I am trying to smooth text rendering using anti-aliasing.
But it's not anti-aliased.
First line is a png image created using pango and cairo.
Second line is just an html <span> tag. It's in firefox, Ubuntu with Gnome DE.
The difference can be better understood if you compare "W" and "v"
between two lines.
The code responsible to draw text can be found on http://pastie.org/1073683
Font options are set on lines 17 and 20 like so:
// setting up antialiasing
cairo_font_options_set_antialias(cfo, CAIRO_ANTIALIAS_GRAY);
// set the font options to cairo
cairo_set_font_options(this->cro, cfo);
Could anyone please tell me how can I make those two lines look the same?
cairo_font_options_set_antialias(cfo, CAIRO_ANTIALIAS_GRAY);
You asked for it, you got it. To get a match you'd need CAIRO_ANTIALIAS_SUBPIXEL. This is however not appropriate when you draw text to an image that might be displayed on another machine. There's no guarantee that the monitor on that machine is an LCD panel with the RGB stripes in a predictable order. Or that it in landscape orientation. Or that it is displayed with the exact original size. When there's a mismatch, the text will look quite poor.
Related
To set the icon for a settings window, I use this code:
SDL_WM_SetIcon(IMG_Load("settings.png"),NULL);
settings.png is a 32x32 PNG image with semi-transparency placed in the same folder as the executable. The problem is that the title bar looks like this:
In case you can't see the black on the screenshot above the first time you look, here's a zoomed version of it with an arrow showing an example of the parts that I'm talking about:
As you can see, at the places where there is supposed to be semi-transparency, there is black instead which makes it a bit ugly. I tried to take away the semi-transparency with IcoFx using "Insert new image" and then selecting "True Color (24 bits)" (which takes away the semi-transparency and replaces the image with what it would look like on a white background without changing parts that are totally transparent), but I still get the same result. I also tried reducing the size of the image to 16x16 but that makes it even worse. How do I fix this? If it's possible, I would prefer to keep the semi-transparency since the taskbar supports it just fine.
I would happily provide a screenshot of this, however the problem is the captured image, is much larger than my actual desktop.
I am completely frustrated with this as I have tried using BitBlt with the desktop hdc AND the new "Graphics" commands.
My actual desktop resolution is 1920x1080 - 1080p .
BitBlt and "Graphics" both return that my resolution is 1536x864 # 96 DPI.
A form (WinForm), Maximized, borderless, and irrelevant of scaling mode the form is set to, also shows 1536x864 # 96 DPI.
Now the image that is captured, is like it is being done from 1920x1080, but clipping the region 1536x864 as the screenshot.
If I do PrintScreen directly using Prtscn button, I get the entire image, but still it is about 1.5-2x larger than what I actually see.
What I am looking for -- is a resolution for how I can take a picture of what is on my screen in the scale/dpi/whatever is going on here that it visually looks like. I have written a screen capture program, and using a few different examples for the RubberBand form (overlay form to select a region of the screen by drawing a box), and as you can imagine, this scaling crap is causing those box captures to be offset, and the contents are zoomed.
This is very annoying -- even to explain, however I am positive that most of you are familiar with the terms I use, and also know what to expect from taking a screenshot, so my explanation above should be pretty clear as to what my problem is.
Example/Consideration
Imagine, taking a picture of a window that is 300x300, and getting the top left 150x150 of that zoomed to 300x300 completely skipping the remainder of the window. Resulting image is still 300x300, but it's not what you selected.
Now imagine, you grab a picture of your screen by the only dimensions you can get programmatically, and then put the image into a picturebox. Even though both your screen and the picturebox claim to be the same dimensions and dpi, the image in the picturebox requires scrolling even if the picturebox is maximized to fullscreen on a borderless with no borders / etc. -- again, the picture is zoomed, but how is it still reporting that it's the same size as the form XD (comparing Graphics or BitBlt dimensions with the actual form. also tried comparing picturebox contents, and still same effect)
This, is EXACTLY what the effect is that is happening. When I try to capture a region or segment of the screen. I am not sure why windows api/crl is lying about this seemingly trivial stuff, however there must be a way to accurately obtain screenshots/capture regions without this faux zoom effect -- across all resolutions.
Thank you Hans Passant for pointing me in the right direction.
To add "true" dpi scaling support to a winforms application, you can make it so by adding the following block to your manifest :
Project > Add New Item > Visual C# Items > Application Manifest File
One the file has been added, open it up and look for a line like
</asmv1:assembly>
Whatever the "asmv" number is, (in the example above it is 1), use that to format the code:
<asmv1:application>
<asmv1:windowsSettings xmlns="http://schemas.microsoft.com/SMI/2005/WindowsSettings">
<dpiAware>true</dpiAware>
</asmv1:windowsSettings>
</asmv1:application>
Paste the above code (changing the asmv1 to whatever version the manifest is), just above the final closing line for the ""
Also, make sure your forms are set to AutoScale to dpi (and all sub-elements).
