required subtitles/captions features

[ileanadumitrescu95]

SourceForge feature request #23 written by Lukas Huber on 2003-08-27

Hiya Folks,

we wanted to extend thank you for setting up such a
nice application. :)

The consortium "Barrier Free Media" made an informal
site about displaying and recording subtitles/captions.
For your information please check out:
http://www.deaftv.de/tvtuner_en.htm
http://www.deaftv.de/tvtuner_test1.htm (german)

We tested Zapping 0.6.7 so we can see which
"deaf-friendly" features should be implemented.

• record captions either: 1.) directly onto the movie
  images, or 2.) into a separate subtitle file, using
  time codes that link the captions to specific
  movie-frames in an established data format.
• display captions at single and double text height.
• correctly select a teletext page that carries
  captions, since different broadcasters transmit their
  caption codes on different pages, especially if they
  provide them in more than one language.
• create a table of broadcasters with teletext page
  numbers that carry subtitle codes.
• copy not only movie-frames but also the associated
  subtitles into the clipboard using screen capture software.

Keep up the great work!!

best regards,

• Lukas Huber
• Bernd Schneider
• Fritz Mainzer
• Nikolaus Riemer
• Bernd Rehling
  Members of the consortium "Barrier Free Media"
  http://www.deaftv.de/tvtuner.htm

Comments by Michael H. Schimek:

Logged In: YES
user_id=46861

Thanks for your feedback.

> - record captions either: 1.) directly onto the movie
> images, or 2.) into a separate subtitle file, using
> time codes that link the captions to specific
> movie-frames in an established data format.

I planned to add MPEG-2 interlaced recording with subtitles in
a vbi stream, the format used for digital TV transmissions. It's
the most efficient and easiest to implement, but I don't know if
there are currently any MPEG players with Teletext decoder.
The DVD standard only permits a Closed Caption stream and
subtitle bitmaps. Don't know about other formats. Open caption
should only be the last resort.

Recording subtitles in clear text format was planned to create
transcripts. What is an established data format? Plain ASCII or
UTF-8 might do, but to indicate who speaks you probably want
to preserve text colors, styles and layout.

> - display captions at single and double text height.

The display leaves much to be desired. Currently CC and TTX
are displayed as on most TVs. Color or character size
changes would be simple. Moving the text relative to video or
a zoom function are more challenging. On the far end of the
scale one could replace the background bars by character
outlines and permit custom fonts with proportional spacing.

> - correctly select a teletext page that carries
> captions, since different broadcasters transmit their
> caption codes on different pages, especially if they
> provide them in more than one language.

I always felt the UI to select subtitles is too clumsy. The user
should enable the Teletext mode, select a page and enable
overlay. For CC the subtitle menu will do. Entering the subtitle
page number in preferences was a stupid idea. Zapping
should have no problem remembering the last page selected
for overlay on each channel, and select that page when the
user actives subtitles in the View menu.

Speaking of different languages, if you open the context menu
you may see a subtitle menu. This is possible because some
stations transmit a machine readable page directory. CC can
transmit language codes, for TTX the decoder tries to guess
from the character set used on the page. You can see this for
example on German station arte.

So actually the user may not need to enter a subtitle page
number even once. An overlay button, menu entry or OSD item
should also indicate if the current program has subtitles. This
information is not directly available from TTX, but one could
monitor known subtitle pages for changes or query an EPG
database.

> - create a table of broadcasters with teletext page
> numbers that carry subtitle codes.

Given the information above one could mark subtitling stations
in the channel editor. The dialog is already quite complex and
will get worse, but if you say it's important. I don't think page
numbers make much sense there. If remembering and subtitle
menu are not enough, bookmarks would be my next choice.
The Teletext browser supports bookmarks, which can point to
subtitle pages, but they are not offered in video mode.

> - copy not only movie-frames but also the associated
> subtitles into the clipboard using screen capture software.

The screenshot plugin can already save Teletext pages, an
option to save images with subtitles is on my wish list, just for
completeness.

