We Keep Coding

Saving DrawingVisual to an image file with a given DPI

I have got a WPF application, and I would like to save a Canvas to a file with a correct DPI value. The Canvas size is the real physical size, eg. 20x10 cm, at 300 DPI, so it's 2362x1181.
I draw images and shapes with DrawingVisual, then I create a RenderTargetBitmap. The size of the rtb is the size of the Canvas. The dpiX and dpiY are 96. I got correct image resolution only with 96 DPI. When I set it to 300, the canvas become upscale, and cropped to 2362x1181. So, it's not good. I tried to modify the canvas width and height value by dpi factor, but didn't work either.
After the RenderTargetBitmap, I use BitmapEncoder, BitmapFrame, and BinaryWriter. See code below. Working great, but the image DPI value will be 96. I've read a tons of topics about reading DPI, resaving image, using SetResolution, but I don't want to loose quality, I don't want to read file and resave it, I don't want to change pixel width/height, etc.
I just really want to save a DrawingVisual with a given DPI. I could write EXIF data for "X Resolution" (uint=282) and "Y Resoulution" (uint=283), but that didn't affect the Image DPI settings, so eg. Photoshop will read 96, not 300.
BitmapEncoder encoder = new JpegBitmapEncoder();
BitmapFrame bFrame = BitmapFrame.Create(rtb, null, meta, icc);
encoder.Frames.Add(bFrame);
using (var stream = new MemoryStream())
{
encoder.Save(stream);
byte[] imageData = stream.ToArray();
using (FileStream fs = new FileStream(path, ...)
{
BinaryWriter bw = new BinaryWriter(fs);
bw.Write(imageData);
bw.Close();
}
}

The following method saves a UIElement to a JPEG file with the specified DPI value:
public static void SaveElement(UIElement element, double dpi, string path)
{
var visual = new DrawingVisual();
var width = element.RenderSize.Width;
var height = element.RenderSize.Height;
using (var context = visual.RenderOpen())
{
context.DrawRectangle(new VisualBrush(element), null,
new Rect(0, 0, width, height));
}
var bitmap = new RenderTargetBitmap(
(int)(width * dpi / 96), (int)(height * dpi / 96),
dpi, dpi, PixelFormats.Default);
bitmap.Render(visual);
var encocer = new JpegBitmapEncoder();
encocer.Frames.Add(BitmapFrame.Create(bitmap));
using (var stream = File.OpenWrite(path))
{
encocer.Save(stream);
}
}
Alternatively, apply the DPI scaling to the DrawingVisual:
public static void SaveElement(UIElement element, double dpi, string path)
{
var visual = new DrawingVisual();
var width = element.RenderSize.Width;
var height = element.RenderSize.Height;
using (var context = visual.RenderOpen())
{
context.DrawRectangle(new VisualBrush(element), null,
new Rect(0, 0, width / dpi * 96, height / dpi * 96));
}
var bitmap = new RenderTargetBitmap(
(int)width, (int)height, dpi, dpi, PixelFormats.Default);
bitmap.Render(visual);
var encocer = new JpegBitmapEncoder();
encocer.Frames.Add(BitmapFrame.Create(bitmap));
using (var stream = File.OpenWrite(path))
{
encocer.Save(stream);
}
}

It seems the solution is the Bitmap SetResolution, after all. I tested it, and it looks like not affect the image quality after the JpegBitmapEncoder()! And the image resolution is untouched, keep the metadata, only the DPI will change!
Helped this documentation: https://learn.microsoft.com/en-us/dotnet/framework/winforms/advanced/how-to-set-jpeg-compression-level
JpegBitmapEncoder encoder = new JpegBitmapEncoder();
encoder.QualityLevel = 100;
BitmapFrame bFrame = BitmapFrame.Create(rtb, null, meta, icc);
encoder.Frames.Add(bFrame);
using (var stream = new MemoryStream())
{
encoder.Save(stream);
using (var bmpOutput = new System.Drawing.Bitmap(stream))
{
System.Drawing.Imaging.ImageCodecInfo myEncoder = GetEncoder(ImageFormat.Jpeg);
var encoderParameters = new System.Drawing.Imaging.EncoderParameters(1);
encoderParameters.Param[0] = new System.Drawing.Imaging.EncoderParameter(System.Drawing.Imaging.Encoder.Quality, 100L);
bmpOutput.SetResolution(300.0f, 300.0f);
bmpOutput.Save(filePath, myEncoder, encoderParameters);
}
}

