[vlc-devel] Fw: zvbi - problem with czech/slovak subset #2

Marian Ďurkovič md at bts.sk
Tue Sep 9 17:40:04 CEST 2008

Copy of my email to Michael H. Schimek for vlc-devel

Hi again,

  I succeded to fix the unicode compose characters by doing slight corrections
in the compose table (src/lang.c). Here is the updated version:

static const unsigned short
composed[12 * 16] = {
  0x1041, 0x2041, 0x3041, 0x4041, 0x8041, 0xA041, 0x0000, 0xB043, 0x1045,
0x2045, 0x3045, 0x8045, 0x1049, 0x2049, 0x3049, 0x8049,
  0x0000, 0x404E, 0x104F, 0x204F, 0x304F, 0x404F, 0x804F, 0x0000, 0x0000,
0x1055, 0x2055, 0x3055, 0x8055, 0x2059, 0x0000, 0x0000,
  0x1061, 0x2061, 0x3061, 0x4061, 0x8061, 0xA061, 0x0000, 0xB063, 0x1065,
0x2065, 0x3065, 0x8065, 0x1069, 0x2069, 0x3069, 0x8069,
  0x0000, 0x406E, 0x106F, 0x206F, 0x306F, 0x406F, 0x806F, 0x0000, 0x0000,
0x1075, 0x2075, 0x3075, 0x8075, 0x2079, 0x0000, 0x8079,
  0x5041, 0x5061, 0x6041, 0x6061, 0xE041, 0xE061, 0x2043, 0x2063, 0x3043,
0x3063, 0x7043, 0x7063, 0xF043, 0xF063, 0xF044, 0xF064,
  0x0000, 0x0000, 0x5045, 0x5065, 0x6045, 0x6065, 0x7045, 0x7065, 0xE045,
0xE065, 0xF045, 0xF065, 0x3047, 0x3067, 0x6047, 0x6067,
  0x7047, 0x7067, 0xB047, 0xB067, 0x3048, 0x3068, 0x0000, 0x0000, 0x4049,
0x4069, 0x5049, 0x5069, 0x6049, 0x6069, 0xE049, 0xE069,
  0x7049, 0x0000, 0x0000, 0x0000, 0x304A, 0x306A, 0xB04B, 0xB06B, 0x0000,
0x204C, 0x206C, 0xB04C, 0xB06C, 0xF04C, 0xF06C, 0x0000,
  0x0000, 0x0000, 0x0000, 0x204E, 0x206E, 0xB04E, 0xB06E, 0xF04E, 0xF06E,
0x0000, 0x0000, 0x0000, 0x504F, 0x506F, 0x604F, 0x606F,
  0xD04F, 0xD06F, 0x0000, 0x0000, 0x2052, 0x2072, 0xB052, 0xB072, 0xF052,
0xF072, 0x2053, 0x2073, 0x3053, 0x3073, 0xB053, 0xB073,
  0xF053, 0xF073, 0xB054, 0xB074, 0xF054, 0xF074, 0x0000, 0x0000, 0x4055,
0x4075, 0x5055, 0x5075, 0x6055, 0x6075, 0xA055, 0xA075,
  0xD055, 0xD075, 0xE055, 0xE075, 0x3057, 0x3077, 0x3059, 0x3079, 0x8059,
0x205A, 0x207A, 0x705A, 0x707A, 0xF05A, 0xF07A, 0x0000

Regarding the problems with 13 characters in national subsets - this can be
workarounded by setting default region to 0. However, this should be transmitted
in X/28 and/or M/29 packets, so they are probably not detected for some reason.

We also noticed that some pages in CZ/SK teletext are missing (zvbi does not
find them), or in case of subtitles sometimes it skips a few subtitles
displaying only every 3-rd text or so.

Do you have any idea what can be wrong here?

    Thanks & kind regards,


----                                                                  ----
----   Marian Ďurkovič                       network  manager         ----
----                                                                  ----
----   Slovak Technical University           Tel: +421 2 571 041 81   ----
----   Computer Centre, Nám. Slobody 17      Fax: +421 2 524 94 351   ----
----   812 43 Bratislava, Slovak Republic    E-mail/sip: md at bts.sk    ----
----                                                                  ----

Previous message:

Hi Michael,

   recently we're trying to get fix teletext support for czech/slovak TV
stations in VLC media player. It looks like for some reason zvbi is presenting
Turkish characters instead of Czech/Slovak ones for  LATIN_G0, LATIN_G2,

   I've put a sample of MPEG TS from DVB-S channel here:


When you select teletext page 898 - the test picture, it should look like this:


However, it looks like this:


It seems zvbi confuses the special characters at rows 08 and 09 - instead
of czech/slovak ones, the turkish special characters appear on output from zvbi.

There's also a slight problem in rows 02 and 03 - the character "Ť" (U+0164 hex)
and "ĺ" (U+013A hex) are not decoded correctly and a white square is displayed
instead. (other white squares are part of that page).

Do you have any idea what can be wrong? We're testing with zvbi-0.2.30

     Thanks & kind regards,


More information about the vlc-devel mailing list