Matroska public alpha release date : 1st May 2003

[ChristianHJW]

Hi all,

i apologize for the very long posting out front, if you're not interested in matroska, well, dont read it ;).

The matroska project, with its predecessor/sister project MCF, was started about 2 years ago. Our aim was to build an open standard and open source A/V container project that was flexible and intelligent enough to last for the next 10 years, and to achieve this we did not just hack some code together quickly and release a first alpha into the wild, but instead we were making a lot of considerations about what such a container should be capable of doing, and we also tried to get advice from many different sides, like other opensource multimedia related projects, what they would consider as necessary and what options could convince them to support our container in their OSS projects.

After this long time, and believe me it wasnt always easy to take this route, the core dev team of matroska was happy and proud to be able to release a first alpha of libmatroska, the basic input/output library that can be used to read and write matroska files, beginning of this year. This library was greatly accepted from guys like Julien 'Cyrius/Suiryc' Coloos, the main developer of VirtualdubMod, and Moritz 'mosu' Bunkus, and they started to work instantaneously with it, building a great set of tools for both Windows and Linux for matroska file creation and also playback ( Linux ). As a result we are maybe capable of releasing a first set of tested tools soon, maybe even before the launch date named above.

So you may ask, what is it about the date that was named in the subject line, the 1st of May ?? Well, unfortunately not all sideprojects are coming along that well. We were maybe a bit naive to believe that every OSS developer out there would stop working on his project immediately, and start implementing matroska support into his tools. But, to be honest, the poor feedback we get from the OSS multimedia world out there is pretty demotivating for us, and we cant explain it at all. Its unacceptable to us that at least 20 - 30 developers in the OSS world are still investing their precious time into hacking strange things into AVI ( like SSA subtitles or DTS audio ), while a library exists allowing them to create matroska files with the very wanted content, but in a spec compliant and technically proof way.

Sure, we have some very good feedback also, like Ronald 'BBB' Bultje from the Gstreamer team announcing he might be able to look at making a Gstreamer plugin for matroska, or Mike Melanson offering to look at supporting at least playback from Xine, Francois 'mmu_man' Revol compiling the mplayer patches for OpenBeOS, etc. But besides that there seems to be not much interest from most other projects out there, and we believ to know why this is : When looking at OGM, the closed source container format based on Ogg ( made by Tobias Waldvogel ), its becoming obvious that the OSS developers only started supporting it after people were actually using it. Its interesting to see in this respect that for doing this there was no nice, clean library, instead the developers looking at it had to more or less reverse engineer the complete container, being major hassle for sure. But they did it, as there were users asking them to support it.

As a result of what is said above, we will change our strategy now. Instead of developing matroska to a very stable status on many platforms, and releasing it then to the public, we decided to release whatever we have on 1st of May. This measure is mainly undertaken to help us, the core dev team, to find the motivation again to work hard and even harder on getting matroska used and accepted, and hopefully that way we will also be capable of attracting the interest of other multimedia OSS developers, so they start supporting matroska in their apps, or maybe even contribute code to the main library.

There is one thing that may complicate releasing the matroska tools on 1st of May, and that is the missing DirectShow parser for Windows. Without this parser filter it will be impossible to release matroska to a broad audience. Our friend BlackSun had always predicted that this huge task is too much for only one person, and it seems our core developer Jan 'myFUN' Schlenker is currently extremely busy with real life stuff ( job ), so he cant contribute as much as necessary to develop the filter into a working alpha status until 1st May. We therefore created a task force yesterday to assist him in bringing the parser filter into a workable state until then. This task force will consist of 6 additional developers :

John 'spyder' Cannon
Ludovic 'BlackSun' Vialle ( DShow consultant, TCMP core dev )
Christophe 'Toff' Paris ( DShow consultant, TCMP core dev )
Steve 'robux4' LHomme
Jory 'Jcsston'
angelfire

All of them are looking at the M$ DShow SDKright now, trying to find out what is necessary to resolve the existing bugs in the current filter. Of course this means that they will stop their work on the current projects. Affected are :

spyder : MPV2MKV ; MPEG1/2 video to matroska
robux4 : libmatroska
jcsston : mpv2mkv GUI , mkvmerger GUI

I hope you all understand the necessity of this step and look forward to the 1st of May, together with us. Comments are welcome.

