topic says it all.

As far as I know noone is working on it, so the question makes no sense.

I thought I remembered reading a while back about it being worked on, that it would streamline the decoding so it would take up fewer resources...

Probably this info is too old but in Ogg Traffic for Wednesday, February 25, 2004 they said:



Oggfile is a nicely working format to store Vorbis, which everyone loves, but this file format is--or, at least, was--not flexible enough as a general-purpose multimedia container. Note that I'm talking about the container (OggFile), not codec (Vorbis). Vorbis is not bad at all.
Among other things, OggFile couldn't store subtitle streams properly as you can read in the above Ogg Traffic, which I pesonally experienced too when handling OGM 0.9.9.6. OGM was not an official xiph format but it was based on OggFile, sharing the exactly same problem (i.e. 2 subtitles can not overlap timeline-wise.)

There was a possible hack for OggWrit, xiph's subtitle format, to solve this problem, but afaik xiph ppl were going to update the Ogg format itself, not just using an ad-hoc hack, to make it more flexible; even so, they were not going to change Vorbis I, as an exception, for backwards compatibility, because it was already very widely used.
In other words, if there was "Vorbis II" it would be probably not backword-comaptible.
But that's just hypothetical.

Let's consider the circumstances overall, doing something simple as changing the window function, would break backwards compatibility. If I understand the situation correctly.

Xiph.org were interested in publishing a subset of Vorbis I, which would make decoders simpler. Nobody's started working on it yet, AFAIK.

It seems to me, like if all Xiph.org projects were dead. Almost nobody is working on Vorbis, Theora isn't moving any further and has really low quality. And I haven't heard any news about FLAC and Speex neither.

I don't think so. For instance, look at ffmpeg2theora and illiminable's filters for Ogg Vorbis, Speex, Theora and FLAC. Theora is not very hq, but it's decent if you are a good encoder and encoding speed is very fast, and it's totally free. Altho being alpha, it should be already usable for some purposes.

If you want to know what's going on at Xiph.org...

http://lists.xiph.org/mailman/listinfo/commits

Thanks for making me realize that Windows was dead as well (I mean, the last release was 2001!). Oh, and BTW, the last Speex release was 2 months ago, but I'm sorry I didn't announce it to you personally. If you leave your phone number, I'll call you before the next release, promise!



While we're at it, could anyone tell me if Windows 2012 will be compatible with Vista? It's really important for me to know that now!

Someone seems really ticked off here.

Anyway, like some pointed out, Theora isn't dead (yet).

Aoyumi is doing a great job optimizing vorbis with his aoTuV encoder. He's working on another alpha right now, I believe.

Also, Thomas Beck has been working on finishing up a really powerful new lossless codec (codenamed YALAC atm) that could potentially be integrated with FLAC at some point.

But even if there hadn't been significant updates recently, that doesn't constitute anything being "dead", depracated, obsolete, etc. Why fix something that works fine?

Speex 1.1.12 was released on February 2006. (2 months ago)
Vorbis 1.1.2 was released on November 2005. (6 months ago)
Theora 1.0 Alpha 5 was released on August 2005. (9 months ago)
FLAC 1.1.2 was released on February 2005. (14 months ago)

In the monthly meeting of March 2006, Monty said that libogg2 is the priority right now, and Theora 1.0 is the end goal... so if you think that all Xiph.org project are dead... then you are totally wrong... Xiph.org is know for not make releases too often... but why do you need a new release every month?

I'm not sure about how are going the things with FLAC but probably Josh is working hard on it...

If you said that nobody is working on Vorbis, then you are wrong again... Aoyumi is working on Aotuv beta 5 right now... yes, he is probably not a member of Xiph but Monty said that probably the next vorbis release will merge the aotuv tunings (I don't know if will be beta 4.51 or beta 4).

Theora is moving slow but if you said that has really low quality then again you are wrong... the quality of theora is quite similar to XviD or DivX.... maybe can't compete with h.264 but based on that theora is based on VP3 that has five years old and the quality is pretty decent... Theora 1.0 is the end goal for Xiph.org right now... and will be ready probably this year...

