones and mp3 players just give me a fucking headache due to listening fatigue. Most live concerts suck at doing sound checks, from cannibal corps to bob dylan, seen em all live, and they all suck at sound checks. lol. Surprisingly, arch enemy does a pretty good job at managing levels though.

        by the way, if you want to be deafened by a harmonica, go see bob dylan.

        either way, mp3 recordings are pretty much shit. here’s an experiment for ya, rip an mp3 at various bitrates straight from the cd, something like dragonforce or any other band that has a lot going on all at the same time, then rip it in flac. listen side by side to the busiest parts of the song, you should hear exactly what I mean.

        if you don’t believe me, go talk to Henry Rollins about sound quality and how he listens to music.

        Straight from the CD is best most people can get, however FLAC is still superior on every device I have ever used vs any mp3 recording bitrate I have ever tried. I honestly just fucking cannot stand hearing someone hit a cymbal and it sounds like noise rather than a clear note. You can have the best equipment in the world, or the worst, but there is zero chance for it to sound close to the original if your recording is sub-par.

        1. An MP3 encoding can be made arbitrarily good; it requires a good encoder at the highest quality setting. If you’re getting bad sound, you’re not doing it right.

          1. Arbitrarily good MP3 means no great benefit in file size so why bother. Flack for the win.
            A ‘good’ audio recording starts with careful, and knowledgeable, microphone choices and placements, in an environment that will allow good recording, It takes skill and knowledge to mix the microphones, it takes skill and knowledge to make recording equipment that doesn’t introduce unwanted artifacts into the recorded sound. Then you want to store the recording in a lossy format.
            Try recording an MP3 recording using MP3 – does the result sound nice?
            My first PC, a real IBM PC delivered by men in grey suits had a 10 Megabyte hard disk. That was the ‘state of the art’ that led to all the interest in reduced file sizes. Long gone.

          1. The opposite of this is also true. MP3s sound crappy at rates less than 192k. In particular there’s an “ess”iness to cymbals that once you’ve heard, you can’t unhear. That’s pretty much gone at 192k.

            We push out the Hackaday Podcast at 192k because I can hear the difference. :)

          2. Decades ago I worked at a recording studio. We did lots of stuff besides what people think goes on in a recording studio. Music was probably one of the least common things. Jingles were semi common. Radio and the audio parts of TV ad’s were very common. Other things included the audio tracks for slide and film strip shows, those of you old enough to remember to change the slide at the beep, and even the narration endless loops at exhibits. And we had a surprising amount of business cleaning up old 78’s for collectors and putting them on tape. I did a lot of that. We had about 4 audio tools in the chain, a special turntable with a special cartridge for playing 78’s and a top secret way of really cleaning them. The audio chain had a 1/3 octive, one of the first commercially available ones in fact, a 2 band parametric, a DBX119 that got used all sorts of ways, and I am spacing on the name, but a really primitive click and pop remover that had a redicon bucket brigade chip in it and would stutter over clicks and pops.

            I knew how all of the tools worked and I could hear every one, and I really did not like the effects that much, I would often do a couple of takes for the customer, one with processing I could barley hear and one that would make my head hurt it sounded so bad. Oddly enough the customers almost universally liked over processed recordings. I got so many “it is just amazing” comments it was not even funny.

            I am not sure if it was just that I knew what went on behind the curtain and that ruins the illusion of if people are more impressed by the parts that beat you over the head and just miss the subtleties.
            It is like something that is easily missed unless it is pointed out and event than it is weighed against other things.

            I also play the guitar and my SO often comments that if I did not say fuck, she would never have noticed a mistake and she listens more critically than most.

    2. This! It’s alright for heavily compressed pop tunes but sounds horrible for almost anything else. And it didnt just hurt the record industry but also strangled the market for decent home audio equiment. A Bluetooth boombox is no altenative for an ampliifer with proper speakers. But that kind of product has disappeared. Yes, there is “specialist” audiophile gear but that is ludricously expensive

      1. You don’t need audiophile grade equipment to hear the compression. A 25 euro philips/sennheiser headphone plugged into a laptop will already reveal a ton of detail, or lack thereof.
        You can also get a set of really quite good speakers for around $180 (Wharfedale Diamond) and some kind of tripath digital amplifier, and end up with a good sounding non-audiophile set.
        Let alone, when you put some money towards 2nd hand equipment…

        1. I briefly worked for a Wharfedale dealer 40 (or so) years ago.
          The customers blew speaker components frequently.
          I hope their dealer repair support has improved in the meantime.

