OGM is not dead - oggds now in xiph CVS

[kxy]

Thanks much to Tobias for donating his DirectShow filters to Xiph.org!

You will find the code from Tobias in the 'oggds' module of CVS. This code represents the 0.9.9.3 version of the filters, which is not quite what the current version was. Apparently Tobias was in the middle of some sweeping changes and was not able to recover the new version for us.

In any case, I encourage everyone interested to play around with them. They are BSD licensed and should make a great starting point to finish a fully function set of filters for the Windows platform.

Jack Moffitt

[Koepi]

Just got a mail 1 minute before your post in this forum where Tobias wrote about that :)

I'm sorry that I had to nag Tobias with that(well, one mail asking about releasing the sources isn't really nagging), but now really noone can complain - OGM is alive and kickin'!

Koepi

[bond]

wow real great news!!!

hopefully someone will continue Tobias' work (matroska team fasten your seatbelt ;) )...

[iago]

Yep, that's good news! :)

[Animaniac]

Sweet!!!

[Ghim]

It's really something good to hear...

There's a lot of great programmers here... Let's hope one of them will take this project and continue it !!!

[Atamido]

Originally posted by Koepi
OGM is alive and kickin'! From hell's mouth I stab at thee. Die OGM, die!! :devil:

Just kidding. But seriously, this is good news. Competition is good. The fact that OGM can do chapters is one of the driving reasons to hurry up and get chapters working in Matroska. We don't want to completely pass up OGM in a week and have nothing to compete against. Maybe we should pull Toff away from Matroska and have him focus on fixing OggDS. ;)

[Garfield]

That is such a good news, i hope Gabest was told about it :D

[avih]

at last :) great news indeed.

[Animaniac]

Originally posted by Garfield
That is such a good news, i hope Gabest was told about it :D

Haha, ditto. I wonder if he'd be willing to work on the filters.

[BoNz1]

Great news, good to see that all of tobias' work and the work of everyone that tested and helped out with ogm is going to be put to good use by the people at xiph.

[robUx4]

Originally posted by BoNz1
Great news, good to see that all of tobias' work and the work of everyone that tested and helped out with ogm is going to be put to good use by the people at xiph.

Mmm, that's not what the announcement says... Xiph haven't assigned anyone to it. And I doubt they will ever do. So they hope someone else will do their work. It will not be dead if someone take the responsability to support it and make it evolve.

But it's good that the situation is finally clear(er).

[BoNz1]

I would hope development would continue, at the very least it seems they are interested. You are right whether or not it actually becomes the container that hold theora, I guess time only will tell.

[ChristianHJW]

I should maybe not comment this thread here, but this is what i liked to say :

I personally see OGMs future as a supported format from Xiph, or not. I may be wrong here, but this is my opinion. Just think of what happens if libogg is no longer supported, and OGM become incompatible with oggfile, which is incorportating Theora and Ogg handling in general.

Monty has recently stated on #vorbis that IHO Tobias has done a great job to stick to the Ogg specs as close as possible, as far as he can tell. Maybe only the 'mapping' of the DivX codec had to change.

Now i bet every sum that Monty

- has no detailled insight how the VCM codec mapping is actually working in OGM
- doesnt know what a BITMAPINFOHEADER is, nor a VIDEOINFOHEADER
- has not really considered how such Windows centric structs can be generated on Linux, his preferred OS ( ask Mosu, he knows about it very well )
- will die a sudden and hard death, presumable from a heart attack, once he finds out that Tobias is fully depending on the information that DirectShow is giving the OggDS muxer filter ( again, BITMAPINFOHEADER and VIDEOINFOHEADER ) to be able to initialize the codec on playback

He maybe has a feeling that this is the case now, thats why he said the 'codec mapping may change', but guys .... if he does change that, all your existing OGM may be supported for playback, but unevitable become out of spec for the future .....

Just my 2 cents, not here to start a flame war ( and i havent mentioned our stuff !! )

Christian

[Liisachan]

this may sound selfish, but personally I would love _any_ container as long as I can softsub in SSA :) :) Yep, that s what I want to do.

To all the godlike developers here: plz support embedded SSA in OGM! And make it unicode-base! That's all what I want....^^; It was once my dream. I gave up once. I m glad to know that it was not dead after all

[Atamido]

Oops, I misread Liisachan's comments. I have some comments later about putting SSA in OGM.

[Affar]

Why have you been so slow in macking these operation?
I have the impression that it´s beacuse Matroska has appeared
I would like to know that had happened if this didn't exist

:confused: :confused:

[robUx4]

Originally posted by Pamel
then insert the 'Matroska splitter' filter

Plz Pamel, no such advertising. We don't want to start any war do we ? I know you're from Texas, but...

[Animaniac]

I'm pretty sure Liisachan knows how to use the Matroska Muxer just as well as the rest of us... She did write a Matroska article on her page, and did some early SSA testing with said filter. -_-;;

Anyways, to bring this back on topic, I think it's only a matter of time now for OGM to match some of Matroska's features--just as Matroska is working to match (and out do) mature OGM features.

[Koepi]

Sorry pals, but could you discuss matroska related topics in a matroska thread instead of hijacking _the_ OGM revival thread? This is rude of you.

But this just proves my oppinion about certain people right again :(

Koepi

[Animaniac]

Originally posted by Koepi
Sorry pals, but could you discuss matroska related topics in a matroska thread instead of hijacking _the_ OGM revival thread? This is rude of you.

But this just proves my oppinion about certain people right again :(

Koepi

I agree that there should be less Matroska advocating across the message board, but when discussing the present and future of OGM, it's difficult not to talk about the compeition. Each format is going to push the other to become better, and ignoring that fact would be a terrible mistake. This would be like discussing the merits and flaws of Codec A without comparing it to rival B (, C, etc.). A better analogy would be not talking about OGM in AVI related threads when it was just beginning to compete with AVI, but that definately did not happen. If anything, OGM advocates should welcome the competition, and help make the format better. Furthermore, forbidding discussion of the competition shows a lack of faith in OGM.

[Liisachan]

Yep, what I meant was "It would be great if I could embed unicode-based SSA in (New) Ogg format" and I hope it ll come true sooner or later now that the codes are open

I know I can do that with other formats like AVI/MKV, but it has nothing to do with my post. That was a feature request for the "new" OGM :)

and, altho it is true that i made a small page to play with MKV, I also made big pages for OGM in 6 languages (http://ld-anime.faireal.net/guide/ogm), "Lazy Man's Guide to Ogg Media (OGM Files)"

Like I said, I ll be a huge fan of _any_ format with which i can do what i want to. I'd say "OGM is not dead" should be good news for everyone including ppl who are developing another format. For instance, if OggDs.dll is updated decently, it is possible that they dont have to work hard for CoreVorbis.ax ; I know this is a delicate topic, but I think, well, basically, there s nothing wrong with having more than one options,
and it is natural that we are looking forward to the new Ogg format which might come with Ogg Theora, as Ogg Vorbis is really nice. (tho, technically, a container and a codec is different)

Go, fish!

[ChristianHJW]

Originally posted by Koepi But this just proves my oppinion about certain people right again :( Koepi

I personally apologize for this behaviour from our team members Koepi. I dont know what was driving Pamel, Steve was even dropping an email to our MLs about how to behave here in this thread, and even then he was posting about our container again, which really shouldnt have happened at all.

Maybe Pamel was surprised to see Liisachan requesting SSA support in OGM, i guess we were all too certain that the whole anime world was just waiting for us, but Liisachan is right if she as a user wants to have the choice, and we have to accept that.

Again, sorry .....

[bilu]

Imagine if the XVID crowd starts flooding DivX forums... it would be as justified as what you matroska guys have been doing in this thread :rolleyes:

Please redirect users to your threads as much as you can, so you can leave space for the competition to grow ;)


Best regards,
Bilu

[corigan]

I for one love OGM, and love the news. I'm glad tobias has given oggds to the open source community for further advancement, cheers tobais.

