(Video version here…)
We’re cutting it fine but right at the end of July it’s still worth pausing for a moment to think some ‘Happy Birthday’ thoughts. But first…..
What’s your take on Suzanne Vega? Like her? Loathe her? Never heard of her? In case you fit the last group she’s an American singer in a folk style with a long and distinguished career. (Confession: I like her music). One of her best-known pieces was written in 1982 called ‘Tom’s diner’ which was resurrected by UK producers DNA and became a word-wide dance remix hit in 1990. It’s the kind of song that keeps getting rediscovered and successfully adapted, most recently by Britney Spears and Giorgio Moroder. So it’s sold a lot.
It has another claim to fame, arguably of an even longer-more lasting kind, one which ensures that it has a place in every music lover’s collection.
But at least one person might have good reason to dislike it. His name is Karlheinz Brandenburg and his job involved him listening to the track over and over and over again. Try it – pick your favourite piece of music and then play it repeatedly 2,3,4 500 times over and over – my bet is you’ll not be such a fan at the end, or at least need a little time before you listen again?
The work in question was research into audio compression – how to make digital copies of music in files small enough to broadcast or share. Brandenburg needed a piece of music to test his latest algorithm. As he explains:
“I was ready to fine-tune my compression algorithm…somewhere down the corridor, a radio was playing ‘Tom’s Diner.’ I was electrified. I knew it would be nearly impossible to compress this warm a cappella voice.”
Which is how it comes about that Suzanne might have quite a good claim to being a Mother of Invention. Because the invention in question here was the mp3.
What’s an mp3?
Without getting too technical it’s what enables the music and other sound files to fit on your phone, tablet, computer or other device. We’ve had the idea of sound reproduction for nearly two hundred years, although Edouard Leon Scott (inventor in 1857 of the phonograph) didn’t think we’d want to listen and would, instead, be content with looking at traces scrawled on paper emerging from his machine as it captured the incoming sound. So while he invented recording it wasn’t until Thomas Edison in 1877 added the possibility of playback that things really began to happen. The rest is history.
Or rather a series of challenges to improve on the basic innovation, not least in terms of how we store and manipulate those recordings. Wax cylinders gave way to shellac discs which improved in capacity and reliability as newer materials became available. Other options emerged such as magnetic oxide coated on tapes, optically readable compact discs using lasers and, eventually to digital media. The trouble is that fidelity (which in Edison’s day was pretty low) can only get higher at the expense of having big files carrying the recorded information. To put it in perspective, by the 1980s when Brandenburg was working it took 10 hours for a mainframe computer to decode one minute of digitized music.
This was a big barrier – many people could see the possibilities in digital media in fields like music and films, but it wasn’t going to happen if the storage problem couldn’t be solved.
Standards can help innovation…
It’s in situations like this that standards become important – not as regulations which constrain innovation but rather as offering a framework within which conversations, research and other stuff happens that shapes the future development of an innovation trajectory. In this case the International Standards Organization (ISO) played an important role by hosting a community of practice which became known as the Moving Pictures Expert Group (MPEG)
It was particularly the brainchild of Leonardo Ciaraglione and it drew in big companies like Philips and AT&T as well as research institutes like the one in which Karlheinz Brandenburg was working. Their initial focus was to help focus the development of the technologies around a CD-ROM, but the group could already see many other applications if digital files for video and audio could be compressed. As their discussions progressed so a series of sub-groups emerged around different layers of the emerging standard model concerned with audio – MPEG -1, Layers 1,2 and 3.
Bridging the worlds of university and industry…
One of the success stories of German innovation is the way in which university research finds its way into practice so effectively, and one of the key channels is the network of Fraunhofer institutes. Set up close to major universities and with specific targeted mission areas these are staffed by a mixture of industry and university researchers, with many professors holding dual roles. So it wasn’t a surprise that Dieter Seizer was both a professor in the respected Friedrich Alexander Universität in Erlangen and also director of the nearby Fraunhofer Institute for Integrated Circuits (IIS). Nor that in 1987 he had Brandenburg working in his team as part of an EU programme – EUREKA – which brought industry and university researchers together. Project 147 was focused on digital audio broadcasting (DAB) and particularly on the audio compression problem
Karlheinz was trying a novel approach based on psycho-acoustics – a field which looks at how humans hear or sometimes don’t hear things. He realised that by filtering out many of the sounds people would still accept the quality of the remaining audio. So he could reduce the size of files needed dramatically – if he could work out precisely which sounds to leave out and encode the audio.
Which is why he needed ‘Tom’s diner’ and specifically Suzanne Vega’s voice.
