Ogg Tarkin?

[Laz]

Is there any news on Ogg Tarkin?
I'm a fan of the the Ogg Vorbis Audio codec so I'm excited about Ogg Tarkin.

Thanks in advance.

[Neo Neko]

Tarkin is years away from having anything in terms of beta or alpha code. They are still in the idea gathering stage by and large. Though we should see some movement on the over due Theora project. By emmett's schedual it was to have hit 1.0 this month IIRC. But emmett is no longer with them so things probably have changed. Though now that Vorbis 1.0.1 has come and gone Theora will probably make rapid progress.

[gino25]

what is ogg tarkin?:confused:

[bilu]

http://www.theora.org/theorafaq.html

[Tommy Carrot]

Xiph.org completely abandoned Tarkin project. I think the closest to what it could've been is Rududu (http://rududu.ifrance.com/rududu) codec, it's the only working wavelet codec i know about.

[Atamido]

Originally posted by Tommy Carrot
Xiph.org completely abandoned Tarkin project. When did that happen?

[Tommy Carrot]

Originally posted by Pamel
When did that happen?

Ok, i don't think they announced it, but the mailing list disappeared, and the source code is untouched years ago as far as i know.

[Neo Neko]

Originally posted by Tommy Carrot
Xiph.org completely abandoned Tarkin project. I think the closest to what it could've been is Rududu (http://rududu.ifrance.com/rududu) codec, it's the only working wavelet codec i know about.

The mailing lists are quite slow. But they (http://www.xiph.org/archives/tarkin-dev/200312/index.html) still exist. And the last few mails give clues to the state of events. No tarkin is not abandoned. Just not a hot topic at the moment.

[DevilsChild]

According to this (http://www.xiph.org/archives/tarkin-dev/200311/0002.html) message, no work has been done on Tarkin since at least November 2002.

[Neo Neko]

Yes but that could mean any number of things. It could coincide with the dates between the release of Vorbis 1.0 and 1.0.1. Very very likely indeed. It could also coincide with the change over of C.E.O.s. This (http://www.vorbis.com/ot/20030429.html#id2750093) tidbit right smack in the middle of the gap on April 29th could be informative. And don't forget their work on the portable firmware and integerised decoders. Xiphophorus has been quite busy. Especially considering that they are a not for profit group with very few actual employees. If a third party were wanting to work on it I am sure they would let them with some supervision. But there are still some things to tackle first. Like how exactly is this all going to be contained in OGG and how will that effect the codec. Which is what Monty should be working on ATM.

[Atamido]

Originally posted by Neo Neko
But there are still some things to tackle first. Like how exactly is this all going to be contained in OGG and how will that effect the codec. Currently, the plan to store Theora and Tarkin in Ogg is the same as Vorbis is stored.

For those that don't know, Ogg does things a little differently than most containers. Most containers store some codec initialization data in the headers. For instance, "Use this codec with these settings to decode." However, in Ogg, you just have a raw data stream from the codec. So, the initialization information is inside of the streams packets somewhere. There is not a generic way to store it either, so it becomes difficult to identify the codec used or anything about it. In the case of Vorbis, the first three data packets will be all of the initilization information. Tarkin/Theora will likely work the same way, with however many initilization packets. When you want to read an Ogg file, a generic libogg will pass the first packet to libvorbis, if it doesn't like it then it will try libtheora, and so on until something works.

Note that Ogm doesn't work this way because a generic method was defined for storing initilization data so that the proper codec could be identified immediately using the FourCC. Ogm is an extension of Ogg in this respect.