I have winapi application with single window which background is like
this
But problem is that shape it non-trivial. It isn't just rectangle. It is
complex mask
What functions do I need to achieve something like this?
I'm interested only in those relevant to masking background. Mask will be moved using mouse but this part is fairly easy once I got mask.
Hope you like my paint skillz
You can change the shape of a window (for painting and hit-testing) using SetWindowRgn.
You can create a region from rectangles, ellipses or polygons and combine them to create more complex shapes.
Related
I couldn't find an example of a block of text not rendered to a rectangular area.
Ideally, it would be nice if ID2D1HwndRenderTarget.DrawText() would let me provide a polygon Geometry instead of a rectangle.
I've tried adding a Direct2D Layer with contentBounds, thinking it might skip rendering text within those layers. It didn't work as expected, it just blocked render to the area still emulating text underneath.
I've also tried applying a rectangular area to hwnd window itself. It too blocked render but didn't shift text.
IDWriteTextLayout only supports rectangular layouts, but DirectWrite supports any shape you can think of by using the lower level functions (text analysis, glyph measurement, glyph shaping). It's no easy task to write your own text layout from scratch, but I wrote a Windows 7 SDK sample containing a "FlowLayout" that demonstrates a circle and a few other simple shapes. It doesn't take arbitrary geometry, but you may be able to adapt it to your needs (see FlowLayoutSource::GetNextRect for computing the width of each line).
https://github.com/pauldotknopf/WindowsSDK7-Samples/tree/master/multimedia/DirectWrite/CustomLayout
DirectWrite only supports rectangular layouts, so you can't get anything more complicated automatically. You'll have to implement layout functionality yourself if you want it to work differently. Clipping arguments, like you already observed, have nothing to do with text layout.
I want to do what they do in Minecraft and that is take a 2d item and make a it 3d.
This:
to this:
Anyone have idea how I would do this without making a model in a seperate program
Well, if you have proper images with a nice contrast to the background, you could try to read the image's pixel value and contrast and try to find the outline of those images. As a next step, you would use Three.Shape or Three.Path and trace the outline with LineTo. Afterwards, you use an ExtrudeGeometry and extrude the shape into a 3d-object.
The main problem is finding the outline, i guess. This would only work with proper images and depending on the quality you want to achieve, you will have to implement the proper algorithms, see edge detection filters maybe.
Or being very simplistic, read the pixel values, if they are white or close to white, ignore them and if they are not, Draw another line.... something like that...
How to spread color Top to Bottom using CCTintTo method in sprite.
Because i am using CCTintTo method to spread color in sprite. But i want to look like spreading color in top to bottom.
What to do for this types of animation.
Thanks in advance.
If you are using cocos2d 2.0+, you could write a shader to do this, and set the shaderProgram property of the sprite. Not that hard, follow the examples in the distribution. My first shader took me .5 work days to get to work, and maybe another .5 workday to properly integrate that technique properly in my overall software architecture. g'luck :).
Look here for an introduction to shaders, play with it in a side project until you are comfortable to integrate in your main trunk.
That's not possible with color tinting. Changing a node's color property or using the tint actions will only tint the entire sprite with a single color, there will be no gradient.
You would have to custom draw the sprite and apply/adapt the gradient rendering code from CCLayerGradient.
Yes CCLayerGradient is what you are looking for.By the way what technique you are using for coloring.If you are using CGContextSetRGBFillColor method to fill the sprite color then it would be tricky but if you are using images to fill color then nice sequence of images plist can be used to produce animation which will give you the exact effect you are looking for.
Take the sequence of images and animate them using CCSpriteBatchNode otherwise if you want to use RGB colors to fill sprite then you have to look for gradient effect.
I'm working on a simple drawing application. My line is constructed using polygons and things look good so far. I would like to add a transparency feature and I used glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA); for that reason. However, because my polys sometimes overlap, I get the ugly result shown in the picture (multiple layers of transparency). What I would like to get is the figure in the left(no overlapping because there is no transparency), with an overall transparency.
I guess I could do this by keeping the polys from overlaping, but that would be a overkill for this task I think. There should be a way to control them at drawing time, but being a beginner with OpenGL doesn't help.
I'm sorry, but the way transparency works does not allow you to do what you want without manually keeping the polygons from overlapping. The way that transparency works is that it takes the colour of the surface below it, and uses the blending function you specify in order to calculate the final colour of the pixel.
In your program you are drawing multiple polygons with alpha on top of each other. That means that their colours add up, giving the result you see.
I've never actually written a drawing application, but you could perhaps take a look at triangle strips to draw your lines. They allow you to extend the line point by point, and make sure the geometry won't overlap with itself.
I am using layered windows and drawing a rounded rectangle on the screen. However, I'd like to smooth out the jagged edges. I think that I'll need alpha blending for this. Is there a way I can do this with GDI?
CreateDIBSection. Fill in the BITMAPINFOHEADER with 32bpp. Fill in the alpha channel with pre-multiplied alpha and youre good to go.
AlphaBlend is the API to actually blit 32 bpp bitmaps with an aplha channel.
You can do this in C# using the LockBits method of the BitMap class (see this question for an explanation). You definitely don't want to use GetPixel and SetPixel for this, as they are hideously slow (even though you'd just be manipulating the edges and not the entire bitmap).
Any chance of using GDI+ instead of GDI? It supports antialiasing and transparency right out of the box.
There isn't an easy way to do such drawing with just GDI calls. What you want isn't just alpha blending: you want anti-aliasing. That usually involves drawing what you want at a larger resolution and then scaling down.
What I've done in the past for similar problems is to use an art program to draw whatever shape I want (e.g. a rounded corner) much larger than I needed it in black and white. When I wanted to draw it I would scale the black and white bitmap to whatever size I wanted (using a variant of a scaling class from Code Project). This gives me a grayscale image that I can use as an alpha channel, which I'd then use for alpha blending, either by calling the Win32 function AlphaBlend, or by using a DIBSection and manually changing the appropriate pixels.
Another variation of this approach would be to allocate a DIBSection about four times larger than you wanted the final result, draw into that, and then use the above scaling class to scale it down: the scaling from a larger image will give the appropriate smoothing effect.
If all this sounds like quite a lot of work: well, it is.
EDIT: To answer the title of this question: you can create a bitmap with an alpha channel by calling CreateDIBSection. That won't on its own do what you want though, I think.