2009年6月1日月曜日

5/31

Good morning !
I solved most of yesterday problem.
I didn't know a content of unknown address at yesterday.
But I could know most of a real identity.
It's alpha value of pixel.
Each values(RGBA) are exist in the each address.
Please see below.
+-----+ _
|alpha| |
+-----+ |
|blue | |
+-----+ Current pixel
|green| |
+-----+ |
|red | |
+-----+ _
|alpha| |
+-----+ |
|blue | |
+-----+ Next pixel
|green| |
+-----+ |
|red | |
+-----+  ̄
| : |
| : |

So, these exist it seem above.
I tried change value of alpha.
But It's not complete.
I could get result see below.
Image1
Image2
Image1+2


These alpha value are 100/255.
About half pixels are transparent.
But remains aren't, more likely, these pixel are seem to broken.
I wouldn't fix it until the WWDC2009 started, because I can't fix it soon.
And I think I should study another that useful for attend WWDC2009.
I took the statistic about differential of touchMoved method count per 10 seconds.
My algorithm of the "Nazca(change name from"YubiFude")" uses three CGLayer objects (depend on the situation, sometime 4layers).
Application speed is slowed by number of layers.
I took the statistic about it.
Result is this.



As a result, application is slowed by increase layer.
It's increase about 20counts per 10seconds.
I'll ask to Apple company member about more good algorithm in WWDC2009.

1 件のコメント: