[vlc-devel] Re: SVG rendering module - advice wanted

[Olivier Aubert]

Hi Sigmund.

Thanks for the answer. Here are some comments.

> I still fail to see why you want to use the text renderer api. Do you expect
> printf() to render svg?
I would reverse your sentence : I expect SVG to be able to render simple
text, which my module does: it allows to specify a template, so that
text is rendered any way you want.

Anyway, the major reason for not defining another API is that the text
API is (and already was, at the time I wrote the plugin) there. It is
already a bit of work to add modules to the existing infrastructure, I
do not have the time to design a new one. In my case, VLC is only a
brick (albeit a major one) in an application, and I prefer to focus on
the application itself, plus its documentation, plus research
articles, ...  I will actively contribute to such an effort, but cannot
lead it.

> Vlc's YUVA format uses equal resolution for all planes. You seem to
> subsample the U and V planes.
Yes, that solved the problem (committed into SVN). I did not see the
annoucement for this change (as the same conversion code was working in
0.7.x), and could not grasp it from the video_chroma/ code.

> The subtitle timing properties are desided by whoever adds the subtitle to
> the subtitle core. This means that the text renderer module cares nothing
> about the subtitle timing. I don't know any part of the code that adds svg
> subpictures, so I can't help proofread this part of the code.
Well, I use VLC_showTextAbsolute/Relative to do this (interactively or
from an external application, through either the CORBA control plugin or
the python-vlc module that I will submit for inclusion one of these
days). You can download the Advene application
(http://liris.cnrs.fr/advene/) if you want to have a look at it.

Regards,
Olivier


-- 
This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
To unsubscribe, please read http://developers.videolan.org/lists.html