A birth is announced…
By 1989 KarlHeinz had his PhD and a German patent on his approach; in 1993 he moved across to work full-time at the IIS where an enlarged team under the new leadership of Heinz Gerhauser was working to try and improve both the technology and its application
On 14th July 1995 one of the team sent an email informing his colleagues about how to refer to their work in future correspondence. In terms of christening cards it wasn’t much, but in terms of giving a name to their offpsring it was the day the mp3 was born:
Subject: File extension for Layer 3: .mp3
Hello,
In light of the overwhelming consensus of the survey participants, the file extension for ISO MPEG Audio Layer 3 is .mp3. That means we should take care to no longer use any .bit suffixes for upcoming www pages, shareware, demos etc.
There’s a good reason, believe me 🙂
There certainly was. People were beginning to see the huge potential for broadcasting and other applications using file compression techniques. The Fraunhofer team should have been in a strong position; in 1996 US patent 5,579,430 for a ‘digital encoding process’ was granted to them. They’d already built a working encoder (I3enc) and launched it in 1994 but while they could see significant potential they couldn’t persuade others that their approach was worthwhile. Big companies like Philips were pursuing their own approaches and potential users weren’t knocked out by the technology. Despite their best marketing efforts they’d only managed to scratch the surface with a licensing deal with the firms responsible for broadcasting the National Hockey League in the USA. Not quite world domination – but at least they had a name.
They decided to try another avenue – exploit the growing interest in internet marketing and offer licenses through that channel. WinPlay3 was launched in September of that year with a modest price of around $50 targeted, as the name suggests, at PC users running Windows. It wasn’t a runaway success; in fact general reception was lukewarm, not least because of its lack of an exciting user interface. But releasing software across the internet carried risks even then, especially in an environment where the ‘warez’ (= illegal sharing of files) community were experimenting with sharing music (albeit in uncompressed formats which didn’t really lend themselves to widespread activity. An Australian hacker under the pseudonym SoloH bought a copy of WinPlay3 (using a stolen credit card), and then worked his way back, rummaging around files at the FAU to find the original source code, tweaked it a little and released a freeware version. (At least he had some remorse because he named the distro ‘thank you Fraunhofer’).
From a business point of view this was bad news; the Fraunhofer team eventually retreated to a position where they continued to licence the encoder but accepted that the player element of their software was now well and truly out there. From a diffusion angle it was the turning point; the software spread quickly and was hacked further. In particular Tomas Uzelacs, a Croatian programmer came up with his AMP mp3 player in 1997. He released it initially as freeware but later partnered with a US media entrepreneur Brian Litman to try and commercialise it.
Unfortunately the cat – or rather the software – was out of the bag and had already taken up residence with, amongst others, a US entrepreneur, Justin Frankel, who licensed the AMP engine in June 1997. By 1998 he’d set up his company, Nullsoft, with Dmitry Boldyrev and created WinAmp, another freeware player for mp3. They quickly learned with the market and adapted their product, adding skins, playlists and other features so that it quickly became a runaway success, despite their charging $10 as a shareware fee. Winamp 1 sold over 3 million copies and Winamp 2, launched a year later was one of the most popular Windows applications. Nullsoft was eventually sold to AOL in 1999 for $80m. By 2000 it had 25 million users, and a year later 60 million.
But something else was happening across this emerging platform which would shake the foundations of the music industry. The big difference between digital music and physical versions is that when you lend your album to someone else there’s still only one copy. But if you share the file there are two copies, and if your friend shares with a couple of others there are then four, then sixteen – and so on.
Peer-to-peer sharing takes off…
These fluid conditions opened up the possibility for further innovation, enabling sharing in more systematic and widespread fashion. Enter Shawn Parker and Sean Fanning who launched a site in 1999 called Napster, named after Fanning’s nickname. This pioneered the idea of peer-to-peer (P2P) networking and opened up the possibilities of widespread file sharing. Within a year of its launch there were close to 30 million users and some network providers to US colleges calculated that two thirds of their bandwidth was being used by students swapping files. (It’s probably not too fanciful to think that one those music files was ‘Tom’s diner’)
At this scale it was inevitable that the music industry would fight back. When Metallica found that their song ‘I disappear’ had been played in illegal mp3 format on radio stations weeks before its official release it triggered the response and on December 6, 1999 the Recording Industry Association of America filed a lawsuit against Napster. They won – but that didn’t stop the problem even though it eventually resulted in Napster’s bankruptcy.
By 2001 digital music using mp3 as the core format was everywhere and Microsoft had bundled the capability in its Windows Media Player available on millions of PCs. The Fraunhofer team finally began to receive a small royalty from anyone using the format based on their patent and they were able to use the money to help fund further R&D.
Happy birthday = happy end?
So here we are, twenty five years on from the dawn of the mp3 age and Karlheinz and his colleagues certainly have a lot to celebrate. But – as we’ll find out in a later podcast – the big changes reshaping the music industry were still to come. And they weren’t going to come from the controlled environment of a laboratory.