On Fri, Aug 13, 2010 at 4:31 PM, Behdad Esfahbod <span dir="ltr"><<a href="mailto:behdad.esfahbod@gmail.com">behdad.esfahbod@gmail.com</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">On 08/13/10 01:09, Soeren Sandmann wrote:<br>
> Krzysztof Kosiński <<a href="http://tweenk.pl" target="_blank">tweenk.pl</a>@<a href="http://gmail.com" target="_blank">gmail.com</a>> writes:<br>
><br>
>> This is the second attempt at bitmap supersampling in Pixman.<br>
><br>
> For people who are curious, this is a big quality improvement:<br>
><br>
> original image (3200 x 2000):<br>
> <a href="http://www.daimi.au.dk/~sandmann/house.jpg" target="_blank">http://www.daimi.au.dk/~sandmann/house.jpg</a><br>
<br>
</div>Great image for testing.<br>
<div class="im"><br>
<br>
> without patch:<br>
> <a href="http://www.daimi.au.dk/~sandmann/house-before.png" target="_blank">http://www.daimi.au.dk/~sandmann/house-before.png</a><br>
><br>
> with patch:<br>
> <a href="http://www.daimi.au.dk/~sandmann/house-after.png" target="_blank">http://www.daimi.au.dk/~sandmann/house-after.png</a><br>
<br>
</div>Interesting that there's still aliasing going on, but blurred.</blockquote><div> </div><div>I think any algorithm will show some sort of aliasing for some images. Look at this page for some comparisons:</div><div>
<br></div><div><a href="http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/down_sample.htm">http://www.xs4all.nl/~bvdwolf/main/foto/down_sample/down_sample.htm</a></div><div> </div><div>The algorithm that he implemented is certainly not the best, but its certainly better than what we have now... </div>
</div>