asked on

Direct2D, Windows 7 and C#

If I want better performance in doing a simple annotation, is it accurate to say that I should be using Direct2D instead of GDI+ under Windows 7?

I am writing a C# console application. At one point, it is currently using System.Drawing.Graphics (which is GDI+ essentially) to do the annotation, like this:

   using (Graphics g = Graphics.FromImage(bm))
            using (SolidBrush solidWhiteBrush = new SolidBrush(Color.White))
            using (SolidBrush solidBlackBrush = new SolidBrush(Color.Black))
            {
                // Write out annotation text
                g.FillRectangle(solidWhiteBrush, 0, 0, bm.Width, bm.Height);
                g.DrawString(AnnotationText, AnnotationFont, solidBlackBrush, new PointF(7f, 2f));

...

...

Open in new window

Is there a faster way to be doing this?

What would be the "Direct2D" way to write these 2 lines of code, specifically:

g.FillRectangle(solidWhiteBrush, 0, 0, bm.Width, bm.Height);
g.DrawString(AnnotationText, AnnotationFont, solidBlackBrush, new PointF(7f, 2f));

And what libraries do I need to include to get to the Direct2D methods?

Thank you!

ASKER CERTIFIED SOLUTION

Bob Learned

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

Tom Knowlton

ASKER

I am abandoning this strategy.

Direct2D is too low level for our purposes.

Bob Learned

May I ask about your "purposes"? What are your requirements?

Tom Knowlton

ASKER

>>>May I ask about your "purposes"? What are your requirements?

-It is a console app, so no winforms or GUI of any kind.
-Said console app is to take a series of TIFF images and annotate them (add text to the upper left corner of the TIFF image.

Here is the current code, which I want to optimize:

public MemoryStream Annotate(MemoryStream InputImage, string AnnotationText, Font AnnotationFont)
        {
            InputImage.Position = 0;
            
            Bitmap bImagem = new Bitmap(InputImage);
            PixImage pImage = new PixImage();
            pImage.Import(bImagem);
            PixTools.Annotation.AnnotationDocument annotationDoc = new PixTools.Annotation.AnnotationDocument();
            PixTools.Annotation.TextAnnotation textItem = new PixTools.Annotation.TextAnnotation();
            textItem.Text = new PixTools.Annotation.ItemText(AnnotationText);
            annotationDoc.Items.Add(textItem);
            Console.WriteLine("LEAD TOOLS ANNOTATE - no print top");
            PixImage newpi = pImage.MergeAnnotations(annotationDoc, MergeAttributes.MergeColorFormatPreserve, null);
            
            Bitmap anotherBitmap = newpi.GetBitmap();

            MemoryStream ms = new MemoryStream();
            
            anotherBitmap.Save(ms, ImageFormat.Tiff);

            ms.Position = 0;

            return ms;      

        }

Open in new window

Having to create a Bitmap from the MemoryStream (and at the end, the reverse happens) seems too "heavy" and inefficient.

Bob Learned

How many images would you need to annotate during a single run?

Tom Knowlton

ASKER

>>>How many images would you need to annotate during a single run?

This is not known until runtime. Typically, 4 to 6 images.

Bob Learned

I was thinking that you needed to process 1000's of files, which would require careful attention to detail. 4 to 6 doesn't sound like you have a big need for optimization. How fast is this running currently?

Tom Knowlton

ASKER

It is taking about 1 second per file.

But I should mention - we are in fact in a scalable environment.

The code should be written as if 1,000's of file might be processed.

Let's proceed as if that were true....if that is agreeable. :)

Bob Learned

Have you profiled this to see what takes the most time in the 1 second?