Develop Reference

Rotate writeable bitmap without creating a new bitmap

Ok so here's how this works:
camera sends me an image as a writeable bitmap
client.WhenMasterFrameCaptured().ObserveOn(SynchronizationContext.Current).Subscribe(frame => UpdateMasterCameraPreview(frame));
I set that pointer to an image source(wpf Image control) in the method that this is subscribed to
if (cameraOneBitmap != null)
{
frame.Image.CopyTo(cameraOneBitmap);
}
else
{
cameraOneBitmap = frame.Image.ToWriteableBitmap();
}
cameraOneBitmap.Lock();
cameraOneBitmap.AddDirtyRect(new Int32Rect(0, 0, cameraOneBitmap.PixelWidth, cameraOneBitmap.PixelHeight));
cameraOneBitmap.Unlock();
cameraOneImage.Source = cameraOneBitmap;
The Problem: The method feeds me a bitmap that isn't the correct angle and i have no way of changing that. I must rotate it. I'm currently using a render transform in xaml. the problem is it shapes the image all weird in certain window sizes it hangs over the grid into the bottom of the window so i don't want to use that. I've tried using writeablebitmapex library but it creates a new bitmap and that gives me an out of memory exception in about 10 seconds... Is there a way to rotate this bitmap with out have to make a new one every time? Somehow the render transform does it without giving me problems. I was doing this to set the source but it still gives me an exception because it makes a new one.
public static RenderTargetBitmap RotateImage(double angle, WriteableBitmap sourceBitmap)
{
TransformedBitmap tb = new TransformedBitmap(sourceBitmap, new RotateTransform(angle));
DrawingVisual drawingVisual = new DrawingVisual();
DrawingContext drawingContext = drawingVisual.RenderOpen();
drawingContext.DrawImage(tb, new Rect(0, 0, tb.PixelWidth, tb.PixelHeight));
//drawingContext.PushTransform(new RotateTransform(270, .5, .5));
drawingContext.Close();
System.Windows.Media.Imaging.RenderTargetBitmap bmp = new System.Windows.Media.Imaging.RenderTargetBitmap(tb.PixelWidth, tb.PixelHeight, 96, 96, PixelFormats.Pbgra32);
bmp.Render(drawingVisual);
return bmp;
}

image quality is blurry

I have a situation where I want to convert some XAML to an image, so I created a RichTextBox and then took the image of it. Now problem is that words in image is blurred, any idea how I might be able to fix it?
public System.Drawing.Bitmap ConvertXamltoImage(string XamlString, int Width, int Height)
{
RichTextBox AdContentRichTextBox = new RichTextBox() { Width = Width, Height = Height };
AdContentRichTextBox.BorderThickness = new Thickness(0);
XmlReader _XmlReader = XmlReader.Create(new StringReader(XamlString));
AdContentRichTextBox.Document = XamlString;
var size = new Size(Width, Height);
AdContentRichTextBox.Measure(size);
AdContentRichTextBox.Arrange(new Rect(size));
RenderTargetBitmap bmp = new RenderTargetBitmap(Width, Height, 300, 300, PixelFormats.Pbgra32);
bmp.Render(AdContentRichTextBox);
DrawingVisual _drawingVisual = new DrawingVisual();
using (DrawingContext _drwaingContext = _drawingVisual.RenderOpen())
{
VisualBrush _visualBrush = new VisualBrush(AdContentRichTextBox);
}
PngBitmapEncoder _png = new PngBitmapEncoder();
_png.Frames.Add(BitmapFrame.Create(bmp));
System.Drawing.Bitmap _tempBitmap = null;
using (Stream _fileStream = new MemoryStream())
{
_png.Save(_fileStream);
_tempBitmap = new System.Drawing.Bitmap(_fileStream);
_fileStream.Flush();
}
return _tempBitmap;
}

Hmmmm..there could be lots of things all interacting here:
1st
"Grayscale fall back - if ClearType is disabled or one is rendering text in certain situations where the ClearType algorithm cannot be run, WPF will use a grayscale rendering algorithm to antialias the rendered text."
Rendering Text to a RenderTargetBitmap seems to be one of those situations....(the renderer switches from a hardware to a software path).
2nd
In addition NET 4 switched the default scaling algorithm from high-quality (Fant) to low-quality (Bi-Linear).....now that shouldn't come into play here as it doesn't look like you are scaling the bitmap in any way...but you never know what's going on inside. It's possible to switch the scaler back to the higher quality one.
http://www.olsonsoft.com/blogs/stefanolson/post/Workaround-for-low-quality-bitmap-resizing-in-WPF-4.aspx
3rd
You may need to take into account the parent container of the RichTextBox...see last link below, mentions it can distort the font rendering.
Problems with rendering text as bitmaps using WPF
Some ideas on how to work around this are:
render the RichTextBox at a higher resolution e.g. 600dpi, and then scale down the bitmap (probably will make no difference)
capture the screen....difficult or not practical if your visual is offscreen/obscured, etc.
See related links:
http://windowsclient.net/wpf/white-papers/wpftextclarity.aspx
WPF RenderTargetBitmap downscaling text ClearType to GreyScale
WPF RenderTargetBitmap downscaling TextRenderMode to GreyScale
WPF text rendering inconsistencies