      1. Yes. NOW there are choices.

        So many years ago the choices were CD or much worse than MP3 quality cassette copy of a friend’s CD. Then there was MP3. For a young person on a limited budget, not born with a trust fund overpriced CDs were really only an option for maybe a few discs per year. It was a different time.

        All these people whining about the sound of a cymbal crash being a little off just don’t get it. MP3 really opened up a whole new world of possibilities. Especially when the big radio conglomerates were already ruining broadcast FM. Or at least they were if you didn’t like hearing the same damn song four times in a mere hour along with twice that many commercials. For many of us MP3 was the first and only viable alternative for at least a decade. Even Napster is for the kids. I’m talking about FTP sites!

        Now music is available at reasonable prices and we don’t have to be pirates. Or maybe I just feel this way now that I am on an adult sized budget. Either way, this is much better. Artists should get paid, just never again in the monetary equivalent of human body parts please. Not that I believe it was the artists that actually received the majority of that 1990s money anyway.

    3. I’m glad my cloth ears* can’t hear a difference between MP3 and anything else. Saves me loads on hifi equipment, I can buy normal cheap cables rather than being forced to buy the hand-wound platinum orgainic oxygen-free ones for starters.

      * (probably caused from listening to lots of live music up close)

    4. People always like to trash the mp3 format, yes the quality is less than a lossless format, but i think they made acceptable trade offs to reduce size. The reason the mp3 is no longer the dominate format is not an issue of quality, but quantity. The 2 audio channel of mp3 became an issue as the cost of home surround sound audio became affordable to mere mortals like me. In 2001 the m4a format was introduced with 5.1 channels, 5 independent channels and a subwoofer channel that mux the other channels and filtered out the midrange and the treble. Apple help push this format to become the dominate lossy format. The only reason the mp3 format is even still relevant is that m4a supported drm and initially the easiest way to strip the drm was to convert your music to mp3. If mp3 supported drm it would have died 15 years ago. To many the mp3 format equals freedom, the freedom to listen to music where and when you wanted, on any device you want with out restrictions. For this reason the mp3 will always have a place in my heart.

      1. 5.1 sound isn’t compatible with most -stereo- equipment anyhow, so such formats have little to no use. People only have two ears, and most can’t be bothered to sit in just the right spot in the room to hear the effect.

        It’s annoying enough that Netflix defaults to 5.1 sounds if you have 5.1 speakers, so you can’t switch to headphones without clicking a different voice track every single time. The other issue is that the 5.1 mixing is usually done badly by cramming the narrator voice in the middle channel, which makes it stick out like a sore thumb.

  3. Photo caption “Around the time the MP3’s name was decided upon, the Pentium was cutting-edge technology. Desktop computers at the time with clock speeds under 100MHz would struggle to play CD-quality files.”

    Pentiums were pretty good at it, with a decent player, I have a P5 P60 that could play them smoothly while multitasking, 486es would struggle, DX-2-66 could manage if you didn’t do anything else. DX-4-120 would be okay while playing minesweeper or using wordpad or other lightweights. Also you needed the RAM, if you didn’t have 16MB you were going to have a bad time.

    Memories in this respect may be clouded by the bogging down of systems of the era with the early yahoo toolbar helpers and crap like Bonzai buddy, the dawn of adware, spyware and crapware that many users weren’t very savvy about, (Maybe it even came on their ISP install disks) and if they had one of those, used a resource hog like windows media player to play them and got a few “optional extras” with their MP3s from the dodgy sites, yeah, they’d have a 400Mhz machine and still find it stuttering and dropping out.

    1. I seem to remember that CD-ROM drives of that era had audio-out that you could connect to the motherboard and it would be an analogue audio channel that went directly to the headphone/line-out.

      1. They’re referring to CD quality MP3s, you only needed bare minimum CPU capable of sending drive control commands to play passthrough analog audio. Some drives, you only need power to them because the front panel controls are sufficient. Though sufficient may vary, some could be got with 6 buttons and a track display LCD, some just had the eject button and auto-played audio tracks in sequence. If they had a 3.5mm jack you didn’t need the audio hookup inside the machine to sound card or onboard sound, or header to RCA jack back panel passthroughs like some SCSI “CDROM”” cards had.