I tried to apply DPI scaling to the DrawingVisual, but it produced less image quality. :( Also I tried to set the canvas to 756x378 (the 96 DPI size), then set the RenderTargetBitmap to 2362x1181 and 300 DPI, but that produced less image quality too. So, it seems any downscale then upscale or upscale then downscale are not the best solution. The DrawingVisual must be in the final render size.

Resizing Bitmap to increase image quality

I have an application that gets and saves a picture of an open window. Unfortunately, when the dimensions of the image and window match, the image quality is unacceptably low. I've increased the size the of the RenderTargetBitmap, which has increased the image quality, however, now the dimensions of the resulting image are too large for my purposes.
My question is this: Is there any way to resize the resulting RenderTargetBitmap to the original dimensions of the window? And is it possible to do this without a corresponding loss in image quality? Here's the code I have now.
public static RenderTargetBitmap GetReportImage(Grid view)
{
// Increased dimensions to increase image quality
Size size = new Size(view.ActualWidth * 2, view.ActualHeight * 2 );
// The dimensions that I want to convert back to.
_size = new Size(view.ActualWidth, view.ActualHeight);
if (size.IsEmpty)
return null;
RenderTargetBitmap result = new RenderTargetBitmap((int)size.Width, (int)size.Height, 96 , 96 , PixelFormats.Pbgra32);
DrawingVisual drawingVisual = new DrawingVisual();
using (DrawingContext context = drawingVisual.RenderOpen())
{
context.DrawRectangle(new VisualBrush(view), null, new Rect(new Point(), size));
context.Close();
}
result.Render(drawingVisual);
// Can I resize result to _size.Width and _size.Height?
return result;
}
Thanks in advance for any help.

To resize back you can do:
var resized = new TransformedBitmap(bmp,
new ScaleTransform(newWidth / bmp.PixelWidth, newHeight / bmp.PixelHeight));
You can also try to scale both dimensions and dpi, like this:
public static RenderTargetBitmap RenderBitmap(FrameworkElement visualToRender, double scale) {
RenderTargetBitmap bmp = new RenderTargetBitmap
(
(int) (scale*(visualToRender.ActualWidth + 1)),
(int) (scale*(visualToRender.ActualHeight + 1)),
scale*96,
scale*96,
PixelFormats.Pbgra32
);
bmp.Render(visualToRender);
return bmp;
}
Updated to answer your comment. So I suppose you need to save the result somewhere, probably to jpeg or png. Both RenderTargetBitmap and TransformedBitmap inherit from BitmapSource, so both you can encode to image like this:
var encoder = new PngBitmapEncoder();
encoder.Frames.Add(BitmapFrame.Create(bmp));
using (Stream s = File.Create("test.png")) {
encoder.Save(s);
}
Here "bmp" can be both RenderTargetBitmap and TrasnformedBitmap

