We have some video software that converts BitmapSource into a BitmapImage which is then drawn to screen later using the following code. The problem is on slower machines this process seems to be a little slow, running the MS profiler it turns out that these Bitmap operations (namely the ToBitmapImage function) is in the top 4 most expensives calls we make. Is there anything I can do to improve the efficiency of this?
// Conversion
this.VideoImage = bitmapsource.ToBitmapImage();
// Drawing
drawingContext.DrawImage(this.VideoImage, new Rect(0, 0, imageWidth, imageHeight));
// Conversion code
internal static BitmapImage ToBitmapImage(this BitmapSource bitmapSource)
{
JpegBitmapEncoder encoder = new JpegBitmapEncoder()
MemoryStream memorystream = new MemoryStream();
BitmapImage tmpImage = new BitmapImage();
encoder.Frames.Add(BitmapFrame.Create(bitmapSource));
encoder.Save(memorystream);
tmpImage.BeginInit();
tmpImage.StreamSource = new MemoryStream(memorystream.ToArray());
tmpImage.EndInit();
memorystream.Close();
return tmpImage;
}
The conversion isn't necessary at all. Change the type of your VideoImage property to ImageSource or BitmapSource. Now you directly pass it as parameter to DrawImage:
this.VideoImage = bitmapsource;
drawingContext.DrawImage(this.VideoImage, new Rect(0, 0, imageWidth, imageHeight));
Related
I'am trying convert DrawingImage into MemoryStream. My code looks like this:
public MemoryStream ImageStream(DrawingImage drawingImage)
{
MemoryStream stream = new MemoryStream();
ImageSource imageSource = drawingImage;
if (imageSource != null)
{
BitmapSource bitmap = imageSource as BitmapSource;
if (bitmap != null)
{
BitmapEncoder encoder = new PngBitmapEncoder();
encoder.Frames.Add(BitmapFrame.Create(bitmap));
encoder.Save(stream);
}
}
return stream;
}
But problem is after casting ImageSource into BitmapSource bitmap is always null. Any sugestion how to fix that?
The reason your bitmap variable is always null is because DrawingImage does not extend BitmapImage or vice-versa, so the cast is guaranteed to fail. A DrawingImage does not contain any pixel data of any kind. It references a Drawing that is used whenever the image needs to be rasterized.
How did you find yourself in a situation where you want to rasterize a DrawingImage and serialize it into a stream? I get the feeling you are going about something in an unusual way if you have need of a function like this.
Nevertheless, you could implement this function by drawing the DrawingImage to a DrawingVisual, rendering it to a RenderTargetBitmap, and then passing the render target to the encoder to serialize the raster data to a stream.
public MemoryStream ImageStream(DrawingImage drawingImage)
{
DrawingVisual visual = new DrawingVisual();
using (DrawingContext dc = visual.RenderOpen())
{
dc.DrawDrawing(drawingImage.Drawing);
dc.Close();
}
RenderTargetBitmap target = new RenderTargetBitmap((int)visual.Drawing.Bounds.Right, (int)visual.Drawing.Bounds.Bottom, 96.0, 96.0, PixelFormats.Pbgra32);
target.Render(visual);
MemoryStream stream = new MemoryStream();
BitmapEncoder encoder = new PngBitmapEncoder();
encoder.Frames.Add(BitmapFrame.Create(target));
encoder.Save(stream);
return stream;
}
If you want something a little more generic, I would split this into two methods and change some of the types.
public BitmapSource Rasterize(Drawing drawing)
{
DrawingVisual visual = new DrawingVisual();
using (DrawingContext dc = visual.RenderOpen())
{
dc.DrawDrawing(drawing);
dc.Close();
}
RenderTargetBitmap target = new RenderTargetBitmap((int)drawing.Bounds.Right, (int)drawing.Bounds.Bottom, 96.0, 96.0, PixelFormats.Pbgra32);
target.Render(visual);
return target;
}
public void SavePng(BitmapSource source, Stream target)
{
BitmapEncoder encoder = new PngBitmapEncoder();
encoder.Frames.Add(BitmapFrame.Create(source));
encoder.Save(target);
}
Then you could use it with any kind of stream. For example, to save the drawing to a file:
using (FileStream file = File.Create("somepath.png"))
{
SavePng(Rasterize(drawingImage.Drawing), file);
}
I'm building a small app using C#/WPF.
The application receives (from an unmanaged C++ library) a byte array (byte[]) from a bitmap source
In my WPF window, I have an (System.windows.Controls.Image) image which I will use to display the bitmap.
In the code behind (C#) I need to able to take that byte array, create BitmapSource /ImageSource and assign the source for my image control.
// byte array source from unmanaged librariy
byte[] imageData;
// Image Control Definition
System.Windows.Controls.Image image = new Image() {width = 100, height = 100 };
// Assign the Image Source
image.Source = ConvertByteArrayToImageSource(imageData);
private BitmapSource ConvertByteArrayToImagesource(byte[] imageData)
{
??????????
