Re: [Devel] Getting accurate advances

[David Bevan]

Werner LEMBERG wrote:
> > Other similar code (producing rounded x_scale values) can be found
> > at ftobjs.c line 2081
>
>
> Do you mean this (y components omitted)?
>
>   FT_Set_Pixel_Sizes( FT_Face  face,
>                       FT_UInt  pixel_width,  
>                       FT_UInt  pixel_height )
>   {
>     ...
>     metrics->x_ppem = (FT_UShort)pixel_width;
>
>     if ( face->face_flags & FT_FACE_FLAG_SCALABLE )
>     {
>       metrics->x_scale = FT_DivFix( metrics->x_ppem << 6,
>                                     face->units_per_EM );
>
>       ft_recompute_scaled_metrics( face, metrics );
>     }
>     ...
>
> Isn't this already correct?  Here, x_ppem is indeed an integer value.
>
>
> > pfrdrivr.c line 102,
>
>
> Probably yes, but I have no idea how PFR fonts work.
>
>
> > and ttdriver.c line 224.
>
>
> Mhm, rather no.  This code modifies the x_scale value for the hinting
> engine only, as far as I can see.  It doesn't influence
> linearHoriAdvance.

Werner,

Thansk for your feedback.

However, I think your responses explain why I'm unwilling to change the 
calculation of x_scale. I don't know why x_scale is set to different 
values in different contexts or what uses are made of it after it is set.

So, I don't know why you think FT_Set_Pixel_Sizes() is correct, or why 
ttdrivr doesn't need changing, or why you guess that the pfrdrivr code 
should be changed (though I can't actually see how it could be changed).

All I do know is that x_scale is used in various ways and that the value 
stored in it isn't currently acceptable for calculating linearHoriAdvance.

And I worry that if there's a good reason for x_ppem to be rounded to 
the nearest pixel (for hinting to work correctly?), then perhaps it's 
important that the scale factor is adjusted (as it is) to match the 
adjusted x_ppem.

I'm sorry I can't really help you much more than to point out the 
problem, but I don't understand very much about the innards of FreeType.

For now I think I may just stick with my work-around.

Anyone else got any views on this?

David %^>

--
 ________________________________________________________
 David Bevan, Emtex Ltd.
 Emtex House, Station Road, Kings Langley, Herts. WD4 8LH England
 Tel:    +44 (0)1923 270882
 E-mail: address@hidden
 Web:    www.emtex.com