FLOSS Multimedia Support in Fedora

[Martin Sourada]

On Mon, 2009-02-09 at 13:54 -0500, Gregory Maxwell wrote:
> 2009/2/9 Martin Sourada <martin.sourada at gmail.com>:
> > Well, that is an interesting piece of software, but it seems to support
> > either image subtitles (big, binary, suboptimal for most modern use
> > cases) or purely text subtitles with none other effects than karaoke...
> > Well, at least some positioning and basic styling (font, color, borders
> > shadows) would make it an interesting alternative. Also there is AFAIK a
> > lack to attach fonts to the stream to make sure the subs will look the
> >same everywhere.
> [snip]
> > So basically we're left to amateur videos and screen-casts. In
> > screencast at least subtitle positioning is vital to be able to use the
> > subs effectively (you would not want to overlay your subtitles over some
> > important text in terminal).
> 
> Pretty narrow view of the world! How about educational works,
> government materials, porno, etc?  :) There are many things which are
> no anime, pirated movies, amateur video, and screencasts. :)
> 
Ok, ok, there is more than anime, movies and screencasts and should we
win the porno industry, it would be a huge, though controversial, win
for FLOSS multimedia but I suspect that is as likely as Sony starting
releasing on BluRay using Theora+FLAC+ASS in Matroska (they'd probably
need matroska for it's DVD-menu like capabilities)...

> In any case— you've missed what most of Kate can do. It is, in fact, a
> bit overkill for most subtitling needs.
> 
> It can do positions, color, transparency, and even animate them with splines.
> 
> In fact, Kate's excessive flexibility in this regard is a problem
> since it hinders adoption.
> 
> The font issue is somewhat hard: Fonts often have licensing problems,
> they complicate seeking and dynamic access (big enough that you don't
> want to constantly resend).
> 
I think the decoders usually do it but extracting the font and adding it
to cache. I'm not sure about streamed media, there the few MiB of fonts
might matter a lot, but when you target quality (not size) it's usually
less than 1 % of the resultant file so there is practically no
difference... 

Yep, licensing issues are problems, but openly licensed fonts aren't
usually not easy to find - it's much more convenient to embed say purisa
into the file than requiring your users to either find it themselves and
install it or being satisfied with a fallback font (which might not
satisfy the artistic concerns that led you to the choice of the original
font).

Nicolas's idea about CSS like font selection is interesting and might be
worth pursuing for streamed files, but for classic usage (movie,
document, video lecture, ...) case it's probably better to attach the
fonts to the file itself or be satisfied with plain white or yellow
subs ;-)

> Kate is far from great, but it should meet most basic and many
> advanced subtitling needs.
> 
I stand corrected. 

> I used quotes only because you discovered that it actually wasn't
> lossless for your input. :)
> 
> You'd probably have better results if you could force it into 4:4:4
> mode, but I'd put money on the resulting files working in just about
> nothing. :(
> 
I was rather interested about the real cause of the blurriness effect
than by a way how to workaround it ;-)

Martin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
Url : http://lists.fedoraproject.org/pipermail/devel/attachments/20090209/b1d26601/attachment.bin