[advocacy] Re: Subtitles and Code Pages (Text Encoding)

[ChristianHJW]

Paul Bryson wrote:
> "Arc Riley" wrote...
>>First of all, there is *NO* Xiph codec available at this time which
>>provides subtitles.  This is the first indication that there's some
>>serious FUD comming from the Matroski PR spin doctors.  If you
>>investigate some of the other claims they've made about Ogg you'll find
>>they've been running this long-term campain to spread misinformation
>>about Ogg's capabilities and expandability.
>
> Yeah, he popped into #matroska and asked about it.  I let him know that OGM and
> Ogg are not the same thing.  A lot of people confuse the two.  Unfortunately the
> Matroska PR Spin Doctor is enroute from China to Germany, so he was unable to
> spread more FUD at the moment.  If you like, I could point him towards this
> discussion?

Not true. Writing this from a hotel in Beijing, i had to watch the F1
race in Bahrain with chinese commentary ... lol :). But soon i'll be on
the plane home, hopefully, even if it was a very successful trip ;) ...
About OGM and Ogg, well it was Xiph insisting on naming the OGM files as
.ogg at first, and many tool makers for OGM files followed their advice
and made .ogg a possible extension for them. Its only logical that
people are confused a lot now, as OGM is definitely a wide spread format
already.

>>There are two Xiph subtitle formats being worked on, Writ and MNG (Ming).
>>Neither have been seen outside a few examples, neither is a "standard"
>>as neither are in a finished state.
>
> None of us have heard of an MNG format for Ogg.  Do you have any links to
> discussion or papers about it?  There is already a PNG video codec and I wonder
> if it uses any of the same principles.

We discussed both PNG and MNG as allowed subs for matroska, but first
tests showed that the necessary decoding power requires a very strong
CPU to allow real time decoding during playback of a compressed movie.

>>Writ, as documented at http://wiki.xiph.org/OggWrit , uses Unicode
>>exclusivly.  It doesn't use seperate codepages, much less the "current
>>codepage of the software", as we instead chose a design which allows
>>every language to be properly represented.
>
> Thats good.  Generally speaking, code pages are just retarded.  UTF-8 was a good
> choice.

There are 5 existing subtitle editors for USF already, another XML based
subtitles standard, some even opensource. You might want to have a look
at them, maybe they can easily be modified to allow Ogg Writ output ?

>>MNG, as you are probobally aware, is just a lossless animation format.
>>It will allow all sorts of weird animated/colorised subtitles to be
>>supported.  Since subtitles in this codec are done in graphics only,
>>well, 'codepages' doesn't apply here.
>
> Again, I would be interested if you could point me to any links here for MNG in
> Ogg.

I also have been interested in this, but only until i learned more about
vobsubs, the subtitles standard used in DVDs. I doubt you can do
anything better than that based on PNG or even MNG, and really achieve
significantly smaller file sizes, gain similar widespread support, and
without consuming at least 100x more necessary CPU during realtime
playback. Not everything must be reinvented, even the MPEG group has
come up with good things already in the past ;) ....

>>There are a few third party subtitle formats being implemented in the
>>wild for Ogg, none are anywhere near a standard, and none are
>>representitive of the capabilities of Ogg.  Ogg is a container format.
>>What some third party (ie, non-Xiph) subtitle format supports or does
>>not support is not an issue with Ogg, but rather, an issue with that
>>codec alone.
>
> I don't know of any supported subtitle codecs in Ogg at the moment.  The closest
> thing would be OGM, but that is definately not the same thing.

... although some people at Xiph found it quite convenient, these days,
to make the existing OGM support group keep believing that one day OGM
could become the official Ogg specs for 3rd party codec support ;) ....

>>Where the Matroska guys got it all wrong is trying to combine the
>>subtitle format and scripting language with the container... this, IMHO,
>>is *WRONG* *WRONG* *WRONG* (beating lame matroska script kiddies over
>>the head with a rolled up newspaper).  These are seperate functions that
>>should be handled in their own layers; a container to hold everything
>>together and codecs to provide media of various types and formats.
>
> *WRONG* *WRONG* *WRONG* Actually, I don't know.  In fact, noone really knows
> what in the world you are refering to.  There is no scripting language in
> Matroska, though it has be theorized on a bit in the past.

... definitely the worng place to discuss about, but i still feel
matroska *SHOULD* have an 'official' subtitles format, and my vote goes
for USF. Yes, i even like the idea to convert USF subs to EBML when
muxing them, but what do i know, i am just a 'PR spin doktor' as we all
know ;) ...

>>Anyone can implement any subtitle format with Ogg because the subtitle
>>format is treated as an independent codec, as it should be.  This allows
>>for the flexibility to add new subtitle codecs in the future when new
>>features are needed, etc.
>
> This is only partially true.  The problem is that Xiph still hasn't defined how
> to handle third party codecs.  Until that time, nothing is valid.

Is it really necessary for Xiph to define a way how to handle 'non-Xiph'
content ? I am not convinced. The Ogg format, from what i understand, is
technically not really a good container for all kind of different video
and audio stuff. Its also mainly 'competing' with MPEG, RM, WMV and MP4
as a free and powerful video/audio content distribution format, and as
such their first aim should be compatibility and playability. So why on
earth should Xiph aim to, or even enforce to allow a lot of stuff to be
put into Ogg at all ? MP4, in principal, could also serve as a 'general
use' container ( after all, its based on MOV ), but do you think the
MPEG-LA would want to enforce that, risking to confuse people if their
brand new MP4 unit doesnt play the nice MP4 movie they just got, while
on their PC it's playing fine ?

>>I'll note, too, that Writ was designed specifically to allow future
>>expansion without breaking compatability with older software/etc.  You
>>should really take a look at it if you're interested in the nuts and
>>bolts of subtitle formats.
>
> You may also consider SSA and its bigger brother ASS.  Both are fairly mature
> subtitle systems that allow for extremely complex effects and have a reasonable
> following.  I'm not sure what you mean by compatibility as there isn't any
> current OggWrit system out there to be compatible with.  Though, I am pretty
> sure that even mplayer will display plain text when it detects an SSA/ASS
> stream. Pamel

Couldnt have said it better. We designed USF, for this purpose. Reality
bites, nobody is really interested in using it, because SSA/ASS can do
the same and even more, and have wide spread support and acceptance
already .....

Christian
matroska project admin

--- >8 ----
List archives:  http://www.xiph.org/archives/
To unsubscribe from this list, send a message to 'advocacy-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.