PNG size growth when using RenderTargetBitmap and PngBitmapEncoder

I'm using a method to merge some tile images into a single image. But if I apply it to four 30kb- PNG images, the resulting PNG image would be 500K (~4x more than what I expect)
This is some part of the code that I'm using (suggested by Cédric Bignon):
BitmapFrame frame1 = BitmapDecoder.Create(new Uri(path1), BitmapCreateOptions.None, BitmapCacheOption.OnLoad).Frames.First();
BitmapFrame frame2 = BitmapDecoder.Create(new Uri(path2), BitmapCreateOptions.None, BitmapCacheOption.OnLoad).Frames.First();
BitmapFrame frame3 = BitmapDecoder.Create(new Uri(path3), BitmapCreateOptions.None, BitmapCacheOption.OnLoad).Frames.First();
BitmapFrame frame4 = BitmapDecoder.Create(new Uri(path4), BitmapCreateOptions.None, BitmapCacheOption.OnLoad).Frames.First();
// Gets the size of the images (I assume each image has the same size)
int imageWidth = frame1.PixelWidth;
int imageHeight = frame1.PixelHeight;
// Draws the images into a DrawingVisual component
DrawingVisual drawingVisual = new DrawingVisual();
using (DrawingContext drawingContext = drawingVisual.RenderOpen())
{
drawingContext.DrawImage(frame1, new Rect(0, 0, imageWidth, imageHeight));
drawingContext.DrawImage(frame2, new Rect(imageWidth, 0, imageWidth, imageHeight));
drawingContext.DrawImage(frame3, new Rect(0, imageHeight, imageWidth, imageHeight));
drawingContext.DrawImage(frame4, new Rect(imageWidth, imageHeight, imageWidth, imageHeight));
}
// Converts the Visual (DrawingVisual) into a BitmapSource
RenderTargetBitmap bmp = new RenderTargetBitmap(imageWidth * 2, imageHeight * 2, 96, 96, PixelFormats.Pbgra32);
bmp.Render(drawingVisual);
// Creates a PngBitmapEncoder and adds the BitmapSource to the frames of the encoder
PngBitmapEncoder encoder = new PngBitmapEncoder();
encoder.Frames.Add(BitmapFrame.Create(bmp));
// Saves the image into a file using the encoder
using (Stream stream = File.OpenWrite(pathTileImage))
encoder.Save(stream);
Anyone knows what is going on?

It's difficult to say without knowing the original images. PNG supports several color models and compressions. If, for example, the 4 original images have (different) palettes, and the composition will have to resort to a true colour format, and the total size will probably be way more than 4 times the original sizes.

Resize bitmap image

I want to have smaller size at image saved.
How can I resize it?
I use this code for redering the image:
Size size = new Size(surface.Width, surface.Height);
surface.Measure(size);
surface.Arrange(new Rect(size));
// Create a render bitmap and push the surface to it
RenderTargetBitmap renderBitmap =
new RenderTargetBitmap(
(int)size.Width,
(int)size.Height, 96d, 96d,
PixelFormats.Default);
renderBitmap.Render(surface);
BmpBitmapEncoder encoder = new BmpBitmapEncoder();
// push the rendered bitmap to it
encoder.Frames.Add(BitmapFrame.Create(renderBitmap));
// save the data to the stream
encoder.Save(outStream);

public static Bitmap ResizeImage(Bitmap imgToResize, Size size)
{
try
{
Bitmap b = new Bitmap(size.Width, size.Height);
using (Graphics g = Graphics.FromImage((Image)b))
{
g.InterpolationMode = System.Drawing.Drawing2D.InterpolationMode.HighQualityBicubic;
g.DrawImage(imgToResize, 0, 0, size.Width, size.Height);
}
return b;
}
catch
{
Console.WriteLine("Bitmap could not be resized");
return imgToResize;
}
}

The shortest way to resize a Bitmap is to pass it to a Bitmap-constructor together with the desired size (or width and height):
bitmap = new Bitmap(bitmap, width, height);

Does your "surface" visual have scaling capability? You can wrap it in a Viewbox if not, then render the Viewbox at the size you want.
When you call Measure and Arrange on the surface, you should provide the size you want the bitmap to be.
To use the Viewbox, change your code to something like the following:
Viewbox viewbox = new Viewbox();
Size desiredSize = new Size(surface.Width / 2, surface.Height / 2);
viewbox.Child = surface;
viewbox.Measure(desiredSize);
viewbox.Arrange(new Rect(desiredSize));
RenderTargetBitmap renderBitmap =
new RenderTargetBitmap(
(int)desiredSize.Width,
(int)desiredSize.Height, 96d, 96d,
PixelFormats.Default);
renderBitmap.Render(viewbox);