[avih]

it's not a black and white world. i don't think mkv discussions should be forbidden on this thread, but mkv supporters should apply some common sense before advertising (not talking about a specific post). that's it. be nice and considering to eachother and the rest will come along.

imho, having 2 able 'competing' containers will only do well for our community.

nevertheless, i still think that this thread should keep the focus on ogg/ogm.

[Hiro2k]

A bit back on topic.

We should make a quick checklist of all the features that are broken in OGM right now, and get to work on those ASAP.

Muxing AC3 and Ogg together
Supporting the <color> tag in SubtitleDS

Keep adding guys

[Liisachan]

Originally posted by Hiro2k
A bit back on topic.

We should make a quick checklist of all the features that are broken in OGM right now, and get to work on those ASAP.

Muxing AC3 and Ogg together
Supporting the <color> tag in SubtitleDS

Keep adding guys

Something that I want to see in the new OGM (I am not a developer, so some of my wishes may be too much or might be silly)

SUBS:
1. <font> tags in SRT(not only color, but fontsize etc)
2. positioning like ASS's {\an}
3. "subs collisions" so that 2+ ppl can speak at the same time in the movie (today's OGM don't accept SRTs which include overlapped data in timeline)
4. UTF-8 support; without which, OGM has no future about i18n/m17n. Many minor languages don't have their proper code pages, but have all Code Points needed scattered in Unicode table. For example, I happened to get a script in Tatar (a minor language spoken in a part of Russia), and was disappointed to find I can't put Tatar subs as SRT in today's OGM no matter what. (Hence I softsubbed in Tatar using another format.)
5. SSA/ASS and coming USF

AUDIO:
1. AAC (already ok?)
2. MP+
3. FLAC

CHAPTERS:
Which charset should we use in Chapters? I guess UTF-8, while SRT doesn't accept UTF-8. This should be unified.

PS. I do love .OGX import/export--one of the OggMux 0.9.5's new features; that is really convenient when you want to mux the same files again and again for some reasons (correcting typos in an srt etc)--especially, with OGX, you have to type about chapters only once. Really cool.

[Animaniac]

Originally posted by Hiro2k
A bit back on topic.

We should make a quick checklist of all the features that are broken in OGM right now, and get to work on those ASAP.

Muxing AC3 and Ogg together
Supporting the <color> tag in SubtitleDS

Keep adding guys

I'm pretty sure the audio problems are because there's a built-in splitter, since the splitter needs to output the same channels/bit/frequency, therefore 2 channel Vorbis with AC3 is not allowed. A swichter-less splitter (as found with all other splitters) is what we need. At least that's what I think.

[zulu]

Chapters:
The chapter seeking algo seems to be of by one.
Chapter keyframes have to be inserted one frame delayed in order that ogm dsf finds them.

[DKDIB]

My wishlist:
1- spilt Ogm/Ogm spilter, Vorbis encoder & Vorbis decoder in 3 filters (CoreVorbis is wonderful, so I want to use it also for Ogm/Ogg ;));
2- Unicode;
3- AAC;
4- RealVideo/Audio.

[Koepi]

The SubDS sources are there already as well? I thought Tobias wanted to release a new version first, 15times faster as the old version due to mmx optimizations :)

Before you create the todo-lists please take into consideration that the sources are from OggDS 0.9.9.3 - not the latest ones :-/ So fixes which are necessary for that version should make it into the lists as well.

and OT again: I have nothing against competition. In fact I like it. But abusing every possibility to advertise matroska ("you don't need ogm, use matroska, it can already do it.") - especially when competition arises in the special announcement/revival thread - is misplaced.
I didn't forbid anything btw., I don't know how that rumour came up. I wrote that this kind of behaviour is rude. Please rethink your positions concerning me forbidding discussions. I _asked_ for something. that's completely different.

Regards
Koepi

[robUx4]

Originally posted by Koepi
and OT again: I have nothing against competition. In fact I like it. But abusing every possibility to advertise matroska ("you don't need ogm, use matroska, it can already do it.") - especially when competition arises in the special announcement/revival thread - is misplaced.