Regards

Christian

[zulu]

congratulations! looking forward to may 1st :D

[Selur]

Cool, thx for the info :)

Cu Selur

[NoLogo]

I wish you all the best.
I've been using OGM (and praid it) for about one year, but I recently decided to give it up, considering the lack of progress with the DShow filter (especially with subtitles and some other things).
And you couldn't imagine (or maybe you could :)) the number of people I know who are not ready to switch to OGM because of some issues (subtitles again).
So my advice is: go ahead, concentrate on DShow. Remember OGM: first, the only tool for a long time was OggCut (by Tobias), then OggMux came later.
One more idea: if Matroska can really handle with .ssa (and do it good), I guess you can have fansubbers behind you, and kick out AVI.

So, once again, all the best.

Regards
NoLogo

[stax76]

a DirectShow parser is one thing, VirtualDub the other, any information about that?

[BlackSun]

we have a fully working VirtualDubMod that support Matroska. Now just expect some basic features to be available, like playback and muxing of simple matroska file atm...

[Palikrovol]

Go on with it guys, hope may 1st were today :D

[HarryM]

Excelent.
I like OGM (mainly for native suporting of the best audio format- Ogg Vorbis), but competition in this category is essential. :cool:

What advantage have Matroska actually against OGM?

[ChristianHJW]

Originally posted by HarryM What advantage have Matroska actually against OGM? ... well, the current status of the parser filter makes it much much worse than OGM, believe me :D .....

[Acaila]

Originally posted by ChristianHJW
... well, the current status of the parser filter makes it much much worse than OGM, believe meI thought you were a salesman? Such a comment is not really conducive to a good business. Better lie and hope no one finds out :D.

Just kidding.

But anyways, great to hear the project is moving along and bearing fruits :).

[spyder]

As you can probably tell from Chris's post, the Matroska team has lost alot of motivation. The best thing this community can do is show suport for us. We do all this work for you also, not just ourselves. I personally won't let this project die until I am the only one left and I cannot proceed any further. :)

So please keep the good comments and feedback coming. And for any of you developers out there, this is an excellent chance to become part of our project. :)

WE WILL TAKE ALL THE HELP WE CAN GET! (As long as it's productive of course).

Spyder

[mikeson]

@ChristianHJW & all matroska devels:

It's GREAT! Please don't give up and keep working. I'm really looking forward to see (and use, of course ;) ) matroska.

Thank you and take care...

[BoNz1]

ChristianHJW, I think it would be difficult to not be disappointed about some developers lack of interest in your container format, and people's crazed obsession to hack all kinds of things into avi when they could help work on a much better format like matroska. I want to encourage you guys though, you have all put a lot of work into the project and I believe you are heading in the right direction. Releasing an alpha is a good idea I think too, it will spark a lot of interest which can only be good for you guys. I think then the support will come. :)

[CaptainCarrot]

Best decision you could have made, imho.
There is no real alternative besides ogm, which is not really developed further (at least at the moment) and was never actually meant to be what it already is. So just make it public, it'll spread since it's superior to anything else. I mean there isn't even anyone trying to compete with you. So just release it (ok, get that ds-parser working first, but then...).
Feels a little bit like christmas is on 1st may this year... :D
And for your motivation... how about
"Go, Matroska, go, get that freakin DirectShow" :D

Just ask for more if you feel demotivated (or did i scare you) :D

[vinouz]

Dont worry, guys, because we are a number eagerly waiting for this nice format to come. And be sure we'll be numbers to use it.

Because a good container is needed, and there you are :).

So, hope for the 1st of may. If there's VdubMod & Dshow filter working, I'm OK for it. I can wait for the rest. As we did for OGM (but too long, in my mind, now, without progress).

See all the post for DSPGuru's depart. That shows how many people read, think, use, and don't need to tell it, often simply not to disturb discussions with such notes. But the silent crowd is there.


Vince

[robUx4]

Thanks for all the kind words. This is really appreaciated !
We'll make our best for 01/05.

:sly:

[BlackSun]

Originally posted by spyder
As you can probably tell from Chris's post, the Matroska team has lost alot of motivation. The best thing this community can do is show suport for us. We do all this work for you also, not just ourselves. I personally won't let this project die until I am the only one left and I cannot proceed any further. :)

So please keep the good comments and feedback coming. And for any of you developers out there, this is an excellent chance to become part of our project. :)

WE WILL TAKE ALL THE HELP WE CAN GET! (As long as it's productive of course).