Haven't heard any news about Speex?? yes... you should give your phone number to Jean-Marc... 1.1.12 was released 2 MONTHS AGO.


Back to the main question... there is a simple answer... probably not... but sice Vorbis II is so far vapourware you should not worry right now...

I'm sorry if I have been to hard by saying the projects are dead. I certainly didn't mean it so. I'm glad that the work on Speex codec continues, thank you for it. Unfortunately I haven't looked at Speex's webpage lately. I expected somebody to announce any new releases here at HA.

Btw. have you ever talked with Matroska developers about Speex support? This would be very useful for storing speech-only audio tracks.

Anyway, what I think is really a shame, is the fact, that there is not much progress in Theora development. Compare Theora to XviD or x264, there is a huge difference in quality. I know it is easier to develop a codec if you have defined standard, but Theora didn't start from zero. It's based on VP3 codec. What might help to spread Theora more widely (and hopefully accelerate it's development) would be a browser plugin (Firefox, IE, Opera at least) for OGG-Theora/Vorbis/Speex with http streaming support. Now the only 2 formats which allow this are ASF-WMV/WMA and AVI-DivX/MP3.

I think what all Xiph.org projects need is more support for and users. I think a very good example is the Matroska team, who coded a lot of usefull applications for their format.

jorsol> I posted my answer before I saw your post. I know that Ayoumi is tuning Vorbis encoder, but it seems to me like if he's the only one working on it. Anyway he does a great job and thanks to him Vorbis is the best lossy audio compression for most bitrate ranges.

It would be really nice to see Theora 1.0 release this year. But I don't agree with you that you can compare Theora to XviD or DivX. Have you seen Doom9's codec comparison? I think Theora needs a lot of work before it can compete with XviD. In all four main aspects - Quality, Speed, Size accuracy and usability.

I'm a big fan of all Xiph.org projects, so I was quite dissapointed when I haven't seen much progress in the projects lately. I'm glad to hear that all the projects are still in active development.

Btw. What will be new in libogg2?

Yes, Theora need a lot of work... and thats why it is still in alpha...

for me the quality of theora is ok... remeber that is based on vp3 that has 5 years old and don't have almost any quality improvements since then (imagine the potential of theora if was based on vp6 or even vp5).... XviD on the other hand is constantly improving it quality.... but when theora hits 1.0, I really hope that Xiph focus in improve the quality.... also note that maybe only 3 or 2 people are working on theora.... I agree that Theora can't compete with h.264 (here is a huge diference in quality).... but still have a chance to compete with XviD...

I agree with you that Theora need lot of work before compete with XviD, but is not so behind it...

Xiph.org don't have an army of developers, is a is a non-profit corporation and they make pretty much without requesting nothing in return.... everybody that want to help is always welcome... I don't imagine how can vorbis compete with modern codecs without the aotuv tunings....

Probably the notable change is low memory footprint.... I'm not sure about other things....

Come on, are you talking about Ogg Vorbis without even knowing about Lancer? *sigh*
And not Ayoumi , A-O-yumi, hence A-O-tuv


Have you ever encoded video with Theora yourself? Or have you just read or heard about it in doom9's? What you are assuming is basically true, but you were wrong when you said "really low quality"
Check the newest comparison, and play the real sample clips. Theora is not HQ, but can be about as good as XviD. In other words, today's XviD is not the best codec anymore, when compared with x264, rv10, vp7, SNOW, and eventually Dirac. Note that I love XviD, I use it most often myself, and I'm even hosting the newest cvs versions. and XviD is getting better too. Still, the fact is the fact. Again, theora is not very hq. Maybe it was the worst among the 7 or so codecs in that test. Btw, Theora is based on VP3, and here's the same video encoded in Theora (~VP3), VP6, VP7. Considering that the quality today's users would want is the one of VP7 or x264 (AVC), Theora is really handicapped. VP7 is obviously better than VP6. And probably VP6 is better than VP5...and so on, down to VP3 ^^;

Nevertheless it's an exaggeration to say "really low quality" about Theora. Theora's quality is decent. Besides, in the above testing, Theora was encoded by 1-pass. The quality can be significantly better once ffmpeg supports 2-pass Theora encoding.

