Wow, a breakthrough! I read in another post that Android's framebuffer uses the RGB_565 color mode. So I tried creating a ARGB_8888 Bitmap and did all of my rendering to that Bitmap. After I was done I copied the Bitmap to the Canvas that I get from the onDraw()-method. This resulted in all semi- transparent images looking MUCH better. Unfortunately this means slower GUI updating, since I have to draw a full 320x460 (skipping the top bar) image to the screen, after each render.
From this I draw the conclusions: * If you draw images into an ARGB_8888 Bitmap, all alpha blendings are done in 8 bit precision. Then you can draw the resulting image to the Canvas provided by the onDraw()-method, and it will be converted to RGB_565 * When drawing an image to the Canvas provided by the onDraw()-method, the result is immediately converted to RGB_565 format. This means that the next image, drawn on top, actually does its blending to a background that is already in the RGB_565 format and the result looks horrible. I have the following request to the developers of Android: Is there ANY possibility that the conversion to RGB_565 16-bit graphics can be done AFTER all the rendering is finished? Perhaps at the same time as the double buffer swapping takes place, or such? If it doesn't mean a big performance hit, of course. Best Regards, Daniel On 5 Apr, 19:05, hanni <[EMAIL PROTECTED]> wrote: > Oh, I just noticed that if I try to do it the other way, that is, > extract the pixels form the image and applying my own blending > algorithm, another problem arises. There seems to be a bug in the > Bitmap.getPixels() implementation, that causes any pixel that has a > non 0xFF alpha, to automatically lose all its color and become black. > So if the image has a pixel with a full red colour of 0xFFFF0000, it > stays red, but if it is a semi-transparent red color such as > 0x7FFF0000, it is returned as the color 0x7F000000, with both the > getPixels() and the getPixel() methods. > > On 5 Apr, 18:02, hanni <[EMAIL PROTECTED]> wrote: > > > > > Hi, my application makes heavy use of semi-transparent .png's for > > interesting graphics. I'm used to this, developing J2ME applications. > > The alpha blending algorithm in Android, however, seems to be less > > than adequate. Everything that is semi-transparent is restricted to > > very few shades of color, and that goes for everything behind a semi- > > transparent image aswell. > > > Does anyone know of a setting to enable a more accurate alpha blending > > algortihm, even if it might require more CPU power? Currently, when I > > really need something to look accurate, I would have to create an > > array from that image and run my own blending algortim on it, > > converting it back to an image afterwards.- Dölj citerad text - > > - Visa citerad text - --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Android Developers" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] Announcing the new M5 SDK! http://android-developers.blogspot.com/2008/02/android-sdk-m5-rc14-now-available.html For more options, visit this group at http://groups.google.com/group/android-developers?hl=en -~----------~----~----~----~------~----~------~--~---