}
I've been working on this for a bit here and haven't been able to figure this out. I've tried several solutions that I've found by goolging around. To date, I haven't been able to figure this out.
I've tried:
1) Creating a BitmapSource
var stride = ((width * PixelFormats.Bgr24 +31) ?32) *4);
var imageSrc = BitmapSource.Create(width, height, 96d, 96d, PixelFormats.Bgr24, null, imageData, stride);
That through a runtime exception saying buffer was too small
Buffer size is not sufficient
2) I tried using a memory stream:
BitmapImage bitmapImage = new BitmapImage();
using (var mem = new MemoryStream(imageData))
{
bitmapImage.BeginInit();
bitmapImage.CrateOptions = BitmapCreateOptions.PreservePixelFormat;
bitmapImage.CacheOption = BitmapCacheOption.OnLoad;
bitmapImage.StreamSource = mem;
bitmapImage.EndInit();
return bitmapImage;
}
This code through an exception on the EndInit() call.
"No imaging component suitableto complete this operation was found."
SOS! I've spent a couple of days on this one and am clearly stuck.
Any help/ideas/direction would be greatly appreciated.
Thanks,
JohnB
Your stride calculation is wrong. It is the number of full bytes per scan line, and should therefore be calculated like this:
var format = PixelFormats.Bgr24;
var stride = (width * format.BitsPerPixel + 7) / 8;
var imageSrc = BitmapSource.Create(
width, height, 96d, 96d, format, null, imageData, stride);
Of course you also have to make sure that you use the correct image size, i.e. that the width and height values actually correspond with the data in imageBuffer.
I'm loading a BitmapImage from a BitmapSource and i always find the BitmapImage's format is Bgr32 instead of Bgra32 which the BitmapSource is. There is a similar thread on this here:
BitmapImage from file PixelFormat is always bgr32
but using BitmapCreateOptions.PreservePixelFormat didn't work for me as was suggested in the thread. here's what i'm doing:
// I have verified that b is a valid BitmapSource and its format is Bgra32
// the following code produces a file (testbmp2.bmp) with an alpha channel as expected
// placing a breakpoint here and querying b.Format in the Immediate window also produces
// a format of Bgra32
BmpBitmapEncoder test = new BmpBitmapEncoder();
FileStream stest = new FileStream(#"c:\temp\testbmp2.bmp", FileMode.Create);
test.Frames.Add(BitmapFrame.Create(b));
test.Save(stest);
stest.Close();
// however, this following snippet results in bmp.Format being Bgr32
BitmapImage bmp = new BitmapImage();
BmpBitmapEncoder encoder = new BmpBitmapEncoder();
MemoryStream stream = new MemoryStream();
encoder.Frames.Add(BitmapFrame.Create(b));
encoder.Save(stream);
bmp.BeginInit();
bmp.StreamSource = new MemoryStream(stream.ToArray());
bmp.CreateOptions = BitmapCreateOptions.PreservePixelFormat;
bmp.CacheOption = BitmapCacheOption.None;
bmp.EndInit();
stream.Close();
Is there a way to create a BitmapImage from the BitmapSource and preserver the alpha channel?
Update:
I used the same code posted in the the other thread to load testbmp2.bmp from file and after loading, the Format property is Bgr32.
BitmapImage b1 = new BitmapImage();
b1.BeginInit();
b1.UriSource = new Uri(#"c:\temp\testbmp2.bmp");
b1.CreateOptions = BitmapCreateOptions.PreservePixelFormat | BitmapCreateOptions.IgnoreImageCache;
b1.CacheOption = BitmapCacheOption.OnLoad;
b1.EndInit();
so maybe i'm not saving the FileStream/MemoryStream correctly? That doesn't seem right since after saving the FileStream, I can open testbmp2.bmp in Photoshop and see the alpha channel.
