Re-writing this SKBitmap code with C# Pointers rather than using SKBitmap.Pixels?

[jonmdev]

I am using the following SkiaSharp code to apply a multiply effect (tint) from a given tintColor to the image and create a new SKBitmap with the applied tint:

                SKColor[] pixels = skBitmap.Pixels;
                for (int i=0; i < pixels.Length; i++) {
                    pixels[i] = new SKColor((byte)(pixels[i].Red * tintColor.Red), (byte)(pixels[i].Green * tintColor.Green), (byte)(pixels[i].Blue * tintColor.Blue), pixels[i].Alpha);
                }
                tintedBitmap = new SKBitmap(skBitmap.Width, skBitmap.Height);
                tintedBitmap.Pixels = pixels;

This is working fine, but it seems evident to me from the documentation here that there is some inefficiency.

In particular, the line SKColor[] pixels = skBitmap.Pixels copies the entire pixel data over into this array which is not necessary. I could alternatively create an empty array like SKColor[] pixels = new SKColor[skBitmap.Width * skBitmap.Height] and then fill this in the for loop by accessing the skBitmap color data with pointers.

However, I am having a hard time understanding how to use pointers to do this. I read the documentation linked above but the C# pointers and byte management of the colors is a bit above my head.

What might the for loop in this example look like if using pointers to reference the color data from skBitmap instead?

Thanks for any help.

[mattleibow]

You can try using the pixel data directly.

This is much easier through the SKPixmap type, and you can get that with skBitmap.PeekPixels() and then with SKPixmap.GetPixelSpan<SKColor>().

One thing to not is that you will need to verify if this is RGB or BGR pixel ordering. You can do this by checking the color type.