Koepi, this matroska vs OGM has been going for a while. We first contacted Xiph a long time ago to join our efforts/knowledges but they simply wanted to push OGG so it didn't end well. What happens here is what happened in many places. The latest was on a Gentoo forum (http://forums.gentoo.org/viewtopic.php?t=63722) where you can see Monty saying that noone needs matroska when there is OGG, in a matroska related thread. I'm really tired of all this shit and I would like everyone to work instead of bashing the other.

Now if anybody want to compare things, you can reach me to know why I think in many cases matroska is the way to go. ;)

[Koepi]

I am not monty.
Monty even dislikes OGM (in the meantime it seems it is really just the extension that bugs him - the implementation seems to be correct.)

The OGM implementation has nothing to do with monty and/or xiph (ok, it's hosted there now. but noone of their crew will work on it, that's for sure.) I had my troubles with xiph as well (nice flamewar with emmett.)

So don't mix up things here please.

OGM is started by Tobias (-> going back where this thread belongs). It's coded by one from us. _He_ never mentioned anywhere we wouldn't need matroska. I didn't write that either (I only wrote that OGM suffices for my needs for now and I may have no use for it ;) ) OggMux will be renamed and support matroska as well. But you only keep bashing at OGM - which is very weak and I loose some of my interest in supporting matroska.

Could we please concentrate on the topic of this thread now and assemble a todo-list (and maybe find someone interested/with a little spare time in coding the OggDS filters further)?

Thanks
Koepi

[ChristianHJW]

Originally posted by Koepi
Could we please concentrate on the topic of this thread now and assemble a todo-list (and maybe find someone interested/with a little spare time in coding the OggDS filters further)? Thanks Koepi

Good idea. Lets start with SSA ... how are you going to handle overlapping timecodes in OGM ? Any ideas here ?

[bond]

i am voting for discussing ogm development in this thread only and create an own "ogm vs. matroska" thread to discuss the differences, advantages, disadvantages (this would be much clearer, fairer and would make our lives (and reading) much easier)

now back to topic:

missing changes between 0.9.9.3 and 0.9.9.5 (from tobias' hp):

- Changed everything to Unicode(tm) to ensure that other languages are handled correctly.
- Support for Vorbisgain tags
- The OggSplitter supports now the IAMMediaContent interface to pass some tags like "TITLE" a.s.o. to the player.
- The tray icon shows now also the title instead of "Ogg DSF"
- Fixed some problems with tray icon, which caused some players to crash when closeing
- Fixed vorbis.dll (the project on the Xiph CVS has an error, which results in vorbisenc to be compiled in vorbis.dll and to get an unusual big size)
- Support for 639-2 codes in the language tag. The corresponding LCID is then reported to the player. ( E.g. English[eng] )

[robUx4]

Well, let me express my deep apologies for what happened here. I don't like it as much as you, Koepi (who's currently the only active person on the OGM side).

I explained to all our team why we should calm down on such issues just when this thread started. Apparently it was not enough (especially for Pamel). So I tried to be a bit more convincing on our IRC channel. I hope this kind of things will be the past now.

Of course we both (OGM and Matroska ppl) get our heads hot quickly when something in our field is mentioned. And we would like our work to be appreciated and used as much as possible. But it should not be agressive or by trying to steal users on the back. But it should be only done when comparison is involved not in OGM or matroska specific threads... Let's make the competition creative, not destructive.

[bond]

Originally posted by DKDIB
My wishlist:
3- AAC;
4- RealVideo/Audio. i just tried muxing aac and rm into .ogm with tobias' multiplexer and aac seems to work without any problem :)
read more here (http://forum.doom9.org/showthread.php?s=&threadid=57327)

realmedia doesnt seem to work (as gabest already wrote)

[Koepi]

AAC should work if you write headerless aac-streams (which option got added to the latest aacmachine-releases). I wrote about that some time ago?

Bond, can you try using OggMux as well please and post any errors occuring here?

Regards
Koepi

[bond]

Originally posted by Koepi
AAC should work if you write headerless aac-streams (which option got added to the latest aacmachine-releases). I wrote about that some time ago?i used .mp4 files (splitted by 3ivx) from nero as input
i guess 3ivx doesnt pass the header to the ogm multiplexer...