Finally, when you said "Quality, Speed, Size accuracy and usability" you missed one point; unlike MPEG (XviD is an MPEG-4 Part 2 codec in case you didn't know), Theora is patent free. You can distribute theora encoders and decoders freely. If you are going to compare MPEG with Ogg, please consider that point too. Beause of that very reason, some games use Ogg Vorbis, because they don't want to pay the MPEG license fees. The similar thing might happen in video, and so there is one important point about Theora besides "Quality, Speed, Size accuracy and usability".

This was posed a few days ago.

http://wiki.xiph.org/index.php/Summer_of_Code



Ogg Dirac ! sounds interesting...

The Schrödinger Project

Ogg Dirac work is being done by the BBC and Fluendo.

Liisachan> First of all sorry to Aoyumi for misspelling his name, I know his name, just misspelled it. Unfortunately I don't know about Lancer. I'm not so much involved in audio compression formats. But I know more about video compression (you've probaly seen me at doom9). I always do some tests myself if I want to prove something. And I was really thinking about OGG Theora for distributing video.

I work for official webpages of an ice-hockey club in our city. I want to set up a new and better multimedia part of the web. But unfortunately I need all people to be able to playback my video files withou to much effort from their site. This is the main point where my idea about OGG/Theora/Vorbis failed. I think that the support for end users sould be also one of points that Xiph should concentrate in.

Btw. OGG Dirac sounds also interesting for me.

Isn't that more up to the creators of the software being used to play it back? I have this funny feeling that Microsoft, for example, isn't going to go too far out of their way to support the enemy. Ogg is philosophically polar opposite to WMA, and it competes with their interests. The same can likely be said for other manufacturers, and many other codecs.

yes... Microsoft will never add native support for Ogg (or others open formats) but you can always download illiminable directshow filters, and have great support for Ogg codecs on all directshow players including Windows Media Player (that BTW sucks)....

And I'm pretty sure that if the format get more popular, some software developers will add native support for Ogg... and hardware players are going to support Theora just like DivX, with the advantage that they dont have to pay expensive patent licenses....

PS: Ogg Dirac sounds really great!!!

Sure I knew your nick, I even quoted you recently there, right? Gabest's VP62-clone decoder working better than On2's official decoder, that was a really good one And I can tell that you are not interested in xiph projects too much from the fact you keep spelling Ogg as OGG.

You are right, but that's another thing. You were talking about "really low quality", not about player support, weren't you?
I agree with your point. Recently I had to post a sample clip in the doom9 forum and here in HA, to confirm the fact that LAVC in ffdshow is buggy for 6-ch Vorbis. I could have tried to make my sample (Theora+Vorbis).ogg but I didn't. I just made it (XviD+Vorbis).ogm. I even avoided .mkv. Because a long explanation would be needed to play theora, whereas for ogm, I can just say "Use MPC".
http://www.hydrogenaudio.org/forums/index....showtopic=43983

Anyway, if you are "not so much involved in audio compression formats" I don't think you can judge the whole situation and say "It seems to me, like if all Xiph.org projects were dead."
But on a second thought, it might be that the Czech word for "dead" is not that strong, and you might have just meant "inactive" or something.




There was an intersting "happening" about that

[Advocacy] Is Microsoft being cruel to OGG?
http://lists.xiph.org/pipermail/advocacy/2...ber/001223.html

last I heard of vorbis II was back before I was even final

some ppl are actually using theora and it 'streams' pretty well using java applet, examples:
http://www.kiberpipa.org/video2.php

You can also track Xiph commits in real-time:
http://cia.navi.cx/stats/project/xiph

It's not working very well for me. The movie is choppy and I can't seek in it or change volume. Also there are many people who don't have Java installed. If I compare it to DivX plugin (see example), it brings much better comfort. I would really love to see a plugin like this for Ogg.