Spyder

Instead of posting you should be coding :D

[robUx4]

Originally posted by BlackSun
Instead of posting you should be coding :D

That's a coder who say that ???
:D

[jonny]

robux4 : libmatroska

I want to put my dirty hands on this! :)

ChristianHJW and all the developers: keep up! A lot of people is waiting for the first release!

[Atamido]

Originally posted by HarryM
What advantage have Matroska actually against OGM?
The ones that pop to my mind immediately are:

1. Variable frame rate video.
2. An actual set of specs.
3. Active developers.
4. Opensource tools.
5. Defined specs for every informational MP3 ID3 tag, and almost every other tag in other containers including APEv2, Ogg, RIFF, AVI, Extended RIFF, WMP9, CD-Text, and a few others.
6. More extensible for the future. (This is subjective as the future isn't here, but it was a top priority to ensure anything* could be stored in Matroksa for the next 10 years.)

Okay, I thought of 5 and 6 afterwards.


Note that Xiph may be working on OGM, but we have no indication of this. Also, Tobias's DS filters were supposed to become opensource, but who knows what happenned to that.

* 'anything' = timecode organized packetized data as well as various forms of pasta

[Blueseb]

alpha tester ready to go!:D
don't even think your work isn't appreciated by (potential) users
long life to matroska!!!!

[ChristianHJW]

Originally posted by Pamel The ones that pop to my mind immediately are:
1. Variable frame rate video.
2. An actual set of specs.
3. Active developers.
4. Opensource tools.
5. Defined specs for every informational MP3 ID3 tag, and almost every other tag in other containers including APEv2, Ogg, RIFF, AVI, Extended RIFF, WMP9, CD-Text, and a few others.
6. More extensible for the future. (This is subjective as the future isn't here, but it was a top priority to ensure anything* could be stored in Matroksa for the next 10 years.)

7. matroska can hold any number of audio streams, even if different audio codecs and/or sampling rates were used. Also the DShow filter will support this ( hopefully this year still )
8. matroska can deal with any VfW and/or ACM codecs out there, without having to know what codec it actually is ( see WMV9 VCM )
9. matroska has defined 'native' MPEG4 modes, for simple, simple avanced and advanced mode
10. matroska is supporting b-frames natively, without simply taking the VfW hackery with dummy frames and the like. In addition to that 5 more frame types are defined for future use
11. matroska specs do support h.264 with all the advanced features like improved streaming using NALUs ( video packet distribution with a priority ) already
12. matroska files can be stuffed with additional EDC/ECC(FEC) elements in a very flexible way, for enhanced mode2form2 or UDP streaming support

Just a few other advantages from the top of my head .... ;)

[Mgz]

Originally posted by HarryM
Excelent.
I like OGM (mainly for native suporting of the best audio format- Ogg Vorbis), but competition in this category is essential. :cool:

What advantage have Matroska actually against OGM?

I don't thinks Mastroka vs Ogg, becuase IIRC ChristianHJW said that Ogg is good for streaming but Matroska aim to storage :)

[ChristianHJW]

OGM <> OGG

OGM has the same target group of people as matroska, because all video streaming solutions from Xiph will be built on Ogg Theora, not on OGM for sure .... so a comparison between the 2 is valid.

[robUx4]

OGG is designed with streaming in mind, unlike matroska.
For TCP based streaming we are equivalent (I'm being diplomatic here ;)). For UDP based (RTP) OGG is better... Until we have specified our RTP payload format (that we already have roughly in mind).

[BlackSun]

Originally posted by Mgz
I don't thinks Mastroka vs Ogg, becuase IIRC ChristianHJW said that Ogg is good for streaming but Matroska aim to storage :)

Storage, features and editing (and love :D)

[haibane]

That's the greates news I've heard today........
so the long wait is finally over.....
Rock On, Matroska...........

[alexnoe]

and people's crazed obsession to hack all kinds of things into avi Yesterday, I've finished adding SSA support to AVI-Mux GUI :D

[crOOk]

I can't wait for May 1st! I hope you get the DS done by then... I'm glad to hear all the specs of your container. It really sounds awesome!

[BoNz1]

Originally posted by alexnoe
Yesterday, I've finished adding SSA support to AVI-Mux GUI :D

You are a bad bad boy. :)

[spyder]

Our DirectShow task force is spending every free minute debugging this filter. We hope to have a very stable filter by the time we release. :)

[Siku]

I read those advantages what Matroska have over AVI/OGM and I'm really impressed! Can't wait the 1st May :D

[masken]

ChristianHJW, I think it's definetly the right way to go. Get it out so that it gets attention. Unfortunately, most people don't have any kind of patience with "paper products" (meaning no binaries). Unheartly to the developers it may seem, but believe me, when there's some binaries out there, even in early alpha-stages, you'll get the whole community behind you in development.

