Click here to download a Zipped copy of
(624K)

One of the least discussed parts of Xara, at least judging by it's lack of mention in the Xara Conferences, is Bitmap Effects found in the Utilities pull down menu. Those who have used older bitmap editing applications may be familiar with many of these effects. Your editor was not familiar with most of these effects and so in everyone's best interest, decided to delve into this menu item more deeply.

The Bitmap Effects section is divided into five areas, shown on the right in the Bitmap Effects fly out menu. I'll cover them in the order of appearance in the fly out menu.

Each of the menu selections opens a separate dialog which consists of a before and after window, some slider adjustments, a drop down list of bitmaps (this is the same list found in the Bitmap Gallery) and the usual buttons including context-sensitive Help which explains the contents of the dialog. Many dialogs also have a Test button which is in effect a Preview effect button. The first menu item is probably one with which most of us are familiar, Brightness and Contrast.

Brightness makes a selected bitmap image lighter or darker. You can use this to brighten a dark image. Contrast makes the darks darker and the lights lighter. 0% contrast produces a white image while 100% contrast produces a black image. Color (or Colour outside of the United States) increases or decreases the saturation in the colors. 0% color produces a grayscale image while 100% color produces an almost fluorescent appearance. Used in moderation, these three settings can improve the appearance of an overexposed, underexposed, or dull looking image.

There are eight Color Depth settings. Color depth refers to the number of bits needed to display a color. Most full color photo images use 24-bits of information to describe color, 8 bits (256 values) to describe the red, green, and blue components of a color for a grand total of 16.7 million colors in all (256 x 256 x 256 = 16.7 million give or take a few).

The first three options are 1-bit color, black or white only, and use three different dithering methods to produce the appearance of a grayscale image. (Examples of all eight settings are shown on the left in horizontal sections and all are of the same image (some of the colorful rocks found in my neighborhood in New Mexico). Grayscale uses 8 bits of color information to display an image in 256 shades of gray. 16 color Bayer and 16 color diffusion are two types of dithering that attempt to create the impression of full color (but not a very good attempt in my opinion). These two methods are reminiscent of the early days of PC graphics when computers could only display 16 colors and used these patterns to try to create the illusion of more colors. The last two methods work with a palette of 256 colors and two different methods to display a more convincing image and are the early equivalent of a 256 color GIF image. To be honest, the tools in Xara's GIF bitmap export are much more robust and do a far better job. I think this set of color depth settings, with the exception of grayscale and halftone, could be eliminated entirely and nobody would complain.

Flip and Rotate There are no surprises here. It is just as easy to produce these results on the Infobar and there is no need to dig through the menus to get there.

Resize is a very curious effect to say the least. I would expect, and I suspect we all would expect, that reducing the Height and Width by one half, as shown on the left, would produce an image that is exactly one half the size. Right?

When you press OK, however, the resulting image is the same size. So what gives?

Xara displays the image the same size while in reality the image is only half the size in terms of pixel dimensions. If you exported this image and loaded it onto a web page, the image would be the size it is shown here in pixels or half the size. Odd, or what?

Linear Interpolation is an algorithm used in determining what pixels to keep and which to discard when reducing the image. It is recommended in the context-sensitive Help menu that you use this setting as a "chunky" image may be the price to pay for not using the option. (Why is it an option, then?) Who wants a chunky image? I don't do you? Lock aspect ratio forces the image to scale in the same height to width ratio as it currently is. Not using this option lets you scale the image non-proportionately, for example, you could make the image taller and maintain the same width.

You can also use this to enlarge an image by increasing the size values. In this case the Linear Interpolation routine decides how best to arrange the extra pixels to produce the best image.

So why would you use this resizing dialog when you can do the same thing entering values on the Infobar? Beats me. The two images on the left were resized by 1/2. The image on the far left, which was resized in the Resize dialog, then reduced to the actual pixel amount, is not as sharp as the image reduced by entering 50% on the Infobar. The result is the resizing algorithm used in resizing from the Infobar, or in the bitmap export dialog, produces superior results.