Update 2:
I think I need to re-word the issue I'm having. I'm trying to display separate channels of a texture. I am displaying the bitmap via an Image control. I have a simple HLSL pre compiled shader assigned to the Image's Effect property that will mask out the channels based user input. Not being able to get the BitmapSource into a BitmapImage while retaining the alpha channel was only part of the problem. I realize now that since my original BitmapSource's Format was Bgra32 I could assign that directly to the Image's Source property. The problem appears to be that an Image object will only display texels with pre-multiplied alpha...? Here's my shader code:
sampler2D inputImage : register(s0);
float4 channelMasks : register(c0);
float4 main (float2 uv : TEXCOORD) : COLOR0
{
float4 outCol = tex2D(inputImage, uv);
if (!any(channelMasks.rgb - float3(1, 0, 0)))
{
outCol.rgb = float3(outCol.r, outCol.r, outCol.r);
}
else if (!any(channelMasks.rgb - float3(0, 1, 0)))
{
outCol.rgb = float3(outCol.g, outCol.g, outCol.g);
}
else if (!any(channelMasks.rgb - float3(0, 0, 1)))
{
outCol.rgb = float3(outCol.b, outCol.b, outCol.b);
}
else
{
outCol *= channelMasks;
}
if (channelMasks.a == 1.0)
{
outCol.r = outCol.a; // * 0.5 + 0.5;
outCol.g = outCol.a;
outCol.b = outCol.a;
}
outCol.a = 1;
return outCol;
}
I've tested it pretty extensively and I'm pretty sure that its setting the outCol value properly. When i pass in a channel mask value of (1.0, 1.0, 1.0, 0.0) the image displayed is the RGB channels with the areas of the alpha that are black drawing as black - as you would expect if WPF was pre-multiplying the alpha on to the RGB channels. Does anyone know of any way to dispaly a BitmapSource in an Image without pre multiplied alpha? Or more to the point, is there a way to have the effect receive the texture before the alpha pre-multiply happens? I'm not sure what order things are done in, but apparently the pre-multiply happens before the texture is written to the s0 register and the shader gets to work on it. I could try to do this with WriteableBitmap and copying the alpha out to another BitmapSource but that would all be using software rendering i think...?
I ran into a similar problem lately. The BmpBitmapEncoder does not care about the alpha channel and therefore does not preserve it. An easy fix is to use a PngBitmapEncoder instead.
I'm working with image files over 1 GB, creating a Bitmap from a large BitmapSource and attempting to dispose of the original BitmapSource. The BitmapSource stays in memory. Normally this is an inconvenience as it is eventually collected, but with these large files, clearing the memory immediately is a necessity:
System.Drawing.Bitmap bitmap;
using (MemoryStream outStream = new MemoryStream())
{
BitmapEncoder enc = new BmpBitmapEncoder();
enc.Frames.Add(BitmapFrame.Create(bitmapsource)); //large image
enc.Save(outStream);
bitmapsource = null;
Bitmap temp = new System.Drawing.Bitmap(outStream);
bitmap = temp.Clone(new Rectangle(0, 0, bm.Width, bm.Height), bm.PixelFormat);
temp.Dispose();
}
GC.Collect();
GC.Collect();
//both bitmap and bitmapsource remain in memory
I've come across a work-around here, but the BitmapSource still persists until much later in the pipeline.
I've been trying to Serialize and Deserialize BitmapImages. I've been using methods which supposedly works which I found in this thread: error in my byte[] to WPF BitmapImage conversion?
Just to iterate what is going on, here is part of my Serialization code:
using (MemoryStream ms = new MemoryStream())
{
// This is a BitmapImage fetched from a dictionary.
BitmapImage image = kvp.Value;
PngBitmapEncoder encoder = new PngBitmapEncoder();
encoder.Frames.Add(BitmapFrame.Create(image));
encoder.Save(ms);
byte[] buffer = ms.GetBuffer();
// Here I'm adding the byte[] array to SerializationInfo
info.AddValue((int)kvp.Key + "", buffer);
}
And here is the deserialization code:
// While iterating over SerializationInfo in the deserialization
// constructor I pull the byte[] array out of an
// SerializationEntry
using (MemoryStream ms = new MemoryStream(entry.Value as byte[]))
{
ms.Position = 0;
BitmapImage image = new BitmapImage();
image.BeginInit();
image.StreamSource = ms;
image.EndInit();
// Adding the timeframe-key and image back into the dictionary
CapturedTrades.Add(timeframe, image);
}
Also, I'm not sure if it matters but earlier when I populated my dictionary I encoded Bitmaps with PngBitmapEncoder to get them into BitmapImages. So not sure if double-encoding has something to do with it. Here's the method that does that:
// Just to clarify this is done before the BitmapImages are added to the
// dictionary that they are stored in above.
private BitmapImage BitmapConverter(Bitmap image)
{
using (MemoryStream ms = new MemoryStream())
{
image.Save(ms, System.Drawing.Imaging.ImageFormat.Png);
BitmapImage bImg = new BitmapImage();
bImg.BeginInit();
bImg.StreamSource = new MemoryStream(ms.ToArray());
bImg.EndInit();
ms.Close();
return bImg;
}
}
So the problem is, serialization and deserialization works fine. No errors, and the dictionary has entries with what seems to be BitmapImages, however their width/height and
some other properties are all set to '0' when I look at them in debugging-mode. And of course, nothing is shown when I try to display the images.
So any ideas as to why they aren't properly deserialized?
Thanks!
1) You should not dispose MemoryStream, used from image initializing. Remove using in this line
using (MemoryStream ms = new MemoryStream(entry.Value as byte[]))
2) After
encoder.Save(ms);
Try adding
ms.Seek(SeekOrigin.Begin, 0);
ms.ToArray();