Thema: Zukunft von Ogg Vorbis
Emmett ist ex-CEO (entspricht dem deutschen Vorstandsvorsitzenden) der Xiph.Org Foundation, die den Audiocodec Ogg Vorbis entwickelt hat und zu der auch FLAC und Speex gehören. Im heutigen Beitrag seines Blogs Emmett Writes Things berichtet er über seine Befürchtungen zur Zukunft von Xiph.org und damit auch Ogg Vorbis.
I recently came upon the new listening test put together by Roberto Amorim on Hydrogen Audio. Roberto's a good guy, and while I've had knock-down drag-out arguments with people on Hydrogen Audio in the past, this post isn't about them at all. My deepest respect to everyone there.
So, Vorbis won the 128kbps listening test, and that's quite an achievement, especially with AAC running so popular these days. The problem is that a fear I had every single day when I was running Xiph manifested itself when I read the results.
Vorbis has been taken over by someone else, who is doing new development and tunings. This is supposed to be great. It's one of those great selling points about Open Source. Personally, this leaves me feeling completely horrible, and I'm probably the only person alive who feels this way.
I did my damnedest to make sure that all of the projects we had during my reign of terror were updated and moving in the right direction. I knew that if we chilled out for even a second, something like this would happen. Someone would take our codebase and make it better. And they have, making Xiph entirely irrelevant. It's not the company people care about, it's the code. Why should people donate a dime to Xiph to support new developments if the 'current' stuff is being quickly co-opted by faster projects?
I put an idea on the table fairly early on at Xiph, because I was terrified of these exact events happening. I wanted to open up development into a collaborative, so that people wouldn't feel like Xiph was such a huge ivory tower. I wanted to be able to welcome new developers with open arms, and give them all the credit possible. Eventually, I wanted to make sure that Xiph donations went to hardware and equipment (and honestly, a little cash) to the new people that came on-board and wrote great stuff. I always tried to bring new people into the fold, as soon as possible.
My first reaction to people saying 'Hey, this guy is doing stuff with Xiph technology that we really like' was usually, 'Okay, let's see what happens. If it doesn't break anything, let's integrate it!' This fell on deaf ears. The folks at Xiph seemed more than happy to continue making their ivory tower higher and higher, regardless of how late I stayed up talking to anyone and everyone about Xiph technology.
When you build a very high tower and put the builders in the very top, they're going to die when that tower falls. And Xiph's tower is creaking like crazy right now. While they might be scoring a ton of money from corporate interests, they are quickly losing their relevance to multimedia.
This needs to change. Xiph needs a blockbuster release, and they need it two months ago. If it's not Vorbis or Theora or any existing project, it has to be something new, and whatever it is has to be a total ass-kicker of a project. There is absolutely no point in even having a company to manage projects like this if people are finding the company's best results somewhere else.
I want Xiph to release great things and stay relevant, while keeping their heads above water financially. It's all I've ever wanted, whether I was the guy in charge or just a guy watching from the sidelines.
I'm terrified. I hope they make it.
Wie ihr bestimmt mitbekommen habt, kommt die Entwicklung von Ogg Vorbis von offizieller Seite nur sehr langsam voran, so liegen die Releases beispielsweise sehr weit auseinander, Vorbis 1.0 erschien im Juli 2002 und das erste Update 1.0.1 erst im November 2003. Auch von geplanten Neuerungen oder damals angekündigten Features wie Bitrate Peeling hört man wenig.
Seitdem sind einige Versionen von Drittentwicklern erschienen, die den verfügbaren Quellcode auf eigene Faust verbessert haben. Ich denke da an GT3 beta 1/2 von Garf, der die Qualität bei hohen Bitraten stark verbesserte, und besonders an den aoTuV-Encoder, der aufgrund des Sieges beim letzten 128 kbps-Hörtest plötzlich in aller Munde ist.
Das Problem dabei ist, dass diese Verbesserungen nicht in die offizielle Version integriert wurden, und damit nur wenige Nutzer wirklich von ihnen profitieren, denn nur eine kleine Minderheit dürfte über diese Entwicklungen auf dem Laufenden sein. Xiph.Org hat damit möglicherweise viele Chancen verschlafen. Ein weiteres Problem ist die bisher mangelnde Koordinierung der verschiedenen Encoder, die Entwickler beginnen erst allmählich und noch sehr zögerlich damit, untereinander zu kommunizieren und ihre Ergebnisse zusammenzufügen.
Quo vadis, Vorbis?