Liisachan
By saying "really low quality", I didn't mean that the quality is not acceptable, but it cannot be compared to any modern codecs like XviD, DivX, x264, VP7, RV40 or WMV9. Three years ago it would be one of the best codecs out there. But I think Theora still has about the same quality as VP3. But of course the price is unbeatable :-) I'm really looking forward to see what we'll have after this year's Summer of Code.

Btw. I can also confirm some garbage in libavcodec Vorbis decoder.

I didn't mean "Theora is hard to play." Generally, I could just say "Use vlc" and it works out of the box.
In the above case, I need to ask others to play movie via DirectShow, not just playing but playing via DS, to test a certain DS Vorbis decoder, and playing Theora via DS is a bit more complicated. Essentially it's easy tho. The fact is, Gabest once tried to let MPC splitt Theora in OggFile too, just like OGM, but then:
http://sourceforge.net/forum/message.php?msg_id=3633754


Theora in other formats (like DSM) plays using MPC's internal splitter, but not Ogg Theora. I'm not really sure, but this might mean that, even tho Theora itself is ok, OggFile as a file format needs some modifications. If that happened, the new Ogg Format, including "Ogg Vorbis II", would be not back-compat.

I earlier already mentioned, that Ogg formats are not very user-friendly. I give you one example why. Today I downloaded some files from this site. I'll let you know my experiences.

First of all Firefox offered me to open the files with Apollo (my default music player). I had to manually choose a video player (ViPlay, ordinary DirectShow player). It didn't play. So I downloaded the file to my HDD and did some tests.

Firstly I tried to open the Ogg file in some audio players: Apollo, Winamp, XMPlay, foobar2000 - nothing played. I expected at least one player will start playing the audio track. I tried some video players - VLC - worked O.K., MPC - worked, all other DS players failed.

Why? I have installed Illiminable filters, Haali's splitter and FFDShow (Thera and Vorbis allowed). But all DS players will load Illiminable splitter, then connect audio and video to Illiminable DS decoders. For an uknown reason this graph doesen't work, I even tried it in GraphEdit. AviSynth with DirectShowSource failed too.

MPC played file fine, as long as I used internal Ogg spliter and FFDShow decoder. If I disabled FFDShow, MPC wasn't able to connect to Illiminable Theora decoder. The same happend to Illiminable Vorbis decoder if I disabled MPC's and FFDShow's decoders.

I finally found out, that all I need to do is to rename the file to *.ogm. Then Haali's splitter will be used together with FFDShow and it plays almost fine. I say almost, because both FFDShow and VLC used my whole CPU (Athlon XP 1,3 GHz) and the video was still choppy (standard NTSC 640x480, 30fps). MPEG-4 ASP would use about 20% of CPU on my machine.

So my suggestions to developers:

1) Fix Illiminable filters, also make them so, that they can connect to other Ogg splitters and fix the Ogg splitter to allow connection to other decoders (such as FFDShow).

2) I would really suggest to use OGM extension for video files. For most people OGG = Ogg Vorbis and it is an audiofile. I have my audio files associated to a simple audio player (Apollo) and video files to ViPlay. Most people use different players for audio and video IMHO. Matroska uses MKV extension for video and MKA for audio. MP4 allows the same confusion as Ogg, but many people are using M4A for audio-only MP4s.

3) Why don't audio players play Vorbis streams in Ogg file? It works O.K. with Matroska (at least in foobar2000).

4) As I already mentioned a web plugin for OGG would be a really good idea. Something like this DivX plugin.

5) Haali could add support for OGG extension to his filter.

Ogg Vorbis format is very user-friendly. Theora is not, because it in development now.

Yes, you're right. But I think that some of the problems I had are also related to OGG multimedia container. I'm looking forward for the final release of Ogg v2 and Theora v1. I just wanted to let the developers know, that there are some problems which need to be fixed.

http://www.hydrogenaudio.org/forums/index....showtopic=41494

This is being worked on, in SVN. /branches/vorbis-tahseen/

vlada, let's just stop the whining and say you are not satisfied with the current (player) support for Ogg formats.