      2. I believe that right up to the end of IDE cd-rom drives, they still had four-pin (three populated) headers with matching cables that went from the drive to the motherboard or sound card. I still have a couple of desktops with those, that aren’t all that old, like they shipped with win7.

    2. Ah, the old DX2-66… had one with a SCSI CD-ROM drive and a whopping 16MB RAM.

      Does anyone have any idea to what happened to those sampled MOD files? I had hundreds of files downloaded from several BBS’, but they seemed to vanish after MP3 got us ‘real music’….

      > Ops, nevermind: just found it again. For my surprise, now we can play MOD files directly on web using JavaScript.

    3. speaking of awful crap to put on your computer. Have/do you hear of/remember something called sub-seven? I believe it was from around that time. My memory is a little fuzzy. Funny stuff.

    4. That reminds me – I had to refer to the mpg123 manual not too long ago, and got a kick out of this line:
      “MPEG audio decoding requires a good deal of CPU performance, especially layer-3. To decode it in realtime, you should have at least a Pentium, Alpha, SuperSparc or equivalent processor. “

  4. Was there a separate Fraunhofer encoder available for free use that was limited in some manner? I certainly recall using their encoder to make my own MP3s prior to 1997, but it was mighty painful and slow.

      1. Fraunhofer was time limited unless you had the unlock code. LAME was released a year after Fraunhofer. Fraunhofer did take a long time but the quality was way better than Xing, which took like 1/10 the time.

  5. To anyone interested in a longer form exploration of MP3, its technical development, its effect on the music industry, and the pirates that made it famous, I wholeheartedly recommend the book “How Music Got Free” by Stephen Witt.

  6. I find it really confusing when the principles of using psychoacoustics in the coding is referred to as “eliminating data corresponding to these sounds that aren’t perceived anyway”, when it actually works the other way around: the data is encoded in a quantized format and the psychoacoustics are used to determine if the quantization noise is audible or not. If it’s not, a coarser quantization can be used saving more bits.

  7. Does this mean that MP3s are finally out-of patent in the US? (or did that happen years ago.) It’s a while since I’ve done a fresh Linux install, I seem to remember that mp3 encoders/decoders were in the “bad” codec bundle (the good, the bad, the ugly) that couldn’t be officially distributed.

  8. This was when I was working on CD pre-mastering equipment. My computer had a monster ‘multi-media’ 1Giga byte hard drive. These drives would guarantee continuous streaming access. Standard hard drives in that segment would have to resync their servo track every few minutes (due to heat the platters would expand in size). 700 Mb of that 1Gig was to store an uncompressed CD image :)
    Certain CD-ROM drives (Plextor was the etter one) could actually read a red book CD digitally over SCSI.

  9. I had a mp3 player for nextstep once, running on a 33mhz turbo color slab, but the developer got an s&d letter from fraunhofer and the software disappeared. :( my hd crashed and i’m still looking for that player ever since.

  10. To the discussions re: audio quality. Don’t forget how a 128 kbit CBR MP3 encoded in 1999 sounds notably worse than one in 2010-2020. The techniques have improved considerably over time and with more processing power.

    I have ripped my entire CD library as FLAC files in order to be able to transcode to whatever on demand (I find MusicBee makes this pretty easy, you just say “this flash drive has to be X format and Y quality” sort of thing and bam everything Just Works for my car stereo. On modern hardware the transcoding to VBR M4A is so quick it might as well not be happening.

    1. That’s the reason lossy compression systems like MP3 and h264 etc only specify the format and how to decompress it, not how to encode it in the first place – because then encoding techniques can improve without forcing end users to constantly upgrade their players and libraries.

    2. I do the same. I have a master library that is ripped to flac and I rip mp3’s from that for the car and RV and phone etc. Oddly enough I tend to use the best quality rips for the phone because I tend to use earphones with that. My car and RV have enough noises etc that the rips can be pretty sloppy and still not be offensive.

  11. Fraunhofer issued a nasty press release saying “mp3 is dead” when all the patents expired. They did so in order to push people to adopt their new patented format, and get money flowing into their pockets again.