Print a WPF control to an Image control

I'm writing an application for the surface that requires displaying data in a table (i.e. DataGrid). This is great, except the table captures the touch interactions for the ScatterViewItem control (basically a panel that can be spun, shrunk, and moved by the user). This prevents the user from easily manipulating the ScatterViewItem.
To solve this problem, I thought it would be easy to draw the control to an image and just put that up. It seems I was wrong. Here are all my attempts:
http://pastie.org/private/gfkkv9f6apgrqi1ucspwpa (no need to read this, unless you think it will be useful. That's why it's in pastie and not on here)
I'm putting the DataGrid inside of another Grid, because otherwise it won't measure properly:
Grid g = new Grid();
g.Children.Add(dataTable);
SurfaceScrollViewer viewer = new SurfaceScrollViewer();
viewer.Content = Utility.SaveWPFControlAsImage(g);
If we change that last line to
viewer.Content = g;
We get a good table:
If we don't, we get:
SaveWPFControlAsImage is as follows:
public static System.Windows.Controls.Image SaveWPFControlAsImage(FrameworkElement e)
{
e.Measure(new System.Windows.Size(double.PositiveInfinity, double.PositiveInfinity));
RenderTargetBitmap targetBitmap =
new RenderTargetBitmap((int)e.DesiredSize.Width,
(int)e.DesiredSize.Height,
96d, 96d,
PixelFormats.Default);
targetBitmap.Render(e);
BmpBitmapEncoder encoder = new BmpBitmapEncoder();
encoder.Frames.Add(BitmapFrame.Create(targetBitmap));
MemoryStream stream = new MemoryStream();
encoder.Save(stream);
BitmapImage bmp = new BitmapImage();
bmp.BeginInit();
bmp.StreamSource = new MemoryStream(stream.ToArray());
bmp.EndInit();
return new System.Windows.Controls.Image()
{
Source = bmp,
};
}
So maybe I'm just not rendering it right, or, maybe, I'm just going about this at the wrong angle...

In WPF you have a VisualBrush which allows you to capture a live preview of a given control. Also if you don't want any input of a given control, you can always set IsHitTestVisible="False".

Capture each WPF MediaElement frame

Is there a way to capture each WPF MediaElement frame? Like an event that fires at each rendered frame and allows me to access it. If MediaElement does not provide such functionality, how could it be implemented or what other control could I use? On a side note, is there such a control or method that would allow for off-screen fast rendering of media clips with frame capture? (so I could process frames as fast as possible)

Try out my WPF MediaKit project. Allows you to do pretty much anything in WPF with Media. Try out the MediaDetector.cs, it allows you to extract out frames from any time in the media. It's a little buggy as I've never put a lot of time in it, but should work for what you need.

There is no built-in WPF way:
MediaElement does not have this ability.
BitmapDecoder has the API to request this but using BitmapDecoder to extract frames from arbitrary media is not implemented: It can only extract frames from a few animated bitmap formats like .gif.
I was able to get frame images from .mpg, .wmv, .mov, .flv, .avi and other movie formats using DirectShow. I constructed a filter graph using DirectShow's COM graph builder interfaces. The resulting filter graph decoded the movie and connected it to a custom renderer filter written in C#. My custom filter received the frame data and converted it into BitmapSource objects for display using BitmapSource.Create.
The DirectShow solution performed quite well, and the managed to unmanaged transition was no big deal, but it took a while to figure out the details of the DirectShow graph building.