Good luck, this is a *very* interesting project, and the work that has been done, together with the specs, indeed truthfully aims at the goal it was made for. And a focus oon a working DirectShow filter is of course one of the most essential things :)

[spyder]

Progress Update:

We (The Matroska DirectShow Taskforce) have debugged the filter a bit and gotten demuxing working properly with a few small problems concerning timestamps which cause a slight stuttering in the playback. Our current working code is not in the CVS yet. We hope to soon have the timestamp issues fixed and look into removing some more annoying bugs in the filter that cause crashes. Only then can we move on to seeking. We have developers working on subtitle muxing, tagging(viewing and editing at post mux stages), the DSF, libmatroska, and MatroskaDub.

Matroska will have a very powerful tagging system, initially including all tags defined by ID3(v1.x and 2.x), APE, Vorbis, WMP9, CD-Text, RIFF info, etc.. More types will be able to be added seamlessly later on. Tags will be able to be used for single tracks or per chapter, depending on the author's desire. Also, I think Pamel is pushing for multi-language tagging to be added to support storing tags in different languages for the same file/track/chapter. It's flexible enough to hold your whole MP3 collection in a single file. Pamel is working on a list of tag elements to be defined in the Matroska spec. There will be more than 100 existing tags supported, but the total variations of tags will be several hundred. If you know of a tagging system not mentioned in this post which you think should be supported please post here.

The DSF taskforce has been steadily working on debugging and repairing the filter to work properly. Yesterday we had playback of video+audio files (Xvid+MP3) with only a small stuttering due to incorrect timestamps we believe. We will be adding block buffering which should make variable framerate playback a reality and also clear up some calculation issues. With block buffering we can use timestamps to calculate the duration of a frame based on the presentation time of the next block.

In the coming days we will be working hard to deliver the best product possible. Our aim is to provide an all-purpose container that will suit any user's needs. Whether it be storing an entire collection of videos in a single file or simply combining video and audio in a synchronised container. We will not yet support a few things that will make Matroska stand out from other containers. We do have limited coding resources. For example, menuing will not be supported at the initial release. The initial release will most likely be no more than simply making a combined stream of audio and video and being able to play it back. We look to the community to join in and help with the next phases. After all, this is our container( yeah, including you ;) ).

John Cannon
Matroska Developer

[mikeson]

@all Matroska devels:

It's wonderful to hear that. I can't wait till 1st May. Please don't lose enthusiasm for working. I (and sure not alone) believe in you and Matroska. :)

Best wishes to you.

Date edited.

[CaptainCarrot]

Hi,
will the bit error correction be supported in the 1st may version, or is that one of the "to be added"-features?

[JimiK]

Why is everybody waiting for March 1st? That was a long time ago, or do you mean March 1st 2004? ;) Nevermind, it's just that I saw this date in another thread, strange coincidence.
After I'm also part of the community, I just wanted to say that I'm also anxiously waiting for the initial release. Just a question about Matroska and players: will it be easy to support it in players using shortcuts like Ctrl-A to switch audio. I think this would be an important factor, because the user wants to have it as comfortable as possible. Of course I know that you're programming a container and not a player. Still I want to now: how can we access different streams? Will it work like the OggSplitter menue so that "Language" or another entry shows up where you can select your stream?
Thank you for pushing that hard towards the initial release,
JimiK

[spyder]

Error correction will not likely be supported in the initial release. The only error protection will probably be CRC32. The ECC will have to be added later.

The release date is May 1st, not March 1st. ;)

We will likely support the same interfaces in the filter as the OggDS filter for player configuration.

[Atamido]

Originally posted by JimiK
Just a question about Matroska and players: will it be easy to support it in players using shortcuts.... Of course I know that you're programming a container and not a player. Still I want to now: how can we access different streams? Will it work like the OggSplitter menue so that "Language" or another entry shows up where you can select your stream?
You're right, this is mostly a player feature, however it will be taken into consideration how to have the player interact with the filter to switch streams.

The interface for switching the stream directly through the filter has not yet been decided on yet, but it will likely take use an interface similar to one of those filters.