Michael

Comments by Lukas Huber:

Logged In: YES
user_id=853683

> > - record captions either: 1.) directly onto the movie
> > images, or 2.) into a separate subtitle file, using
> > time codes that link the captions to specific
> > movie-frames in an established data format.

> [...] but I don't know if there are currently any MPEG
players
> with Teletext decoder. [...]

> Recording subtitles in clear text format was planned to
create
> transcripts. What is an established data format? Plain
ASCII or
> UTF-8 might do, but to indicate who speaks you probably
> want to preserve text colors, styles and layout.

Regarding 2.) there are good tutorials about miscellaneous
data formats which contain subtitle informations as described:
http://ncam.wgbh.org/cdrom/guideline/guideline2.html
http://ncam.wgbh.org/richmedia/index.php
I think that e.g. SMIL and SAMI are one of most interesting
standard data formats for showing subtitles on video clips.

Logged In: YES
user_id=853683

Hi Michael,

Recording subtitles in clear text format was planned to
create
transcripts. What is an established data format? Plain
ASCII or
UTF-8 might do, but to indicate who speaks you probably want
to preserve text colors, styles and layout.

FYI, there are different subtitle file formats. Maybe you
can look
at section "Current list of supported formats:"
http://www.urusoft.net/products.php?cat=sw
I think at first you can pick up some most used subtitle
formats
such as SAMI Captioning (.smi), SubViewer 2.0 (.sub),
SubRip (*.srt).
I use a free player www.bsplayer.org running on Windows which
can display subtitle files with movies. I will test if any
of Linux
video player can display subtitle with those formats.

Demo screenshots:
http://members.kabsi.at/l.huber/images/cor1978/cor1978b.jpg
http://members.kabsi.at/l.huber/images/cor1978/cor1978g.jpg

For further questions and feedback please send me mail to
lukbo at a1 dot net

~Lukas

Comments by Michael H. Schimek:

Logged In: YES
user_id=46861

Hi Lukas,

Here are examples recorded from Teletext with libzvbi 0.3
(should work with Closed Caption and Movietext as
well):

http://zapping.sf.net/screenshots/caption.smi (SAMI)
http://zapping.sf.net/screenshots/caption.sub (MPSub)
http://zapping.sf.net/screenshots/caption.txt (QText)
http://zapping.sf.net/screenshots/caption.rt (RealText)

This is the SAMI file translated into proper HTML. Of course
players don't have to render the black boxes, it's just the
default background color of paragraphs.

http://zapping.sf.net/screenshots/caption.html

Is that what you expected?

Michael

Comments by Lukas Huber:

Logged In: YES
user_id=853683

Hi Michael,

yes, of course I do expect such recorded subtitle formats as
one addition feature displaying subtitles in player
applications! :)

Do you also think about implementing an additional feature
recording subtitle text directly onto the movie images? It
would be fine also. The user should be able to choose
between both features. But implementing subtitle text
feature seems to be easier.

Please also consider the subtitle format *.srt (SubRip),
it's one of most used formats. Your subtitle demo files
seems for me to be ok, but caption.sub (MPSub) looks unusual
for me. I can't play a movie with your *.sub file. Maybe
could you try to record a *.sub (SubViewer 2.0) file ?

Here are demo subtitle files made with SubtitleWorkshop:

http://members.kabsi.at/l.huber/videos/cor1978.mpeg (video film)
http://members.kabsi.at/l.huber/videos/cor1978.sub
(SubViewer 2.0)
http://members.kabsi.at/l.huber/videos/cor1978b.rt (Realtext)
http://members.kabsi.at/l.huber/videos/cor1978.smi (SAMI)
http://members.kabsi.at/l.huber/videos/cor1978b.srt (SubRip)
http://members.kabsi.at/l.huber/videos/cor1978b.txt (QuickTime)

It's also ok with transparent only subtitle text without
rendering black boxes. This depends on player applications
which display subtitle text from files.