| Time |
S |
Nick |
Message |
| 02:14 |
|
|
nanoz joined ##friendlyjava |
| 02:18 |
|
nanoz |
<nanoz> there is a project called evolys this has many set of java projects independent/dependent |
| 02:18 |
|
nanoz |
<nanoz> so on svn checkout via subclipse, eclipse didnt know the subproject were project |
| 02:18 |
|
nanoz |
<nanoz> has anyone encountered similar problem? |
| 07:53 |
|
|
kirua joined ##friendlyjava |
| 12:39 |
|
mr_lou |
Hmmm..... |
| 12:42 |
|
mr_lou |
It appears as if BufferedImage.setRGB() doesn't actually SET pixels, but rather adds them to the existing pixels? When using alpha, it doesn't actually SET a pixel with that value, but rather merges it with the pixel value that was already there. At least it looks that way. |
| 12:43 |
|
mr_lou |
But even when trying to "clear" the image by called setRGB() with an array of 0xff000000 pixels - or using fillRect() to clear - doesn't change anything. |
| 12:43 |
|
mr_lou |
I feel I'm missing something. |
| 13:13 |
|
* mr_lou |
fiddles with AlphaComposite without effect |
| 15:02 |
|
* mr_lou |
pokes aditsu |
| 15:02 |
|
aditsu |
ow! |
| 15:02 |
|
mr_lou |
Halp! :-( |
| 15:03 |
|
aditsu |
can you paste your code? |
| 15:03 |
|
mr_lou |
Working on my rotation stuff. Wanna make pixels that are farther away look darker. Thought I'd use alpha to do this. And I think that part works. But when using alpha, it's like the previous image isn't cleared, so the transparent pixel is just drawn on top of the previous pixel, when using setRGB(). I don't get that. |
| 15:04 |
|
mr_lou |
Shouldn't setRGB() actually SET new pixels? And not merge them with the ones that are already there? |
| 15:05 |
|
aditsu |
not sure.. you were creating new images though, or did you change it? |
| 15:05 |
|
aditsu |
btw, the rotation code has lots of room for optimization |
| 15:05 |
|
mr_lou |
Doesn't matter. It'll just be used to create image that'll become a video. |
| 15:05 |
|
mr_lou |
*images |
| 15:06 |
|
mr_lou |
Previously I used one of those MemoryImageSource yes. Now I used BufferedImage. |
| 15:07 |
|
aditsu |
well, some code would help, but have you considered just changing the colors instead of using alpha? |
| 15:08 |
|
mr_lou |
Tried that too. Changing alpha just seems easier. |
| 15:08 |
|
mr_lou |
But apparently not. |
| 15:08 |
|
mr_lou |
Can't think of any clever stuff to do changing colours. |
| 15:08 |
|
mr_lou |
I'm calculating a "zindex", that ranges from 0 to 255. |
| 15:08 |
|
mr_lou |
Depending on z position of each pixel. |
| 15:09 |
|
mr_lou |
Would be easy to just make that alpha. |
| 15:09 |
|
aditsu |
you could simply do alpha mixing in your code |
| 15:09 |
|
mr_lou |
Not sure how to use that value to change colours. I'm not getting any useful results with the things I've tried. |
| 15:09 |
|
mr_lou |
That's what I'm doing. |
| 15:09 |
|
mr_lou |
alpha mixing |
| 15:09 |
|
mr_lou |
But that's when I get "ghost effects" because the previously image apparently isn't cleared. |
| 15:10 |
|
aditsu |
I mean do it manually instead of passing it to the image |
| 15:10 |
|
mr_lou |
Even when I try setting all 0xff000000 pixels BEFORE I set the other pixels, I still have a ghost effect. |
| 15:10 |
|
aditsu |
but either way, give me some code and I can try to fix it |
| 15:10 |
|
mr_lou |
Hm |
| 15:10 |
|
mr_lou |
You will need the image too. |
| 15:11 |
|
aditsu |
you could try to just reproduce the problem in an test program |
| 15:11 |
|
* mr_lou |
cleans up code a bit |
| 15:11 |
|
aditsu |
a* |
| 15:13 |
|
mr_lou |
The code: https://pastebin.com/hhrJLY4f |
| 15:14 |
|
mr_lou |
The image: http://www.villumsign.dk/logoinvert.png |
| 15:14 |
|
mr_lou |
Should mention though, that it's all coded to run on a 1920x1080 screen. Will probably not look right on other resolution. |
| 15:16 |
|
aditsu |
1920x1200 here |
| 15:16 |
|
mr_lou |
hm |
| 15:16 |
|
mr_lou |
Well... |
| 15:17 |
|
mr_lou |
I guess we'll find out then. :-) |
| 15:18 |
|
aditsu |
ok I need some time |
| 15:19 |
|
mr_lou |
It's 95% the same code as last time. |
| 15:20 |
|
mr_lou |
Just with a PNG image now, instead of an IndexColorModel MemorySourceImage. |
| 15:24 |
|
aditsu |
mr_lou: you may want to clear g before drawing the image |
| 15:24 |
|
mr_lou |
You mean with a fillRect()? |
| 15:24 |
|
mr_lou |
Tried that. |
| 15:24 |
|
mr_lou |
Changed nothing. |
| 15:24 |
|
aditsu |
or clearRect |
| 15:24 |
|
mr_lou |
Hmm. |
| 15:24 |
|
mr_lou |
Didn't try clearRect(); |
| 15:24 |
|
aditsu |
g.clearRect(0, 0, w, h); |
| 15:26 |
|
mr_lou |
doh... dumbass me.... I never cleared anything on g.... I was doing it on screenG. |
| 15:28 |
|
mr_lou |
That was it of course. |