  12. In 1978 I read in a little news column that Phillips was working on a disc with pits read by a laser, it had no name. Upon reading the encoding specs I put a finger down for each bit and wasn’t happy till 21 to 24 bits. Just 16 bits, no way! 24 is here now and at a higher sample rate than Nyquist. The CD isn’t a hifi media, the top third of the spectrum is “pixelated”, Nyquist doesn’t work the way they teach. Arguing over the CD’s superior quality is moot.

    Speaking of that word “quality” in these contexts of course it means NOT. In an age of megabit drives they should have waited rather than deal with the mess they have made. Analog media have been transferred to this short shrift stuff and the originals have been thrown out thinking they done good, it’s even worse with film and video.

    MPEG, the folks that put a round one in a square hole and hammered it in tight. How many times to peck at Tom’s Diner till it’s “OK”? A generation’s recordings on that short 16/44 standard straight from the studio never recorded well. Add to that the vault fire and lots have been lost to the future.

  13. Love the article, except for the comment about how Facebook defeated MySpace. Facebook defeated MySpace through a ruse. Once MySpace became mainstream and everyone’s family members started jumping on, Facebook opened up outside of Harvard with the guise of privacy and secrecy by requiring a college email address to join. Once they suckered everyone, they opened it up their family members and became the juggernaut they are now.

  14. “truly loved it, because it looked like a cool vintage stereo.”

    While, obviously, they really loved it because it really whipped the llama’s ass, I had no connection in my mind to vintage stereos while using it. It was clean and usable, and while they kept making it more difficult with each version to get back to the classic theme and functionality, each install I would change it all back, even when playing with Linux desktop I imported the classic winamp theme into xmm2.

    I don’t know why so many players changed to the default of having a double click on a track in the library starting a new playlist with the entire album and playing the track you selected.

    Also milkdrop was an excellent visualizer.

    1. AND Winamp’s queue management was/us the best. You can add songs from the playlist to “play after this song” then when it’s finished it goes back to where it was. Great for parties – your friend could queue up their favourite track without ruining the rest of the playlist.

  15. Creative pioneered the mp3 player with their ZEN player, then Apple came along and stole the patent and idea from them to make the iPod. Creative sued Apple and Apple had to pay Creative

    1. 6,928,433 “automatic hierarchical categorization of music by metadata”, filed in 2001-01-05
      ‘patent covers an interface that lets users navigate through a tree of expanding options, such as selecting an artist, then a particular album by that artist, then a specific song from that album ….on
      a portable media player’

      Brilliant patent!!1 Its something computer music players have been doing up to that point, but creative filed one of those “on a mobile device” patents and keeps suing companies to this day. Google even preemptively sued Creative in 2017 just in case.

  16. I was one of those OGs who was ripping and encoding my audio CD library (from the Windows 95 command prompt) and collecting MP3s pre-Napster. I used all of the software shown above (Winamp was my favorite) and also had the Diamond Rio player! I couldn’t even fit an entire album’s worth of MP3s encoded at 128kbps on the storage! I was one of the first with a CD-R, so I ended up having to store all my MP3s on CDs since I had only 1.6GB hard disk space between the two hard drives I put in my computer. I remember reaching the milestone of over 24 hours of unique music. I’m sure I have way more than that in my Amazon Music library now. Now that I am an adult, it’s just easier to buy it, even if I have the original CD somewhere in a box.

    1. I also later worked at an audiobook publishing company where I was the sysadmin, but due to my prior working knowledge of MP3 encoding, I was also later tasked with R&D on creating some of the first MP3 audiobooks. I used a digital master of The Adventures of Tom Sawyer and some of the other original ones were also classics since the text of those were public domain and we had no royalties to pay from using it. (It was kind of like the wild west at the time, and I think the lawyers hadn’t yet worked out how author royalties were going to be paid for MP3, since it was to be an imprint with lower media cost and it wasn’t specified in prior contracts.) We were planning to distribute them via CD-R, and I was tasked with picking the optimal bitrate and format, and calculated the capacity of a 700MB disc at that bitrate. Along with some probably now defunct metadata, we could fit most books on one disc, extra long unabridged books could take two. The studio recorded the voiceover in mono, so even if intro music may have been in stereo, to save space we mastered the MP3 files in mono. 64kbps mono optimized for voice sounded as good as or better than a joint stereo music recording at 128kbps. Now this was somewhere close to 20 years ago, and encoder software has changed a lot in that time, so the results may vary if this experiment were done today. I wonder if they still sell that first book and if they’re still using the files I created. I should try to buy it and find out. I’m sure there’s metadata there in the ID3 tags so I could tell.