Keep in mind that OGM is not a Xiph-blessed container derivate. It was born to allow storing other (non-Xiph-format) streams by signalling its use via an AVI-like stream header so one could couple DivX with Vorbis for example. You're not the first one that is suggesting different extensions for different media types. Of course, you can rename the files like you want if it makes your life easier, but the official ending for (real!) Ogg streams is "ogg". See here.


Possibly because they're expecting a so-called "degenerate ogg stream" (a stream consisting of exactly one logical stream). Degenerate Ogg streams are easier to parse / seek in. I'm pretty sure that most mp3 players can't play an mp3 which is inside an MPEG1 system stream. Foobar2000 is possibly the only audio-only player that supports Matroska. Usually audio-only players don't give a sh** about containers. I'm sure WinAMP has no abstraction/interface for containers -- just input plugins


Good idea! Are you volunteering?

Something on-topic: I'd like to invite you to read Wikipedia's definition of "backwards compatible" which states "Backward compatibility is a relationship between two components, rather than being an attribute of just one of them [....] Data does nothing in the absence of an interpreter, so the notion of compatibility does not apply to document files, it only applies to software" So, strictly speaking the thread title doesn't make much sense. The OP either wants to know if a Vorbis 2 decoder is able to play Vorbis 1 streams (in this case the vorbis 2 decoder would be backwards compatible) or if a Vorbis 1 decoder is able to play Vorbis 2 files. Concerning the latter title interpretation: Can't tell since there isn't a Vorbis II yet. Though, it's highly probable that they won't be. And even if they are playable you won't get the full quality (like HE-AAC playing on an LC-AAC decoder).


Sebi

No, it is not a problem of a certain players. There are players, which have a workaround (like MPC). But generaly no DirectShow player will work with Illiminable filters (at least on my computer). Those filters are linked from Xiph.org page. There are no official DirectShow filters. There is one more alternative from Radlight, but doesen't work for me either. What I'm trying to point out is, that there is no support for end users from Xiph.org and that third party players/decoders don't work correctly.



I know exactly what is OGM and I also know, that the only official extension for Ogg Theora is OGG. But I don't think it is a good idea. You have .FLAC and .SPX extension for FLAC and Speex. So I think there should also be another extension for Theora, OGM was just a sugestion because most people know, that OGM is a video file. Having the same extension for video and audio files is really inpractical a nobody else uses it.




Winamp supports Matroska through DirectShow. But you're right, that most audio players have a separate plugin for each single format. But why an OGG Vorbis plugin shouldn't be able to play a Vorbis soundtrack and ignore all others tracks. Unfortunately I don't know much about degenarate ogg streams. But as kjoonlee said, Xiph.org knows about this issue and works on it.


I would, if I were a better programmer.

P.S. I'm not requesting anyone to do anything. I'm just giving suggestions and I expect a discussion about them.

For Ogg Theora I have been using *.ogv (.ogg opens in winamp, and it can't handle theora)

Illiminable's filters DO WORK. Just make sure they are the last ogg filters you have installed (install them after oggds/corevorbis). Otherwise the registry entries will be overwritten. Illiminable's filters are extention dependent, they will handle .ogg, .ogv, and .oga. *.ogm shouldn't be used for 'real' ogg files.

That reminds me of the file association dialogue of WMP that doesn't know *.ogg, but *.oga and *.ogv. I never heard of these suffixes before.

The big problem is not to have only one extension for all formats, if I have an Ogg Vorbis or an Ogg FLAC or an Ogg Theora with the *.ogg extension, illiminable filters will open all without problem in any directshow based player.... the problem is that audio-players that are plugins-based, don't care the container format.

This is a workaround for winamp...
- install last illiminable filters...
- open winamp and go to preferences.
- go to Plug-ins - Input
- configure "Nullsoft DirectShow Decoder"
- add ";OGG" (without the quotes) to the list...
- done, now you can play Ogg Vorbis, Ogg FLAC and Ogg Theora using Winamp.

