Question

I’m developing a game client in which texture colors are stored as int values. Sometimes, however, I have to use 16bit colors to draw some texture… so I would like to know how to convert 32bit color to 16bit color and vice versa… preferably avoiding managed functions like Color.FromArgb() and similar. I prefer to achieve those operations as fast as possible with byte shifting.
Do you also know how to get the grayscale value of a 32 bit color?

Answer 1

    public static Int32 16To32(UInt16 color)
    {
        Int32 red = (Int32)(((color >> 0xA) & 0x1F) * 8.225806f);
        Int32 green = (Int32)(((color >> 0x5) & 0x1F) * 8.225806f);
        Int32 blue = (Int32)((color & 0x1F) * 8.225806f);

        if (red < 0)
            red = 0;
        else if (red > 0xFF)
            red = 0xFF;

        if (green < 0)
            green = 0;
        else if (green > 0xFF)
            green = 255;

        if (blue < 0)
            blue = 0;
        else if (blue > 0xFF)
            blue = 0xFF;

        return ((red << 0x10) | (green << 0x8) | blue);
    }

    public static UInt16 32To16(Int32 color)
    {
        Int32 red = ((((color >> 0x10) & 0xFF) * 0x1F) + 0x7F) / 0xFF;
        Int32 green = ((((color >> 0x8) & 0xFF) * 0x1F) + 0x7F) / 0xFF;
        Int32 blue = (((color & 0xFF) * 0x1F) + 0x7F) / 0xFF;

        return (UInt16)(0x8000 | (red << 0xA) | (green << 0x5) | blue);
    }

For what concerns grayscaling… you could try the following function but I just wrote it without testing so I cannot guarantee it will work perfectly:

    public static Int32 Grayscale(Int32 color)
    {
        Single red = ((color >> 0xA) & 0x1F) * 8.225806f;
        Single green = ((color >> 0x5) & 0x1F) * 8.225806f;
        Single blue = (color & 0x1F) * 8.225806f;

        Int32 grayscale = (Int32)(((red * 0.299f) + (green * 0.587f) + (blue * 0.114f)) * 0.1215686f);

        if (grayscale < 0)
            grayscale = 0;
        else if (grayscale > 0x1F)
            grayscale = 0x1F;

        return grayscale;
    }