  17. You can take my Pontis SP600 from my cold dead hands! Streaming is nice but wait till that serverfarm is no more and all your music is gone, i can still listen to my favorites streaming in from my very own CF card.

  18. Winamp it really whips the lamas aSS :P
    Best music player ever, I still use it Today and plan to use it for the next 20 years as well. I think this is clearly shows that we do NOT have to replace good old things with new shiny actually shittier technologies such as:

    init system->systemd

    We don’t need this new garbage and won’t need it 20 years from now either why not just stick with which works well (I certainly will).

    1. Do you realize that IPv4 addresses are no more available for years already in several regions, and some of the big routers had to be replaced because they could “only” store a million entries for that table. That limit was only reached because ever smaller blocks of IPv4 addresses were given out to different locations.

      Maybe you should go back into retro computing and buy a ZX80 or commodore64.

  19. An MP3 jukebox project in 1999 was my introduction to GCC and AVR processors. An internet user somewhere in Europe put together the YAMPP project and sold PCBs connecting a Hard Drives to an AVR and a ASIC mp3 decoder.

  20. I think MP3 and other encoding formats will always beat streaming in one thing: you need connectivity in order to play music, i.e. from Spotify, while can bring your favorite music along with you, EVERYWHERE you go!

    Plus, as others said in the comments, for Streaming you need to rely on a service provided by someone who could decide to shut down the servers at any moment, and you favorite music is gone.

    Personally, I like to have my music converted from the CDs I own, to listen to it from my phone and car stereo, while the original supports are well protected in a drawer, at home: in 2020, memory cards are not so expensive after all, you can buy a 64 GB MicroSD for about $25 or less, and you can put a bunch of high-quality (MP3, 320Kbit/s) songs there!

    Exact Audio Copy + LAME Encoder forever! :)

  21. Back in 2006, there was a high-profile murder case in Belgium, where a schoolboy was stabbed to death for refusing to hand over his MP3 player to his assailants. The killer was caught on security cams, apprehended and sentenced to 20 years in prison. Less than a decade later, an MP3 player was standard functionality in everyone’s phone.

    1. The real news was that dedicated hardware MP3 players could be got for $20 5 years later, but sure let’s gloss over that and laud the achievement of a $500-$1000 phone being able to emulate one. (Talking about 2011 here, I am aware that $100 phones do it now.)

  22. “The MUSICAM technology became the basis for much of the original MPEG 1 Audio Layers I and II.”

    You left out one “tiny” step there: The Digital Compact Cassette (DCC) by Philips and Matsushita. It used Precision Allocated Subband Coding (PASC) encoding, which was the predecessor of MP1. To be precise: MP1 and PASC are identical except PASC is limited to the 384 kbps bit rate.

    The DCC documentary ( tells the story of how DCC failed to catch on because it got delayed, and how Philips made millions on the intellectual property by selling the rights to companies like Fraunhofer.

    === Jac

    1. To this day, there still does not exist a format that provides a universally supported container for both uncompressed PCM and the requisite metadata. WAV and AIFF are both problematic with regard to metadata.

    2. I think it’s a testament to how the speaker quality has been slowly creeping into our lives. 25 years ago we were using some crappy overpriced PC speakers to listen to music. Now, I can hear bad encoding even when listening in the car.
      I would even go as far as saying that an Echo Dot has higher fidelity than an old soundcard with those beige desktop Creative speakers.

  23. :D I bought the Mp3 grabber on 1997 and I made a backup of all CD’s and cassettes; I still have that bounch of recorded media. Also I made th first portable mp3 player on 1999 with and old computer and wireless keyboard. that was an odyssey

  24. I refuse to jump on the streaming bandwagon. For a simple reason. Eventually that server farm is gonna go live on a farm and bye bye. No thanks, still got all my mp3 files from the day I started my music collection. Still use WinAmp 2.x. The only habit I dropped is listening to music on the go.

    I bought my first mp3 player for around 180 eur in 2004 in Brussels. That thing was basically a 128MB usb stick with an lcd display. Also had a couple other features, like a voice recorder and an FM radio. Neat thing. Too bad it died within 2 years of *heavy use*

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

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