Bond, can you try using OggMux as well please and post any errors occuring here?failed to render input aac! i tried aacs from nero, faac and aacenc (of course i never used any option to be headerless ;) )

[Koepi]

Thanks bond! :)

Ok, seems like it's necessary to track that down. If you use "render media file" in graphedit with your nero generated aac, which works in your example, which graph gets built? I'm sorry, I don't have the time to play around with aac and thus am dependant on your info :)

if the error message is really "failed to render input .aac" then there's a bug in the OggMux GUI, not sending the proper file name.

Regards
Koepi

[bond]

Originally posted by Koepi
Ok, seems like it's necessary to track that down. If you use "render media file" in graphedit with your nero generated aac, which works in your example, which graph gets built? I'm sorry, I don't have the time to play around with aac and thus am dependant on your info :)no problem :)
no graph can be built (i guess because there isnt something like an "aac splitter")!
when muxing in graphedit i used
input .mp4 -> 3ivx splitter -> ogg multiplexer -> output .ogm

if the error message is really "failed to render input .aac" then there's a bug in the OggMux GUI, not sending the proper file name.sorry my fault
the error message of course mentions the used file name

[el00343]

having just read the thread, all I can say is that this is great news indeed! So much for the "death of OGM".

[DAvenger]

Originally posted by el00343
having just read the thread, all I can say is that this is great news indeed! So much for the "death of OGM".

Let's hope we will see some progress on the above mentioned ogm issues :rolleyes: Competition is vital ;)

[Atamido]

Originally posted by ChristianHJW
Lets start with SSA ... how are you going to handle overlapping timecodes in OGM ? Any ideas here ? I don't think that this will be to hard. In Matroska we strip the timecodes/duration out of the subs and use them at the container level. In OGM it would probably be easy to just leave them in, or create a field specifically for duration. Then you could store an event at the appropriate timecode, and using that timecode, and the stored duration, you should be able to reconstruct the original timecodes. I think at that point it just becomes a matter of the support for those structures being added to VSFILTER.

I'm a bit rusty with the OGM structure, can anyone tell me if there would be a problem with that?

A bigger question in my mind would be how to use forward referencing frames in OGM. I think that currently MPEG-4 codecs are using packed bit-stream only, but there should be a way in OGM to store the frames individually. Any ideas for this?

@Koepi: Chill. I misunderstood Liisachan's comment.

[unmei]

oh i'm glad to see the storm in here seems a over ;D

i don't know much about the philosophy in OGM but i would understand if the concept was not that radical to attempt to break with old bad avi habits. I think at the moment it is not a big downpoint not to support native (err how you call it ? independent frame packing ?) as all the current codecs support the packed bitstream. This should not mean i would oppose someone implementing it, but i guess it's a rather big work someone has to find time for first.

There are sure a lot of small improvements easier coded that could give the end user more obvious instant improvment in "usage easeness" (the mentioned extended audio and subtitle format support for example)
à la "simple" things first...

[Atamido]

One other thing. OGM is currently using a "samplerate" that is set the the fps of the video. So, then the "granulepos" becomes the frame number. You would probably want to change this to use a default behaviour of using milliseconds as the samplerate, and then the timecode in ms for the granulepos.

This would allow vfr video as the default config. (and I'm not refering to dropping drupeframes that has recieved so much attention recently) I don't THINK this would break anything, but I'm not that familiar with how the DS filter works.

[Stux]

Packed Bitstream is not MPEG-4 compliant.

[Atamido]

Oops, I meant to say, "I think that currently MPEG-4 codecs are using packed bit-stream only in AVI". I would not be inclined to call packed bitstreams in AVI, MPEG-4 compliant. However, I have heard some compelling arguments for otherwise.

[DSPguru]

Cheers to Tobias !
i don't know about Xiph's plans, but it is vital that the following code will stay inside the OggDS sources :
if (vorbis_comment_query_count(&m_vc, "LWING_GAIN"))
{
m_bPostGain = true;
m_fPostGain = (float)atof(vorbis_comment_query(&m_vc, "LWING_GAIN", 0));
}