PS. btw illiminable filters are the most official directshow filters from Xiph.org, they dont have to be created by Xiph.org or a top member to be the official one... illiminable has access to SVN of Xiph and all the progress is done there, illiminable is practicaly a member of Xiph and he work close with Xiph.

and remeber that the 0.x means that the filters are still in beta... is to be exprected that the 1.0 version of the filters will have a complete support for ogg codecs, user-friendly and stable.

gameplaya15143
I don't have OggDS or CoreVorbis installed. I had Radlight filters, but I unregistered them. I also have Haali's splitter and FFDShow with Vorbis and Theora enabled.

jorsol
I know how to enable DirectShow decoding for a certain extension in Winamp. But it won't work for me as I already mentioned.

All
So the videos I linked work for you with Illiminable filters? Then there must be something wrong in my computer. But all other filters work fine.

I don't know about speex, but .flac is for native FLAC. Ogg FLAC is just .ogg. For degenerate streams, .flac.ogg (etc) might be useful way to do it.

Josh

Stupid waste of time. Developers need to write software that recognize the difference between a Vorbis stream and a Theora stream. You only need one stream container. It makes life so much easier. What is going to be tricky is designing libogg to support Theora + Speex data, chaining physical bitstreams, etc.



If you change the window function or implement some sort hybrid filterbank it won't be possible. The focus for Vorbis II should be rewriting the encoder to support VQ books more effciently and to have full 5.1 coupling. That's just my honest opinion.

What's the problem with "VQ books" ?
Also, the Vorbis I format looks like you can do proper 5.1 coupling. I guess it's just the lack of an encoder that's smart enough.

No one's going to use something like Vorbis II if it's just Vorbis I with some incompatible minor fixes. And that's probably why Monty's currently on a complete different path. See here or here.

MP4 does. You regularly see both audio only and video mp4s online, audio only almost always AAC but now and then mp3. While it caused some confusion a year or two ago, the dust seems to have settled down now.

It can be done with effort on splitter makers' parts; perhaps Illiminable hasn't done it because theora simply isn't efficient or fast enough for real use compared to other codecs (unlike vorbis) and is hardly found anywhere. Although they did make the DS decoder... *shrug* Maybe you should email them about it.

I don't think so. Most people have no idea, what MP4 is. People who have already seen it usually think it is a video compression. I see a lot of computer users looking for "MP4 codec".

And it is similar with OGG, most people think it is a sound compression. They're confusing OGG and Vorbis.

I really believe, that 99% of computer users have no idea what a multimedia container is. Most people never heard about OGG Vorbis a most people don't know that Windows is not the only one operating system in the Universe. I think this is the sad truth....

AFAIK, Ghost codec is just a proof of concept (like Ogg Tarkin), and probably will never see the light... or at least not first than Vorbis II...

Maybe this codec serves for inspiration for Vorbis II, maybe not... but the only thing sure is that Vorbis I still has a lot of room for improvements, like 5.1 channel coupling and in the quality of the low bitrate area... so I don't see the reason why we should bother to us about the compatibily of Vorbis II, when Voribis I will be here at least 5 years more.

@vlada:
this applies to illiminable's oggcodecs and radlight too (IIRC):
If their splitter doesn't parse the file, then their decoder will NOT be used. They will not work with other parser filters. Do you understand what I'm saying here?

The vorbis decoder in oggds, corevorbis, and ffdshow will decode vorbis in mkv.

Since there isn't any standalone directshow parser filter for ogg that knows what theora is, ffdshow can't decode it. If I'm not mistaken, it will only decode theora in AVI. (..at least as far as I know)


/me drools at the thought of tarkin actually being usable

ffdshow can play theora in MKV too (using Haali Splitter)...

Also note... Haali Media Spliter can play Ogg Theora and Ogg Vorbis, using ffdshow only... you don't need to install illiminable filters if you don't want, you only need last ffdshow and last Haali Splitter.... you can even use CoreVorbis for the audio if you disable Vorbis in ffdshow....