Kodi Community Forum

Full Version: Rendering UNICODE in XBMC/KODI Incorrectly
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2
there is an Unicode rendering bug in KODI.

I have tried to change Appearance->International->Language to Tamil (India)
I have seen no text in there. (because no unicode text available in Roboto-Bold.ttf and Roboto-Reguler.ttf)

then I have change Skin->Fonts to Arial Based. Texts are showing boxes (because no unicode text available in Arial Font)

then I have replaced default font to Latha font which is default unicode font in Windows PC.

Then I got this

Most of the fonts are rendering incorrectly. It has to be in red text which I typed in above picture.

I think there is a problem with text engine. It doesn't display correctly with Unicode characters.
For Tamil Language some texts are like this
கோர்ப்பு பட்டியல் -> ே+க+ா+ர+ஂ+ப+ஂ+பு ப+ட+ஂ++டி+ய+ல+ஂ

but it doesn't display correctly..
Please some genius help me..
Hi !

Thanks for reporting. Unfortunatelly it is a known problem and a limitation to the font we are using. I am sure there is a working font suitable for your language which you copy into kodi works ok. Try searching on forums or post a message on Transifex to get help from other users.

I would also try a windows arial font as well.
Thanks for the reply.
But I have tried many fonts. But latha.ttf is the best for Tamil Languages

I have searched alot but it gives me nothing

this is ticket for Myanmar language

Asian Languages its needs to be render it by some logical steps.
for example: 'skin' in tamil as 'தோல்'
but in KODI it says 'த ேல்'

'த' should come after ' ே'

 it also happening in Adobe Photoshop. I dont know why they dont fix that.
I think this is an error with TextEngine that used in KODI.
Do strings appear correct in the strings.po file ?

thanks for your reply and sorry for late reply

i think it is correct in strings.po file because when I look at that file in browser it seems correct.

but if I see it in TextWrangler app, it doesn't appear to be correct

When I see it in MAC default text editor it appear correcly

I think something wrong with text engine that rendering text.
In tamil language font places are different than latin/english font.

Try the fonts from Google early access program, start from the Noto Sans Tamil.

p.s. I don't read nor speak Tamil.
thanks dhead,
but that doesn't help me.. no changes between 1st image and this image
with Noto Sans Tamil

with Latha.ttf
I'm from Sri Lanka. I have same problem. kodi not rendering Sinhalese unicode correctly.

Please Help.
It is a bug. But they don't care about these bugs. No luck..
So the problem is that you can get it to draw with an unicode font, but that leads to the first letter switching places with the second?
I'm translating Kodi to Sinhala in transifex. but kodi not rendering unicode correctly. Please fix that kodi team.
Because, I Love KODI !

Thank you !
(2015-10-25, 21:18)Razze Wrote: [ -> ]So the problem is that you can get it to draw with an unicode font, but that leads to the first letter switching places with the second?

Unicode rendering solutions for Sinhala, Tamil and hindi.

for windows - Uniscriber
for linux - Pango, QT, ICU
for apple - Opentype specification / AAT
(2015-10-28, 06:15)geeth321 Wrote: [ -> ]
(2015-10-25, 21:18)Razze Wrote: [ -> ]So the problem is that you can get it to draw with an unicode font, but that leads to the first letter switching places with the second?


Is that a yes?

Is that a yes?

Yes. but, not always first letter. problem with Sinhalese (and tamil) vowel signs.


this is sinhala unicode character set

here how it works.

in english, sound "sa" has two letters. in sinhala sound "sa" has only one letter.
in english sound "s" has one letter. in sinhala sound "s" has two letters.

english ( s + a ) sa = sinhala ස
english s = sinhala ( ස + ් ) ස්

the problem is, sinhala language using some vowels before Consonants. in english vowels always use after Consonants.

english ( s + e ) se = sinhala ( ෙ + ස ) සෙ
but, for ease of sorting Unicode characters, the vowel ෙ (e) put after the Consonant.
but it should put before Consonant.
ස ෙ is wrong. සෙ is correct.
this correcting happening when rendering unicode.

actually vowel " ෙ" is after the Consonant. but unicode rendering and layout show vowel " ෙ" before Consonant.

Pango is a library for laying out and rendering of text.
GTK / GNOME uses this library to `shape' strings, i.e., to converts strings encoded in different languages into sequences of glyphs (shapes) from fonts.


(sorry for bad english)
Pages: 1 2