Now, I'm having a very difficult problem
I convert a font to hex code, and can resize it
as follows :
1. I use lcd font maker software to create char A font Arial size 18
2. Then I convert to hex code
(see picture to understand)
How do I resize the image when data input is hex code or binary, and result is hex code or binary?
Please suggest me, or document related to this issue.
Here i dont know what do you mean by resize the image
1> In lcd font maker software you have created one image
2> Now you have encoded that image in digital form and made array of hex code.
Now i think you want to pass this array of hex code to some other program and want to recreate/decode that image back ? if then you can write one program using same reverse logic which can decode that data and create original image back.
And if you want to resize image then you need to understand/use the concept of pixcel sampleling and all stuff
Wiki page "Image scaling" could be good stating point. There is list of different scaling methods.
Result of scaling may look ugly, especially if you have 2 color display.
So consider use of two or more presized fonts.
I work on gschem, a free software tool for editing electronics schematic diagrams. Recently we have encountered a problem using a GtkScrolledWindow containing a GtkTextView.
Context
Recent versions of Ubuntu use overlay scrollbars, which mean that GtkScrolledWindows no longer set a minimum height that provides enough room for a legacy scrollbar (in fact, they have a minimum height of 0). Likewise, a GtkTextView with no text to display requests a height of 0. This means that one of the scrollable GtkTextViews in gschem has been being displayed as one pixel in height, and this is obviously unusable.
In the dialog box on the right of the screenshot shown above, note the invisible widget between the "Value:" label and the "Add" button.
This has been reported independently by several users -- see also the bug report.
Question
Obviously, we could fix this by doing:
g_object_set (textview, "height-request", 100, NULL);
However, this is pretty inelegant, and will break for users who set very large font sizes in pixels (e.g. users with vision problems or who use high-DPI screens).
Ideally, therefore, we want to set the minimum size of the GtkTextView relative to the default font size, e.g. tell it to "show at least three lines of text".
Can anyone suggest a sensible/elegant approach for doing this?
Just disable the ubuntu overlay scrollbars in your application by doing:
putenv("LIBOVERLAY_SCROLLBAR=0");
Not ideal, but it's a quite good until you can find a more permanent solution. Alternatively just wait until Ubuntu disables overlay scrollbars...
I would add code to dig out the current/default style information, use that to figure out the font baseline height, and then compute some rough size allocation based on that, around three lines as you mention.
Does it have to be a textview ? If you can use an eventbox instead, then you can make a cairo surface from it, render the text with pango, and then use pango_layout_get_size() to get the text height.
Likewise, a GtkTextView with no text to display requests a height of 0.
Probably you can create GtkTextView with some text inside. Like several spaces, and set empty value after creation.
When a child window of my application is opened and I view the ALT+TAB menu, the application icon looks pixellated. I assume that Windows uses a low resolution version of the icon (16x16 pixel I think). What can I do that Windows selects the right version which would be 32x32 pixel?
I assigned an icon to the window in question that has 16x16, 24x24, 32x32, 48x38 and 256x256 in true color. Please note that VS says in the proterties that 32x32 is used and that it works fine for the main window of my application where I assigned the exact same icon.
The caption bar of a Windows application window display icons in 16x16 pixels. The Alt-Tab list, however, shows icons in 32x32 pixels. It uses the same icon as is set for the window. If you only set a 32x32 one, the caption bar has to scale it down, which tends to look ugly as it uses nearest-neighbor interpolation. So how does one support both?
The answer is the ICO file! It supports embedding multiple icons in one file, typically the same (or similar) icon in various sizes and, less commonly, color formats.
Browsers, Windows, and others are typically designed to smartly use the appropriately sized variant within a given ICO file. So the answer is to have an ICO file with both sizes (or more) inside. The result is that the caption bar correctly uses the 16x16 version and Alt-Tab uses the larger 32x32 one.
The methods for saving an ICO file with multiple icons inside varies from program to program. However, GIMP can easily do it (and it's free). The trick is to have your variously-sized icons as separate layers. When you go to save it as an ICO file, GIMP will prompt you with the ability to set the size and color format of each layer. A good tutorial, with images, can be read here.
If anyone has any links or suggestions for creating multi-icon ICO files in other programs, feel free to add them. Also, I'm unsure if the Visual Studio built-in image editor can do it or not — I've rarely bothered with it.
Fixed the problem. I put the icon on the wrong form. As I don't think that this question will be of any help for anybody feel free to delete it.