New matroska sample file with track names

[ChristianHJW]

Finally !!

Hi,

i uploaded a new old sample, where many of you had been helping with translation a couple of months ago :

http://samples.matroska.org/matrix/index.html

CruNcher's original sample was redone with Xvid 1.0 beta2 and uploaded, as well as a RV9 version with chapters and track names, where Karl himself has done the original RV9 video track some weeks ago ;) ....

The samples contain

1 video track ( RV9 or XviD beta 1.0 )
1 audio track ( HE-AAC )
11 subtitles tracks ( arabian, catalan, dutch/flemish, english, finnish, french, german, japanese, portugese, slovenian, spanish )
RV9 sample only : 2 covers and 3 cover arts as attachements
track names, chapters, etc. ....

Here two ed2k links for the samples, if you want to help us spreading them :

Ed2k link RV9 sample (ed2k://|file|Matrix.Reloaded.Trailer-RV9-Karl640x432_HE_AAC_subtitled_chapters.mkv|21307025|542E0F6C8A2997D84B5464CA2DE81FFE|/)
ed2k link Xvid1b2 sample (ed2k://|file|Matrix.Reloaded.Trailer-640x346-XviD-1.0beta2-HE_AAC_subtitled.mkv|22251742|E548D30D19003A4EA1D7DA9E27FF92E3|/)

If anybody could help me with seeding some bittorrent links, i'd be graceful also and will add them here.

[Liisachan]

Subtitles are actually in SRT (UTF-8), but labeled wrongly as SSA.
As a result, I cannot read subs without setting VSFilter manually, as if it were not a modern container...
This is not how Unicode is supposed to work. ;_;

[ChristianHJW]

Hmmm ..... i took subs unaltered from CruNcher's sample :( .... how is that possible ?

Wouldnt mkvmerge realize the type of subs from the file its opening ? CruNcher, can you give us some insight how you actually made the samples ? BTW, all subs play fine for me, without having to set anything ? Is this because of my European Unicode code page ?

[bill_baroud]

@ChristianHJW : for the upload problem, it's a problem with Free.fr, not you. It's been 2 or 3 weeks i can't access my personal account, even though free is my isp :(

[ChristianHJW]

Originally posted by Liisachan Subtitles are actually in SRT (UTF-8), but labeled wrongly as SSA.As a result, I cannot read subs without setting VSFilter manually, as if it were not a modern container... This is not how Unicode is supposed to work. ;_;

Liisachan, thanks for pointing this out ! In fact, it seems that playback of the smaples is screwed by this, we are investigating what went wrong but apparently CruNcher's way to make the file, first load the subs in Graphedit with matroskamuxer and then tagging them with mkvmerge didnt ork out correctly.

We are redoing the subs now, and all with mkvmerge .... Thanks again !!

[Liisachan]

Another thing:
I'm afraid the Arabic subs are broken.
Maybe...converted into UTF-8 in a wrong way.


Funny thing is I can see Japanese subs easily on my Win XP English version, while I have to re-configure DirectVobsub to see Japanese subs on Windows 2000 Japanese version. These things just happen.
Perhaps this is a problem of VSFilter, or my OS is just screwed up

I guess the muxer used a wrong subsource filter on GraphEdit, but I cannot solve the above-mentioned problem by remuxing the file by myself. I know SSA would solve the problem perfectly but only on Windows (afaik SSA is not supported on Linux/Mac)

[ChristianHJW]

Liisachan, thanks for offering help here, but we hope to be able to remux everything in mkvmerge just fine and without any problems. The problem with this sample is simply that CruNcher made it several weeks ago, and at that time he way maybe fighting with a bug in current mkvmerge, thats why he was using Graphedit ( maybe because this was the only way to mux HE-AAC at that time, cant remember ).

He has all the source SRTs still, so this shouldnt be a problem i hope.

But another thing :

A vobsub track would fit perfectly in the sample, dont you think :D ? Is there any chinese in your team, so you could maybe make a chinese vobsub track for us, so we can mux this into the file also, maybe even compressed ?? I'd be very thankful for your help on this Lissachan :) ... while thinking about it, what is the most difficult to display european language, giving problems on most normal english Windows versions ? Finnish ? Danish ? Maybe worth another vobsub track ?

[Liisachan]

Originally posted by ChristianHJW
Liisachan, thanks for offering help here, but we hope to be able to remux everything in mkvmerge just fine and without any problems.
um, please don't misunderstand. I didn't say "I can solve the problem by remuxing for you" -- I said "I cannot solve the problem by remuxing the file by myself" which means, I had already remuxed it but it didn't help.

VSFilter tries to use the font settings in the DirectVobsub settings, even if the subs are in Unicode. This is paradoxical because there're no "settings for Unicode subs" in DirectVobsub settings box. If there was a setting for "font for Unicode" i'd use "Arial Unicode MS" for instance... As for the arabic subs, mkvextract tracks filename 3:arabic.srt gives me a UTF-8 text file, which is obviously not a proper arabic file. Could you please recheck Arabic sub data muxed?

A vobsub track would fit perfectly in the sample, dont you think :D ?

That is what Yusaku said in September: "VobSub subtitles could potentialy be used for storing generic image subs as well (under the limits imposed by DVD subs, but that is not that bad)"

I agree in a way. But just think about this: "We want to see Unincode glyphs, but it's difficult to use Unicode as text so let's use them as pics!"--This is an easy way, but too easy way, not a cool way. IMHO the way to go is USF, text-base. idx/sub for SSA is the 2nd best. But, yes, it might be better than SRT in a way, tho image subs are not flexibly reusable like text. Anyway, we'll need a powerful unicode sub renderer for SRT/SSA/USF, which can do automatic font matching, like Mozilla can display any unicode codepoint as long as you have at least one font file for the glyph in question.

sorry, I don't know any Chinese translators btw.

while thinking about it, what is the most difficult to display european language, giving problems on most normal english Windows versions ? Finnish ? Danish ?

