So I'm trying to capture a region using SlimDX. What I mean by this is that I'm not looking to capture from [0,0]
to [1920,1080]
.
Preferably I want to pass a Rectangle object that holds the information needed for the capturing.
I would like to do this with SlimDX (DirectX) as it would drastically improve the capturing time if we look at alternatives like CopyFromScreen
.
I need to be capturing around 30 chunks of 100x100 pixels and I think using DirectX might be my best bet. All starting coördinates plus width/height are stored in integer arrays.
I'm currently using the following code:
Rectangle rect = new Rectangle(chunk[0], chunk[1], chunk[2], chunk[3]);
Bitmap temp = new Bitmap(rect.Width, rect.Height, PixelFormat.Format32bppArgb);
Graphics g2 = Graphics.FromImage(temp);
g2.CopyFromScreen(rect.Left, rect.Top, 0, 0, temp.Size, CopyPixelOperation.SourceCopy);
This code is living in a foreach loop that iterates over the chunks
variable. However, this takes about ~500ms to complete with 30 iterations.
as @Nico pointed out, copy once and then split is faster than many small copies. Here is an example
var rects = Enumerable.Range(1, 30)
.Select(x => new Rectangle(x, x, x + 100, x + 100));
var bounds = Screen.PrimaryScreen.Bounds;
Bitmap bigBmp = new Bitmap(bounds.Width, bounds.Height, PixelFormat.Format32bppArgb);
Graphics g2 = Graphics.FromImage(bigBmp);
g2.CopyFromScreen(bounds.X, bounds.Y, 0, 0, bounds.Size, CopyPixelOperation.SourceCopy);
var bmps = rects.Select(rect =>
{
return bigBmp.Clone(rect, PixelFormat.Format32bppArgb);
});