[int 21h]

No offense Chris, I wish you guys the best with this project, however, the negative response from other developers stems from the amount of hype this project and the project it originally sprang from have always suffered. All of those developers are busy working on fixing their own bugs and adding their own new features, and more than likely if they are typical OSS developers, barely have time for that, let alone, implement the support for an untested, unusable, but technically superior file format.

However, that being said, I believe you guys have finally realized this yourself. Releasing a working implementation will speed integration of your format into other tools.

So to all of you, keep working, your diligence should eventually pay off. Rome wasn't built in a day, and certainly no one moved in as soon as the blueprints were done.

[ChristianHJW]

Originally posted by int 21h No offense Chris, I wish you guys the best with this project, however, the negative response from other developers stems from the amount of hype this project and the project it originally sprang from have always suffered. .. well, like Nic says, its ' .. a helluva project ...' :D :D !!
All of those developers are busy working on fixing their own bugs and adding their own new features, and more than likely if they are typical OSS developers, barely have time for that, let alone, implement the support for an untested, unusable, but technically superior file format I am a motivation driven, over-enthusiastic person and it took me some time to realize that ....
However, that being said, I believe you guys have finally realized this yourself. true
Releasing a working implementation will speed integration of your format into other tools. So to all of you, keep working, your diligence should eventually pay off. Rome wasn't built in a day, and certainly no one moved in as soon as the blueprints were done. .. Suiryc got XviD and Vorbis playback from matroska working yesterday, without stuttering ... but i cant say yet how he did it, its a surprise ;) .... if the filter will come along for the next 14 days as its doing now, then we should have something working pretty cool on May 1st, being fully opensource, open to anybody and we really do hope to be able to raise the interest of other developers by then. Mosu even got allowance that his mplayer playback patch based on the C++ lib will find its way into official mplayer CVS, at least as an interims solution until there is a C lib .... i wished our DShow filter had the same level already as this mplayer patch has, Mosu is really a wizzard :) ... his program 'mkvmerger' was ported to win32 already and can be the basis for a couple of command line matroska muxers on almost all platforms ... so how could i ask for more ....

[unmei]

how the hell could i overlook this thread till now ???
The first time i actually dived into this forum i saw "matroska" and loved the idea. But i learned that even the cracks need some time to get great things working, and it seems like stupid users like me finally can get a glance of it :)

I really don't know whether i will be able to find any sleep until 1st may - i love you guys and will gladly take and rape any snippet of muxer, demuxer, dshow filter etc you release (well for windoze that is...) :D

*big endorphine shock here*

[vinkes]

@ChristianHJW,

I have a short question:
will I be able to mux ogg vorbis (5.1) , subtitles and realvideo 9 movie in to a matroska container? Is realvideo 9 supported?

Regards

[Atamido]

Originally posted by vinkes
will I be able to mux ogg vorbis (5.1) , subtitles and realvideo 9 movie in to a matroska container? Is realvideo 9 supported?
1. It will eventualy, but probably not at release. Edit: The Matroska team (Toff) figured out how to connect the Matroska filter to OggDS, so now Vorbis is playing back. If anyone *cough*Xiph*cough* would make make a general Vorbis DS filter, I'm sure everyone in the community would be eternaly grateful. Regardless, 2ch Vorbis will be working at launch, but getting 5.1 Vorbis up will probably just take longer.

2. Different subtitle formats are well on their way to being placed in MKV.

3. RV9 could be supported if Real openned the API so you could play RV9 from anywhere. I mentioned this to karl_lillevold, and he said that he is already aware and actively trying to get Real to open things up. If they do, then you can, if not, then you can't.

[Atamido]

Two weeks and counting.....

[ChristianHJW]

Guys, the parser filter is working :D :D !!!

Toff, Blacksun and Suiryc made it !!!!! Even Vorbis audio, and no buffer underruns anymore ;) .... these guys are just great :) ...

[BoNz1]

Sweet!!! I got a whole bunch of stuff on my hard drive: avis and vorbis waiting to be muxed into matroska :)

[marnum]

Yeah, thanks for your great work! The quick success of implementing all these features can only be a sign of Matroska's high compatibility!

[stax76]

I'm a little bit confused right now, there is VirtualDub, VirtualDubMod and MatroskaDub. I wonder which plans the developers of each of the 3 different versions have

[spyder]

MatroskaDub is only a test version for development of Matroska. Once we release it will become a part of VirtualDubMod.