Finnish and Danish are as easy as English, French or German to display for Windows, because they share the same "Code Page".

[ChristianHJW]

Originally posted by Liisachan um, please don't misunderstand. I didn't say "I can solve the problem by remuxing for you" -- I said "I cannot solve the problem by remuxing the file by myself" which means, I had already remuxed it but it didn't help. ... i got this right, but i was thanking you that you have tried to help us :)
VSFilter tries to use the font settings in the DirectVobsub settings, even if the subs are in Unicode. This is paradoxical because there're no "settings for Unicode subs" in DirectVobsub settings box. If there was a setting for "font for Unicode" i'd use "Arial Unicode MS" for instance...[quote] ... i hope gabest will read this, maybe he can come up with a solution ?
[quote]As for the arabic subs, mkvextract tracks filename 3:arabic.srt gives me a UTF-8 text file, which is obviously not a proper arabic file. Could you please recheck Arabic sub data muxed? ... lets hope the remuxing of the original arabian SRT sub will solve that problem. Other than that, i have no idea how to fix that, as all the originals are at CruNcher ...
That is what Yusaku said in September: "VobSub subtitles could potentialy be used for storing generic image subs as well (under the limits imposed by DVD subs, but that is not that bad)"
I agree in a way. But just think about this: "We want to see Unincode glyphs, but it's difficult to use Unicode as text so let's use them as pics!"--This is an easy way, but too easy way, not a cool way. .. this sample has only one function : to show what is possible with matroska ! I thought it would be a good thing to include a vobsub subs track also, for this very reason, maybe even 2, one compressed and one uncompressed. As i currently have no clue what program to use to make a vobsub subs file from a text sub file, i thought i ask the pro's :D ..
IMHO the way to go is USF, text-base. idx/sub for SSA is the 2nd best. But, yes, it might be better than SRT in a way, tho image subs are not flexibly reusable like text. Anyway, we'll need a powerful unicode sub renderer for SRT/SSA/USF, which can do automatic font matching, like Mozilla can display any unicode codepoint as long as you have at least one font file for the glyph in question. ... USF is finally on the way, i promise ;) ...
sorry, I don't know any Chinese translators btw. ... korean would be fine also :D !
Finnish and Danish are as easy as English, French or German to display for Windows, because they share the same "Code Page". ... its a shame i dont know more about this whole issue, but i am slowly making my way i hope. Thanks again for the advice Liisachan :) !

[Liisachan]

Originally posted by ChristianHJW
.. this sample has only one function : to show what is possible with matroska ! I thought it would be a good thing to include a vobsub subs track also, for this very reason, maybe even 2, one compressed and one uncompressed.

True :)

How to convert SRT or SSA to IDX&SUB:

(1) VirtualDub: use a noise-free, all-black dummy video with TextSub.vdf (for input) and SubLog.vdf (for output) [SSA in "letterbox mode" would be a good idea too]
(2) Edit the resulted .idx manually, if needed. (for instance, to change "Original frame size" and/or "Origin")
(3) the resulted idx/sub can be accepted by vobsub_vd.vdf but MkvMerge just doesn't like this .sub file.
A quick hack I've just found is using binary editor: open the .sub and replace all "00 00 01 BA 00" with "00 01 01 BA 00". Then mkvmerge will be fooled and accept the file. [Warning: If you use this trick to make mkvmerge happy, then vobsub won't accept the hacked file.]

Note: Technically, there should be a better way than that, because this conversion is, in essence, just printing out texts. But I just don't know...as a matter of fact, this is my first try to convert textsub to vobsub. I know I should try harder to configure the sublog filter to make the resulted .sub more noise-free, but I've no time for now.

resulted idx/sub (http://www.faireal.net/matroska/idxsub.zip)

[ChristianHJW]

Files were overworked and uploaded now, but French and Arabian subs seem to be broken still ... any help on that is well appreciated ;) ....

[outlyer]

Originally posted by ChristianHJW
but French and Arabian subsMmmh, I forgot a pair of accents on the spanish subs... they aren't all that important, so again, if you're gonna update the files, please update theese subs.
I've sent a pm to you, Chris, with the url.

[ChristianHJW]

bump !

Samples are finally up, and without bugs ( hopefully ) ! Thanks again to all the guys helping with translation of the subs, and especially CruNcher for timing them all, outlyer for making the catalan and correcting the spanish subs, as well as Toff for fixing the arabian charset and correcting the french bugs :) !

About bittorrent :

I am on a bussiness trip now for 2 days, so if somebody can help with seeding, just paste the links here .. thanks !

[ChristianHJW]

.... can somebody please help me with the bittorrent seeding for this sample, it seems ed2k is not really the right way to spread the files :( ...

[Blueseb]

I would be glad to help, but you should first set up a tracker (http://bitconjurer.org/BitTorrent/guide.html) ;)

[ChristianHJW]

Bittorrent links are up now :

Bittorrent Link RV9 sample (http://corecodec.homeunix.org:6969/torrent.html?info_hash=8a5e60aca1c3deb4e44a41bb275ea5e69c114565)
Bittorrent Link MPEG4 sample (http://corecodec.homeunix.org:6969/torrent.html?info_hash=6dd02f1caf94602b64a2ca36cac1736b268b394c)

[BetaBoy]

In case some tried and did not get them my Tracker was down. It's back up again... seeders need please. thanx.

----
Dan "BetaBoy" Marlin
CoreCodec Founder