SearedIce

Some of you may be familiar with "AutoHotKey" or "AutoIt". These programs give a Windows computer Apple-script-like capabilities. You can automate simple tasks that are boring/time-consuming, bind new hotkeys for things, run software, manage the clipboard, etc. etc. etc. I thought it'd be cool to use the icons on my desktop in my desktop's background in some way ... to mimic next-gen OS GUI effects, if at least on the desktop... I took a screenshot of my current desktop to show what I mean (right now, glowing icons): http://www.seas.upenn.edu/~drake/bgu/BGU01.jpg Anyway... so I wrote a script for AutoIt that does the following: Minimizes all windows Changes the desktop background to "None" Screenshots the desktop icons Opens PDN Pastes the screenshot to a new image Applies a Gaussian Blur Saves the image Applies the image as my computer background The steps could be changed, and any effects (or combinations of effects could be used). You could even have it add an image to the mix (a normal background). Here's the script file: http://www.seas.upenn.edu/~drake/bgu/BackgroundUpdate.au3 With a script, if the icons on my desktop change (move files around or install a new program), I just run the script and it takes less than 30 seconds to produce a new background. Now, this may not seem desireable to most, but I happen to like how it works out and I think the concepts could be applied to other PDN automation tasks. Also, it sure did impress my girlfriend! AutoIt (open source = free!): http://www.autoitscript.com/autoit3/ -John

If you want to write the algorithm yourself, PDN's Codelab is a pretty interesting environment to test in. I did some basic shape-recognition algorithm testing in Codelab...made it easy to see results as I worked...

Not for me...for me, it automatically selects everything on the canvas and begins rotating it...

You may be looking for "feathering," a feature not in PDN. If you truly are talking about antialiasing, there's only really one genuine effect it has, and the first squiggle surely has it. For example, straight parts of the line are prefectly sharp, and angled parts attempt to be perfectly sharp. In the 3rd line there, nothing is sharp.

Try a bit to figure it out, and then (if you need it): To convert from degrees to radians: pi*(degrees/180)=radians Radians to degrees: (radians/pi)*180=degrees

Arctan / inverse tangent That's what you need for finding the angle. y=mx+b m = rise/run rise/run = y/x tan(a)=y/x so... arctan(y/x)=a (what I have there is not at all code...I'm not sure of the syntax for the trig functions in PDN source/C#...I've used them in effects before, though, so if you need, I can help figure it all out) You may need to adjust this in the code a bit (since the images have the origin at the top left not the bottom left), but it should do the job.

Awesome! Thanks so much!

Nice Can't wait.

The "white" areas in the Paint.net logo at the top-left of this page are not actually 100% white. The white on that image has a brightness value of 99. See for yourself It's been bothering me because my laptop screen is actually showing that small difference in brightness...the logo looks like it's in a grey box.

I'm not sure exactly which build this was implemented in, but I've found that the Ctrl+Spacebar shortcut for the thumbnail dropdown is rather inconvenient. I'm used to holding both Ctrl and Space at the same time to both zoom and pan an image for detailed work. Now, however, when I'm doing this sort of work, I have to make sure I totally stop pressing one of the keys before depressing the other. I often accidentally hit them both at the same time. There's no easy way now to activate both zoom and pan, moving around an image efficiently. Thanks for considering this problem, John

If the colors/shades are similar (and around them are sufficiently contrasting ones) you can use the magic wand tool to easily select the colored area. If there isn't enough contrast you could duplicate your image layer, increase its contrast, make the selection there, then switch back to your original image to use the selection how you wanted.

Yeah this is the thruth with such projects. However, it seems Rick is willing to work on PDN without pay for now. I think his donation system even actually sends the money to his college.

LOL. But really, nice effects guys...

Education. Also, maybe the text is ugly...? As long as you aren't copying beyond what you're allowed, it's pefectly legal. This is a very tricky topic because, for example, even your browser technically copies the image when it downloads it. As long as the stuff is for totally private, education-oriented work, no one will ever know and no one will ever care.

You just need to learn the ways of modern programming. There are a lot of approaches to this. I prefer the "teach yourself" method... Buy a book or two. Look up tutorials online. Maybe start with a simpler language, then make the switch. If you really want to do it hard-core style, download a bunch of C# code and figure it out on your own by looking at it. You can take a programming class...if you're not in a school that offers this, your local community college probably has a course you can take now or during next summer. Ask your friends...you might be able to find someone who can help you 1-1 mentor-style.

Removing the text isn't illegal. Using the image beyond what its license allows is illegal. For example, erasing the author's name from a new, copyrighted book isn't illegal, but copying the book (with or without the name) is.

Well... lol

Barkbark's right... You're probably using JPEG files (*.jpeg, *.jpg) so when you go to save the image (or when you do a "File->Save As"), just lower the slider for JPG quality...you should be able to see what it's doing in the preview window...that way you can tell when you've reduced the quality too much. You probably want to keep backups of the originals somewhere too.

Lol...random is funny.

Since the photo itself is a bit blurry, a 1px Gaussian blur does not hurt the image, but it does get rid of some of the noise...see for yourself how you like it.

Leif, try turning off Alpha Blending before you do that next time...sounds like that might be your problem. And Hellfire, I haven't noticed what you're talking about...that's strange.

The GNU license was originally written for the GNU project: The GNU license is commonly called the GPL or "General Public License"

Yeah...this is a problem with the code that currently determines where your cursor is over the image. I think Rick said it will eventually be fixed with a more precise system. I know how annoying this can be nickspacek... If you just have to move it, you can actually click and drag outside of the selection area (as long as your cursor is a black arrow, and not a hand).

Raven, that will work, but it's going to take a long time if the image is the size of any normal photograph...there could be thousands of noisey areas on the image! There are some tricks you can use to make the noise less apparent, but there's no real way to get rid of it entirely. For example, you can blur the image a bit, but then the image will still be blurry and will probably still have noise. Some people have methods they love to no end, but I haven't seen anything that does a perfect job...

The little icon for the Sepia adjustment has opaque edges that should be transparent. I can get a screenshot if you want. Also, it's not obvious anymore that the "Adjustments" only affect a single layer...do you think this will be a problem?