[NTG-context] lines tightly fitting together
luigi.scarso at gmail.com
Wed Mar 2 09:33:36 CET 2011
On Wed, Mar 2, 2011 at 9:22 AM, Hans van der Meer <H.vanderMeer at uva.nl> wrote:
> On 2 mrt 2011, at 07:05, Wolfgang Schuster wrote:
> Am 01.03.2011 um 21:59 schrieb Hans van der Meer:
> Replying directly to your address because the message body is too large.
> Pictures are pixel images of DVB subtitles from telvision broadcast, in case
> youre wondering.
> Why don’t you create a real font for this, this will create smaller files
> and you can use it with other programs.
> That is beside the point. The problem is not the drawing of letters, but the
> extraction of pixel data from a Transport Stream packets. The specification
> of the DVB-subtitles mentions two variants: (1) string of characters, (2)
> pixel colorvalues to be looked up in a 16-value CLUT. The stream I have at
> hand consists of pixel data in a 720x36 rectangle. Furthermore different
> broadcast stations use different lettertypes. Of course it would be nice if
> I could simply extract the letters, but for the moment this surpasses my
perhaps leptonica can help here .
More information about the ntg-context