Re: [ft-devel] freetype rendering problems

[suzuki toshiya]

Dear Mauro,

Could you tell me how to reproduce the reported issue?
I built test.cpp, and trying to print the rasterized
bitmap & dump other parameters, but I could not find
which character and which size should be specified.

In my case, "r" at size 11-13 is similar, but different.

$ ./mauro-test.exe r 13 /usr/share/fonts/truetype/ttf-dejavu/DejaVuSans.ttf

O.OO
OO..
O...
O...
O...
O...
O...

bitmapGlyph->metrics:
bitmapGlyph->metrics.height 7
bitmapGlyph->metrics.horiAdvance 5
bitmapGlyph->metrics.horiBearingX 1
bitmapGlyph->metrics.horiBearingY 7
bitmapGlyph->metrics.vertAdvance 13
bitmapGlyph->metrics.vertBearingX -2
bitmapGlyph->metrics.vertBearingY 3
bitmapGlyph->metrics.width 4

Regards,
mpsuzuki

Mauro Conzato wrote:
> Hello,
> 
>  
> 
> We have a strange rendering issue with freetype library that is driving
> us to ask an opinion from you.
> 
> Our application is using two free true type fonts: DejaVuSans.ttf and
> DejaVuSans-Bold.ttf, in monochrome bitmap way, with horizontal
> alignment; the issue basically consists in the following: the 'r'
> character is being displayed in a higher position, on the baseline.  
> 
>  
> 
>  
> 
>  
> 
> Two different versions of freetype library are producing a different
> result:
> 
> -  a relative recent version of ft (2.4.6) is showing the problem
> 
> -  an older version of ft (2.3.9) is properly displaying the character
> 
>  
> 
>  
> 
> I have made some debug and discovered the following:
> 
> - If I reduce the horizontal bearing Y parameter, the char is properly
> aligned on the baseline
> 
> - the metrics are showing me that there is not difference on the values
> that are managing the bitmap glyph, so it seems that the horizontal
> bearing Y recovered from the font file, through the library are the
> same. So,  since the code base that is managing the application is the
> same, I am wondering where the issue is.
> 
>  
> 
> I have created a prototype application that loads the font, render them
> in a txt file with a specific char in the pixels that are full of color
> ('o') and also prints out the metrics values. The code is in attachment.
> The final result is that old version of freetype load the glyph in this
> way 
> 
>  
> 
> ....
> 
> OOO.
> 
> O...
> 
> O...
> 
> O...
> 
> O...
> 
> O...
> 
>  
> 
> bitmapGlyph->metrics:
> 
> bitmapGlyph->metrics.height 7
> 
> bitmapGlyph->metrics.horiAdvance 4
> 
> bitmapGlyph->metrics.horiBearingX 1
> 
> bitmapGlyph->metrics.horiBearingY 7
> 
> bitmapGlyph->metrics.vertAdvance 11
> 
> bitmapGlyph->metrics.vertBearingX -2
> 
> bitmapGlyph->metrics.vertBearingY 3
> 
> bitmapGlyph->metrics.width 4
> 
>  
> 
>  
> 
> while the new one:
> 
>  
> 
> OOO
> 
> O..
> 
> O..
> 
> O..
> 
> O..
> 
> O..
> 
>  
> 
> bitmapGlyph->metrics:
> 
> bitmapGlyph->metrics.height 7
> 
> bitmapGlyph->metrics.horiAdvance 4
> 
> bitmapGlyph->metrics.horiBearingX 1
> 
> bitmapGlyph->metrics.horiBearingY 7
> 
> bitmapGlyph->metrics.vertAdvance 11
> 
> bitmapGlyph->metrics.vertBearingX -2
> 
> bitmapGlyph->metrics.vertBearingY 3
> 
> bitmapGlyph->metrics.width 4
> 
>  
> 
>  
> 
> So I can see a difference on the bitmap, but since the font file is the
> same, the code is the same, just the freetype libraries are different.
> 
> I am wondering if there is another parameter or test that I can do in
> order to understand what is going on; have you got any suggestions?
> 
>  
> 
> Thanks in advance for your consideration.
> 
> Mauro
> 
> 
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Freetype-devel mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/freetype-devel