Blurry text in WPF / Silverlight
-
Pete O'Hanlon wrote:
You may find this[^] post interesting.
You gotta love this reply: Just to add a comment, when I look at your image on my computer, it looks much better with the antialiasing turned on. Dude. Antialiasing does not change how a JPG looks. :rolleyes: Marc
Marc Clifton wrote:
Dude. Antialiasing does not change how a JPG looks.
It does if you close your eyes, cross your fingers and hope blindly that the pesky users go away.
Deja View - the feeling that you've seen this post before.
-
Pete O'Hanlon wrote:
You may find this[^] post interesting.
You gotta love this reply: Just to add a comment, when I look at your image on my computer, it looks much better with the antialiasing turned on. Dude. Antialiasing does not change how a JPG looks. :rolleyes: Marc
-
Pete O'Hanlon wrote:
You may find this[^] post interesting.
You gotta love this reply: Just to add a comment, when I look at your image on my computer, it looks much better with the antialiasing turned on. Dude. Antialiasing does not change how a JPG looks. :rolleyes: Marc
I think they are saying antialias and meaning Cleartype. And in such case the screenshot may look different on different computers, Because different monitors may have different RGB element layout.. if its RGB or BGR , since cleartype is a trick wich uses just that.. (http://en.wikipedia.org/wiki/ClearType[^] for those who refuse to beleive it ;) )
-
I think they are saying antialias and meaning Cleartype. And in such case the screenshot may look different on different computers, Because different monitors may have different RGB element layout.. if its RGB or BGR , since cleartype is a trick wich uses just that.. (http://en.wikipedia.org/wiki/ClearType[^] for those who refuse to beleive it ;) )
Roger Alsing wrote:
I think they are saying antialias and meaning Cleartype.
It's a JPG! :) The guy was saying that the antialiasing looked fine when he viewed the JPG on his computer! How can that be--there's no ClearType involved rendering a JPG. :rolleyes: Marc
-
Roger Alsing wrote:
I think they are saying antialias and meaning Cleartype.
It's a JPG! :) The guy was saying that the antialiasing looked fine when he viewed the JPG on his computer! How can that be--there's no ClearType involved rendering a JPG. :rolleyes: Marc
If I take a screenshot of my screen here (with cleartype enabled) The text on that screenshot will be pixelated according to _my_ screen rgb layout If I send that screenshot to you, and you happen to have another RGB emelent layout, eg BGR instead, the text on that image would look bad when you view it.. because the pixels in the text on the image was rendered according to another rgb layout But if he is infact talking about real antialias and not cleartype, then ofcourse the image will not change..
-
If I take a screenshot of my screen here (with cleartype enabled) The text on that screenshot will be pixelated according to _my_ screen rgb layout If I send that screenshot to you, and you happen to have another RGB emelent layout, eg BGR instead, the text on that image would look bad when you view it.. because the pixels in the text on the image was rendered according to another rgb layout But if he is infact talking about real antialias and not cleartype, then ofcourse the image will not change..
Roger Alsing wrote:
because the pixels in the text on the image was rendered according to another rgb layout
A pic is a pic is a pic. If what you say were true, then any pic would look like crap simply becaue of some difference in the element layout. What am I missing here? Marc
-
Roger Alsing wrote:
because the pixels in the text on the image was rendered according to another rgb layout
A pic is a pic is a pic. If what you say were true, then any pic would look like crap simply becaue of some difference in the element layout. What am I missing here? Marc
You are missing how Cleartype text is rendered.. Cleartype is based on a trick with sub pixels layout of rgb elements of some monitors are: R G B R G B R G B so you can get a virtual resolution of 3 times your real horizontal res. eg if you have 1024x768 , the cleartype resolution is 3072 * 768 lets say that we render an "A" with 6 x 8 pixels in Cleartype RGB layout:
one pixel = [R G B]
R G B R G B R G B R G B R G B R G B
. . . . . . # # # # . . . . . . . .
. . . . . # # # # # # . . . . . . .
. . . . # # # . . # # # . . . . . .
. . . # # # . . . . # # # . . . . .
. . # # # # # # # # # # # # . . . .
. # # # . . . . . . . . # # # . . .
. # # # . . . . . . . . # # # . . .# # # # . . . . . . # # # # # . .
if you take a screenshot of an image containing an "A" rendered this way. and then present it on a monitor with BGR layout. it would appear as:
included extra spaces to show the BGR elements
one pixel = [B G R]B G R B G R B G R B G R B G R B G R
. . . . . . # # # . . # . . . . . .
. . . # . . # # # . # # . . . . . .
. . . # # . . . # # # # . . . . . .
. . . # # # . . . # # . . . # . . .. . # # # # # # # # # . # # . . .