If you use your imagination perhaps this snippet can give you some ideas:
MediaPlayer player = new MediaPlayer();
player.Open(new Uri(_inputFilename));
player.ScrubbingEnabled = true;
DrawingVisual dv = new DrawingVisual();
for (int i = 0; i < session.FramesList.Count; i++)
{
Frame f = session.FramesList[i];
player.Position = new TimeSpan((long)(f.Time * 10000000));
using (DrawingContext dc = dv.RenderOpen())
{
dc.DrawVideo(player, new Rect(0, 0, 1024, 576));
}
RenderTargetBitmap bmp = new RenderTargetBitmap(1024, 576, 96, 96, PixelFormats.Pbgra32);
bmp.Render(dv);
f.Thumbnail = bmp.GetAsFrozen() as ImageSource;
framesListView.Dispatcher.Invoke(() => FramesList.Add(f));
}

you can try this method to output any UI Element including the MediaElement in WPF
public static void ConvertUiElementToBitmap(UIElement elt, string path)
{
double h = elt.RenderSize.Height;
double w = elt.RenderSize.Width;
if (h > 0)
{
PresentationSource source = PresentationSource.FromVisual(elt);
RenderTargetBitmap rtb = new RenderTargetBitmap((int)w, (int)h, 96, 96, PixelFormats.Default);
VisualBrush sourceBrush = new VisualBrush(elt);
DrawingVisual drawingVisual = new DrawingVisual();
DrawingContext drawingContext = drawingVisual.RenderOpen();
using (drawingContext)
{
drawingContext.DrawRectangle(sourceBrush, null, new Rect(new Point(0, 0),
new Point(w, h)));
}
rtb.Render(drawingVisual);
// return rtb;
var encoder = new PngBitmapEncoder();
var outputFrame = BitmapFrame.Create(rtb);
encoder.Frames.Add(outputFrame);
using (var file = System.IO.File.OpenWrite(path))
{
encoder.Save(file);
}
}
}

WPF Screenshot JPG from UIElement with C#

I'm trying to create a JPG from part of my WPF Applications. Like a screenshot, only of individual UIElements. I started here: http://www.grumpydev.com/2009/01/03/taking-wpf-screenshots/
I am using his extension method, which essential allows you to get a byte[] with UIElement.GetJpgImage(). This can then be written using a filestream to a JPG image. If I make a JPG of the whole window, it looks just fine! However, this is not ideal because it just captures what the user sees. Things that are not visible because of the scrollviewer or because their parent was animated to a small size won't show up.
If I take a "screenshot" of, say, a grid that I use for layout:
alt text http://img697.imageshack.us/img697/4233/fullscreenshot2.jpg
I get this crap with a black background. I don't want that. Furthermore, if I've collapsed this grid's height using animation, I won't get anything at all. Those are actually templated checkboxes, they should have black text above them, and the background of the grid should be white. Here's the code that someone else wrote to return the byte[] array that gets written to a filestream:
public static byte[] GetJpgImage(this UIElement source, double scale, int quality)
{
double actualHeight = source.RenderSize.Height;
double actualWidth = source.RenderSize.Width;
double renderHeight = actualHeight * scale;
double renderWidth = actualWidth * scale;
RenderTargetBitmap renderTarget = new RenderTargetBitmap((int) renderWidth, (int) renderHeight, 96, 96, PixelFormats.Pbgra32);
VisualBrush sourceBrush = new VisualBrush(source);
DrawingVisual drawingVisual = new DrawingVisual();
DrawingContext drawingContext = drawingVisual.RenderOpen();
using (drawingContext)
{
drawingContext.PushTransform(new ScaleTransform(scale, scale));
drawingContext.DrawRectangle(sourceBrush, null, new Rect(new Point(0, 0), new Point(actualWidth, actualHeight)));
}
renderTarget.Render(drawingVisual);
JpegBitmapEncoder jpgEncoder = new JpegBitmapEncoder();
jpgEncoder.QualityLevel = quality;
jpgEncoder.Frames.Add(BitmapFrame.Create(renderTarget));
Byte[] _imageArray;
using (MemoryStream outputStream = new MemoryStream())
{
jpgEncoder.Save(outputStream);
_imageArray = outputStream.ToArray();
}
return _imageArray;
}
Somewhere in there, we're getting a black background. Any insight?
EDIT: If I set the grid's background property to white, the screenshot comes out as expected. However, it's not feasible to set everything's background that I need to take a screenshot of.

Just a guess, I would think that a black background would represent portions of the byte array that are not set to anything in this process. The initial zeros in the array would appear as black.
To avoid this, I suggest initializing the array with 0xFF (byte.MaxValue) values.
UPDATED:
From looking at this closer, I think you should draw a white rectangle onto the image before you render the UI element. That ought to work anyway.
Just before this line of code
drawingContext.DrawRectangle(sourceBrush, null, new Rect(new Point(0, 0), new Point(actualWidth, actualHeight)));
put something like this
drawingContext.DrawRectangle(Brushes.White, null, new Rect(new Point(0, 0), new Point(actualWidth, actualHeight)));

Unfortunately the only thing that worked was just setting the element's background in the XAML. I didn't want to do this, but I guess it is what I need to do in this case. Thanks anyway for the suggestion.