Re: [ft-devel] ftgrays - exact pixel coverage?

[Alexei Podtelezhnikov]

On Wed, Feb 8, 2012 at 11:36 PM, Werner LEMBERG <address@hidden> wrote:
>
>>> I believe there is discrepancy between what's written in ftgrays.c
>>> and its actual functioning.  It is said it "computes the _exact_
>>> coverage of the outline on each pixel cell".  But the following
>>> program shows this is not true.
>
> Good point!  Thanks for reminding us.  Consequently, I think that...
>
>> To speed up the rendering ftgrays flattens the curves, i.e., it
>> renders them as short straight segments.  This explains your
>> observation.  Note that the statement is still true for straight
>> segments.
>
> ... this deserves a comment in the code (near the word `exact' :-).
> Alexei, can you prepare something?
>

What was the situation before the flattening was introduced?
As it stands right now with permitted deviation of up to ONE_PIXEL / 8,
The coverage can be missed by as much as 0.169 of a pixel in the
absolute worst (diagonal) case and less than half of that on average.
So we are still better than 8 levels of gray from Redmond on average.
How much better do we want to be is a real question.