# . . . # . . . . . . # # # . . .
# . . . # . . . . . . # # # . . .
# # # # . . . . # . . # # # . .
(its still the exact same pixels on the image.. its just that the layout of the RGB elments on the monitor are different..) The same does NOT happen for normal images, since they do not rely on this sub pixel trick.. (You better give me a 5 for the ascii art ;-)) On my system a crisp "A" looks like this if zoomed. http://www.puzzleframework.com/Roger/cleartypea.gif[^] But if I make a screenshot of the blurred WPF text containing an "A" , the red and blue would be reversed.. Found this on wikipedia also: http://en.wikipedia.org/wiki/Subpixel_rendering[^] It might explain it better than my ascii art :P
modified on Thursday, May 8, 2008 2:11 PM
-
You are missing how Cleartype text is rendered.. Cleartype is based on a trick with sub pixels layout of rgb elements of some monitors are: R G B R G B R G B so you can get a virtual resolution of 3 times your real horizontal res. eg if you have 1024x768 , the cleartype resolution is 3072 * 768 lets say that we render an "A" with 6 x 8 pixels in Cleartype RGB layout:
one pixel = [R G B]
R G B R G B R G B R G B R G B R G B
. . . . . . # # # # . . . . . . . .
. . . . . # # # # # # . . . . . . .
. . . . # # # . . # # # . . . . . .
. . . # # # . . . . # # # . . . . .
. . # # # # # # # # # # # # . . . .
. # # # . . . . . . . . # # # . . .
. # # # . . . . . . . . # # # . . .# # # # . . . . . . # # # # # . .
if you take a screenshot of an image containing an "A" rendered this way. and then present it on a monitor with BGR layout. it would appear as:
included extra spaces to show the BGR elements
one pixel = [B G R]B G R B G R B G R B G R B G R B G R
. . . . . . # # # . . # . . . . . .
. . . # . . # # # . # # . . . . . .
. . . # # . . . # # # # . . . . . .
. . . # # # . . . # # . . . # . . .. . # # # # # # # # # . # # . . .
# . . . # . . . . . . # # # . . .
# . . . # . . . . . . # # # . . .
# # # # . . . . # . . # # # . .
(its still the exact same pixels on the image.. its just that the layout of the RGB elments on the monitor are different..) The same does NOT happen for normal images, since they do not rely on this sub pixel trick.. (You better give me a 5 for the ascii art ;-)) On my system a crisp "A" looks like this if zoomed. http://www.puzzleframework.com/Roger/cleartypea.gif[^] But if I make a screenshot of the blurred WPF text containing an "A" , the red and blue would be reversed.. Found this on wikipedia also: http://en.wikipedia.org/wiki/Subpixel_rendering[^] It might explain it better than my ascii art :P
modified on Thursday, May 8, 2008 2:11 PM
Roger Alsing wrote:
(You better give me a 5 for the ascii art )
Done. :) <blockquote class="FQ"><div class="FQA">Roger Alsing wrote:</div>The same does NOT happen for normal images, since they do not rely on this sub pixel trick..</blockquote> Well, I must be extra dense today, because your wonderful explanation is still not connecting the dots for me (pardon the pun). I guess if I look at it the other way, that you have to set up subpixeling correctly for your monitor type, then on a monitor with a different RGB layout, it would look fuzzy. But then, that sort of begs the question, why aren't pictures distorted depending on the RGB layout? Or are they, but we just don't notice? Marc
-
Roger Alsing wrote:
(You better give me a 5 for the ascii art )
Done. :) <blockquote class="FQ"><div class="FQA">Roger Alsing wrote:</div>The same does NOT happen for normal images, since they do not rely on this sub pixel trick..</blockquote> Well, I must be extra dense today, because your wonderful explanation is still not connecting the dots for me (pardon the pun). I guess if I look at it the other way, that you have to set up subpixeling correctly for your monitor type, then on a monitor with a different RGB layout, it would look fuzzy. But then, that sort of begs the question, why aren't pictures distorted depending on the RGB layout? Or are they, but we just don't notice? Marc
Marc Clifton wrote:
But then, that sort of begs the question, why aren't pictures distorted depending on the RGB layout? Or are they, but we just don't notice?
Exactly. We can see the effect on cleartype text because there are enough pixels (that exploit the effect) in the same area to make the effect visible. Subpixel drawing is best applied in high contrast scenarios. like text or line drawings. Otherwise the effect will not be noticable.. in a normal picture where the trick is not used, you will not notice if one pixel leans to the left while another leans to the right.
-
Marc Clifton wrote:
But then, that sort of begs the question, why aren't pictures distorted depending on the RGB layout? Or are they, but we just don't notice?
Exactly. We can see the effect on cleartype text because there are enough pixels (that exploit the effect) in the same area to make the effect visible. Subpixel drawing is best applied in high contrast scenarios. like text or line drawings. Otherwise the effect will not be noticable.. in a normal picture where the trick is not used, you will not notice if one pixel leans to the left while another leans to the right.
Roger Alsing wrote:
Subpixel drawing is best applied in high contrast scenarios. like text or line drawings.
That's what I was thinking. OK, now I understand. Thank you so much for sticking with me on this. :) Marc