RenderTargetBitmap is blurry

Hi i'm creating an image in memory from a Canvas using a PngBitmapEncoder.
public void CaptureGraphic()
{
Canvas canvas = new Canvas();
canvas.SnapsToDevicePixels = true;
canvas.Height = IMAGEHEIGHT;
canvas.Width = IMAGEWIDTH;
Draw(canvas);
canvas.Arrange(new Rect(0, 0, IMAGEWIDTH, IMAGEHEIGHT));
member.MemberImage = GetPngFromUIElement(canvas);
}
public static System.Drawing.Image GetPngFromUIElement(Canvas source)
{
int width = (int)source.ActualWidth;
int height = (int)source.ActualHeight;
if (width == 0)
width = (int)source.Width;
if (height == 0)
height = (int)source.Height;
RenderTargetBitmap bitmap = new RenderTargetBitmap(width, height, 96d, 96d, PixelFormats.Pbgra32);
bitmap.Render(source);
PngBitmapEncoder enc = new PngBitmapEncoder();
enc.Interlace = PngInterlaceOption.Off;
enc.Frames.Add(BitmapFrame.Create(bitmap));
System.IO.MemoryStream ms = new System.IO.MemoryStream();
enc.Save(ms);
System.Drawing.Image image = System.Drawing.Image.FromStream(ms);
ms.Flush();
ms.Dispose();
return image;
}
Then i'm sending the image to the printer using the GDI+ DrawImage() method. However the printed result is blurry.
I've tried to match the original canvas size to the printed size to avoid any scaling, equally i've tried to make the original considerably bigger so the scaled image retains the quality however the final printed image is always blurred.
Can anyone offer any suggestions/alternatives. I have a considerable GDI+ print routine already setup and moving to wpf documents is not an option just yet.
Thanks

You're capturing the bitmap at 96 DPI. Instead of using 96 in the constructor of the RenderTargetBitmap, try to match the DPI of your printer output. Alternatively, you could do the math and calculate the difference in width/height and rescale the image on the report accordingly (the result is the image on the report will appear smaller).

I had the same blurry result and have come up with the following piece of code, which applies the same idea with the offset, but using the Offset property on the DrawingVisual (since I was using DrawDrawing, which doesn't have an overload with the offset argument):
public static Image ToBitmap(Image source)
{
var dv = new DrawingVisual();
// Blur workaround
dv.Offset = new Vector(0.5, 0.5);
using (var dc = dv.RenderOpen())
dc.DrawDrawing(((DrawingImage)source.Source).Drawing);
var bmp = new RenderTargetBitmap((int)source.Width, (int)source.Height, 96, 96, PixelFormats.Pbgra32);
bmp.Render(dv);
var bitMapImage = new Image();
bitMapImage.Source = bmp;
bitMapImage.Width = source.Width;
bitMapImage.Height = source.Height;
return bitMapImage;
}

Think i've found the answer.
http://www.charlespetzold.com/blog/2007/12/High-Resolution-Printing-of-WPF-3D-Visuals.html
I just needed to scale up the image size along with the dpi and voila, massively increased file size!

I had the same problem. To avoid blurry text and lines, I had to draw everything with an offset of 0.5 in X and Y direction. E.g, an horizontal line could be
drawingContext.DrawLine(pen, new Point(10.5,10.5), new Point(100.5,10.5));
In my case, I was rendering to a RenderTargetBitmap in a different thread to improve the UI performance. The rendered bitmap is then frozen and drawn onto the UI using
drawingContext.DrawImage(bitmap, new Rect(0.5, 0, bitmap.Width, bitmap.Height));
Here, I needed an additional offset of 0.5, but (strangely) only in X direction, so that the rendered image did not look Blurry any more.