'MP3' Celebrates its Tenth Anniversary

Sachin Garg writes "The Data Compression News Blog reports that on July 14th 2005, the name "MP3" celebrates its tenth anniversary. On this day back in 1995, the researchers at Fraunhofer Institute for Integrated Circuits IIS decided to use ".mp3" as the file name extension for their new audio coding technology. Development on this technology started in 1987, in 1992 it was considered far ahead of its times, then MP3 became the generally accepted acronym for the ISO standard IS 11172-3 "MPEG Audio Layer 3" and no other coding method so far (2005) could uncrown MP3 as the popular standard for digital music on the computer and on the Internet."

Comments

Raise your hand...

[AKAImBatman]

...if you remember using WinPlay3 back in the day!

If you don't, well, maybe you were too young back then. ;-)

Re:Raise your hand...

[jwlidtnet]

And let's not forget the magical search for the "superior" version of L3Enc, the stupendous version 2.0, which had two advantages over the more common later versions:

a) Whereas 128kb/sec (the standard of the day) was a registered-version-only switch in later versions of L3enc, it was in the free and clear in earlier versions.

b) L3Enc 2.0 is one of the few encoders I've *ever* seen that supports dual-channel encoding, in which both channels of the stereo spectrum are dealt with entirely separately. As joint-stereo would occasionally sound like crap on some more dubiously-sourced MP3s, dual-channel was a must in those cases.

Re:Patent Issues?

[AKAImBatman]

I don't know about lawsuits, but this article touches upon the cease and desist letters they sent out. Such a move *could* have killed MP3s, except that Thomson's licensing is very reasonable.

For one, you don't need a license for "private, non-commercial activities (e.g., home-entertainment, receiving broadcasts and creating a personal music library), not generating revenue or other consideration of any kind or for entities with an annual gross revenue less than US$ 100 000.00."

Beyond that, their royalty rates are as little as $0.75 per copy, or a one time fee of $50-60K.

Re:Patent Issues?

[gbulmash]

Well except that every iPod does not support it...and that's a significant number of portable players...

Ummm, iTunes imports and convert it to AAC for your iPod.

- Greg

Re:Patent Issues?

[Peter Cooper]

I just tried it and.. nope.

Re:The tech-better isnt the all-in-wonder-solution

[Anonymous Coward]

Ogg Vorbis may be "technically" superior to MP3, but that's not the only factor that matters. The omnipresence of MP3 playback is far more important than any "technical" superiority. I can't play Ogg Vorbis files on my iPod without some major brain surgery (ie installing Linux). I can't use iTunes to encode Ogg Vorbis files without installing an incomplete and unsupported plugin. It should also be noted that said plugin causes iTunes to slow to a crawl.

Even if I managed to do all of this, why would I WANT to go to all this work? Because someone I don't know claims that Ogg is "technically" superior? So I can save a little space on my two hundred gig drive? All that work, for so little reward. Gee, sign me up!

The other reason(s) to use Ogg would be what, anti-corporate? Well, let's think about this. I've got a computer which was created by a corporation, software which was created by the same corporation and a portable MP3 player which was also created by the same corporation. It seems pretty fucking hypocritical for someone to say, "No software patents" while using proprietary hardware and software.

Why OGG Is "Better"

[jcole]

http://www.vorbis.com/faq.psp

* Vorbis files can compress to a smaller file size and still sound fine
* Vorbis' better compression will cut down on bandwidth costs
* For a given file size, Vorbis sounds better than MP3.
* If you decide to sell your music in MP3 format, you are responsible for paying Fraunhofer a percentage of each sale because you are using their patents.
* Vorbis is patent and license-free, so you will never need to pay anyone in order to sell, give away, or stream your own music.
* Epic Games (the makers of Unreal Tournament, et. al.) have used Vorbis in their games ever since releasing Unreal Tournament 2003 to compress game music without having per-game license fees sap profits from every game sold.
* Vorbis saves developers money by avoiding patent-license fees.
* Ogg Vorbis has been designed to completely replace all proprietary, patented audio formats. That means that you can encode all your music or audio content in Vorbis and never look back.

Need I say more?

-Joe

Re:Why OGG Is "Better"

[thunderbird46]

You don't save much disk space when you need to generate the mp3s anyway, to have something that an automotive CD/MP3 system, an iPod, and Macs running iTunes (in OS X) or XMMS (in Linux) can all handle :)

Re:can we legally play MP3 on Linux?

[innocent_white_lamb]

http://www.real.com/ apparently has MP3 license for their Linux player. I remember reading a notice from them about that when RealPlayer 10 was first released.

The essential mp3 patent

[tepples]

There are a number of patents under the mp3 licensing group http://www.mp3licensing.com/patents/index.html. Some look like they might expire soon, I'd welcome corrections if I am wrong.

The essential MP3 patent is listed on that page as "internal no. P3912605", which corresponds to US Patent 5,579,430. That one was filed in April 1990 and should expire in April 2010.

How Vorbis avoids the MP3 patents

[tepples]

Fraunhofer has patents on psychoacoustic compression. OGG does psychoacoustic compression.

The patents aren't as broad as you think, and the Ogg Vorbis developers have done a good job of inventing a codec that the patent claims do not describe:

• Fraunhofer has patents on multipass encoding of a single frame to achieve a target bitrate. Vorbis does not use multipass encoding, instead strongly preferring variable bitrate encoding at a given noise floor level. (I haven't investigated how Vorbis ABR and CBR work, but presumably it uses techniques in the prior art.)
• Fraunhofer has patents on subdividing a spectral-transform into frequency bands and applying a uniform scale factor, determined psychoacoustically, across each band. This results in a noise floor that looks like a stairstep, with wider steps in the high frequencies. Vorbis uses a set of (frequency, amplitude) points and connects the dots to form a piecewise-linear noise floor.

how about this one?

[YesIAmAScript]

I'm told MP2 is better at higher bitrates than MP3 is. That the extra layer 3 stuff doesn't help at higher bitrates and actually hurts.

True? I dunno.

As to the comments about OGG avoiding the MP3 patents, I have a couple things.

First, at a very low level, all MP3s are VBR. It uses a "bit reservoir", and how you deplete the bit reservoir can be optimized by multi-pass recording. VBR does not preclude multi-pass encoding, nor does it even help you maintain a noise floor level any differently than multi-pass with a bit reservoir does.

The 2nd point sounds interesting to me, it does seem like it avoids that aspect of the Fraunhofer patent.

But in the end, I can't go into this in detail, but Fraunhofer seems to think their patents cover OGG, and when you're trying to get an mp3 license, this is an issue. Are they correct? Are they using illegal means? I'm not sure it matters, it definitely puts the chill on commercial OGG support.