Tips & Tricks

Resolving pale "halo" when printing semi-transparent objects from CorelDRAW

If you worked with CorelDRAW's dropshadows or objects with transparencies you may have probably noticed a problem printing those objects on composite printers (such as laser printers, inkjet printers and others). Even if everything looks perfect on the screen, you still can't get your artwork printed correctly:

halo1.jpg (4967 bytes)
Original image as seen on screen
in CorelDRAW
arrow.gif (307 bytes) halo2.jpg (5471 bytes)
The same document printed to
an inkjet printer

Why does this happen and how to cure it? Let's try to find this out.

The way CorelDRAW treats transparencies

In order to reproduce a transparent object CorelDRAW creates a bitmap image from the object and everything that lies directly underneath:

halo3.gif (6508 bytes)

As a result, we see the dropshadow bitmap bordering on the underlying vector object(s).

Color space mapping

Exterior areas of the dropshadow bitmap has exactly the same color values as the vector object beneath it. Why do we see differences on printout, then? The answer to this question is buried deeply into color representation theory.

As we know, different output devices have different color reproduction capabilities. Monitors emit light using three color phosphors - red, green and blue (thus, representing RGB color space). Most of printing devices uses other technologies, such as subtractive color inks (cyan, magenta, yellow, black - CMYK color space) and so on. Obviously, devices built with completely different techniques have different properties and capabilities. Modern RGB color monitor can display wider range of colors than CMYK printers can render. Because of this we are facing a problem of color transfer between different color models.

There are two ways to translate colors from one color space to another:

  1. Replace colors falling behind the output device's capabilities with the closest ones that can be displayed by the device. This method reproduces colors without alterations unless they are out of gamut of output devices.
  2. Scale all colors in the source image to fit the output color range. This changes all colors in image but keeps relative color differences between colors along the whole source color range. The advantage of this method is to keep shades of continuous-tone images such as photographs.

The two color transfer algorithms can be represented by the following charts:

halo4.gif (6083 bytes)
Method 1: Colors are clipped at the end
of color range
  halo5.gif (6126 bytes)
Method 2: Colors are mapped along
the whole source color range.

On these drawings: S - source color space, D - destination color space, B - the color range that can be represented by the output device, A and C - the color range falling behind output device capabilities, E - output device color space scaled to the source color space.

When using method 1, colors from region B will be printed without changes. Colors from range A and B will be replaced with the closest match from range B. When method 2 is used, all colors printed look unique even if they are not completely the same as the original image.

The method 1 is best for vector illustrations because usually vector artwork doesn't have much colors and it is not likely that clipping will be visible. On photographs, though, it is possible to see the color distortion like shown on the illustration below:

halo6.jpg (7391 bytes)
Original Image
halo6-1.gif (987 bytes)
Color histogram of the original image
halo7.jpg (6177 bytes)
After applying method 1
halo7-1.gif (833 bytes)
Color histogram of result 1
halo8.jpg (5619 bytes)
After applying method 2
halo8-1.gif (1061 bytes)
Color histogram of result 2

On these simulated images, method 1 just clips everything that falls behind the output color space. Both resulting images have the same color range (as seen from the histograms above). The result #2 shows more details even if it differs more in color from the original.

When CorelDRAW prints a document on a composite printer, it applies one of the color mapping algorithm described above. By default, it uses method 1 for vector images and method 2 for bitmaps. And this is the cause for that pale bounding box around the dropshadow.

To remove the halo, you can do one of the following:

  1. Convert the underlying object(s) to bitmap. You can add the dropshadow to that combined bitmap, or print it as separate object - this doesn't matter. The most important thing is not to have vectors and bitmaps of the same color contact each other.
  2. In CorelDRAW, go to Tools>Options>Global>Color Management>General and you will see a drop-down list named "Mapping Mode" (in CorelDRAW 9 it is named "Rendering Intent"). Set it to either "Illustration" or "Photographic" ("Saturation" and "Perceptual" respectably in Draw 9), but make sure it is not "Automatic".

This should fix the problem. Note, that some printer drivers have color correction too. For example, HP DeskJet 850C uses automatic color mapping by default. Make sure you set it to either of the two methods, rather than automatic. HP DeskJet's driver names method 1 as "Vivid colors" and method 2 as "Match colors on screen". Unfortunately not all drivers allow you to turn off its color correction...


[ CorelDRAW Tips and Tricks | CorelDRAW Users' Page | Oberon Home Page ]
Copyright 2000 by Alex Vakulenko. All rights reserved.
This page was last revised on 01/13/00.