What is going on with Vorbis? (or Xiph overall) Options

[rjamorim]

Today I was checking RareWares' links page to find broken links, and happened upon Xiph's page. I noticed the "news" there were 3-4 years old.

So I was wondering, does anyone, by any chance, know what is going on there? Are they working on Vorbis2, or Tarkin, or Theora? Have they disbanded and only the web pages remain? I remember that in Emmett's time, he would feed us news non-stop, mostly from #vorbis (he was a little too loud, matter of factly :B). Now, we listen nothing but crickets.

Do we have some insider here that knows if something is going on there, at all? Maybe Coalson or Valin?

And what the hell is Emmett up to anyway?

Thanks for any info.

R.

[jorsol]

Well I try to follow the progress in Xiph by the stats of SVN Xiph CIA and the Monthly Meeting.

I found that there is almost null development for Vorbis and Theora...

Tarkin is just vapourware and is not developed anymore... Theora has almost no development (sice alpha7), there is a fully functional decoder in the theora-exp branch but the main library (libtheora) is just incomplete, and over the years mainly a few bugfixes and no quality or improvements.

Vorbis II is just a bunch of ideas and graphs but nothing funcional (vapourware too) or in active development, note that the name of Vorbis II will probably not be used because Monty dont like it, it will be know as "Ghost", so the correct name for Vorbis II is Ghost...

In the mailing list Ralph Giles said that they have not merged aoTuV R1 (or probably b5) to the main libvorbis because they have to check the changes and Monty is the only one that can do that... this is what Ralph said:

We agree the changes should be merged. The last time that was done, Monty
was able to further improve the code on both sides. We'd like that code
review to happen again before merging the aoTuV branch into the
official reference encoder. However, Monty is the only one of the
current volunteers we trust to do that, and he doesn't have time in the
foreseeable future. So while it's on the todo list there is no schedule.

and because he has not time to do it they have not merged it yet. Is just curious that Monty don't have time to do it when I see a lot of submits from him to SVN for "sushivision"...

In other words there is no future for ogg codecs if their creators don't care about it... they work on it someting like a hobby... probably they don't even know the potential that their codecs have in the world... is a shame...

This post has been edited by jorsol: Mar 5 2007, 15:06

[kjoonlee]

Vorbis II is just a bunch of ideas and graphs but nothing funcional (vapourware too) or in active development, note that the name of Vorbis II will probably not be used because Monty dont like it, it will be know as "Ghost", so the correct name for Vorbis II is Ghost...

I thought Ghost was Speex II, with the Vorbis I psymodel.

[xiphmont]

Vorbis II is just a bunch of ideas and graphs but nothing funcional (vapourware too) or in active development, note that the name of Vorbis II will probably not be used because Monty dont like it, it will be know as "Ghost", so the correct name for Vorbis II is Ghost...

I thought Ghost was Speex II, with the Vorbis I psymodel.

No, there was a seperate thing where JM borrowed some parts of the Vorbis psymodel to add to Speex, and we presented a paper about it. Ghost has nothing to do with that.

Ghost is at the same point right now that Vorbis was in late 1998, early 1999. There will be nothing anyone outside the org will see [well, care about] for the most optimistic minimum of a year. Vorbis's failures in hardware adoption was not because the industry hated us. There were several technical reasons Vorbis was hard to adopt for low end hardware manufacturers, and that's something we have to fix in time for Ghost. The other reason for Ghost is that I don't want to keep tweaking the same old, I want to make a very big tech leap... the difference between replaying old levels over and over to squeeze out an extra second on the timer, and actually levelling up...

The primary reason it will not be named Vorbis II is because the codecs have virtually nothing in common. I know that doesn't stop MPEG, I don't really care.

And for those who care, Sushivision is a visualization tool I built as part of the Ghost dev effort. It's not meant for external consumption, at least not yet. It's still cumbersome to set up.

(Oh hi, I'm back. And it's nice to see you're still around Emmett.)

>This sadly means that there's apparently little time to e.g. review the AoTuV changes.

Then the reigns should be passed on to a suitale replacement (advertise for one), having someone who is effectively a spanner in the works (nothing against monty, but that is how it appears) is counter productive.

Well, in terms of Vorbis, that is what Aoyumi has done. I offered making AoTuV an official Xiph project (or an officially sanctioned external project, or some sort of official status) and he balked at that. I don't blame him. He's doing his own thing, doing it well, and enjoying himself. Why screw it up?

I've told anyone who's asked that AoTuV is the best of the Vorbis encoders out there that I know of. There is nothing dodgy or unofficial about it. I'm enjoying having a healthy third-party project advance the encoder.

As for me, I'm going to continue doing what I enjoy, and that's the codec research. I have no interest in being an executive. It doesn't make me happy. Being an R&D engineer makes me happy.

This post has been edited by xiphmont: May 11 2007, 23:44

[Fandango]

The primary reason it will not be named Vorbis II is because the codecs have virtually nothing in common. I know that doesn't stop MPEG, I don't really care.

"Common folks" will still call Ghost "teh .ogg" anyway, just like it is with Vorbis at the moment (although that has changed a bit over the last couple of years).