How to Use Corel PhotoPaint 8 to Make Web-Friendly GIF & JPEG Images

Web Works Part 8 - GIFs 4

Reducing the number of colours in paletted images.

Reducing the number of colours in a paletted image can be accomplished using the Convert to/ Paletted 8 Bit dialog box.

Only the Optimized and the Adaptive palettes allow us to reduce the number of colours. If either of these palette types is selected, the Colours numeric control adjustment in the Convert to 8 Bit Paletted dialog box becomes selectable and operational.

Change the number of colours and observe the result in the preview pane.

To see the effect of reducing the number of colours in an image clearly, ensure that Dithering is set to None at this stage.

Reduce the number of colours to a point just above any obvious reduction in image quality is reached.

Now choose a number about half of this. The image loses considerable quality.

At this point choose 'Error Diffusion' as the Dithering type. Note that the image regains some of its former quality. Dithering tricks the eye into smoothing out contrasting areas BUT also increases file size.

The Optimized palette provides a great deal more control over the selection of colours during colour reduction and is described in greater detail separately below.



The Optimized Palette in Detail

The Optimized palette controls features make it is possible to produce almost photographic-like images with as few as 16 colours, depending on the starting image. It is also possible to make high quality, realistic skin and natural tone images while simultaneously reducing the number of colours to an absolute minimum.

Even the versatile Adaptive palette cannot match the level of control the Optimized palette provides over almost every aspect of palette construction.

Indeed, an entire tab on the Convert to Paletted (8 bit) dialog box is devoted to controlling the Optimization engine's features.


Reducing the number of colours in Optimized palette images.

Reducing the number of colours in an Optimized palette image provides a much greater level of control than the same task in an Adaptive palette image. If you haven't yet read the section on the general techniques for colour reduction for both the Adaptive and Optimized palettes then follow the discussion above and return here for specific use of Optimized palette images. The Optimized palette controls allow us to produce almost photographic type images with extremely low numbers of colours without needing to resort to dithering which always makes an image look grainy. When the Optimized palette is selected, clicking on the Colour Range Sensitivity to: checkbox not only allows you to use the eyedropper to select the sensitivity range colour (in the Options tab) it also now activates all of the features for colour selection on the Range Sensitivity Tab. For any other palette, or if the Colour Range Sensitivity to checkbox in the Options tab is not selected, the features on the Range Sensitivity tab remain unavailable. For the following exercise, select a high quality 24 bit image with regions of fine colour change such as sand, sky or skin tones, then select Image/Convert to Paletted (8 bit) and the Optimized palette as described earlier.


Controlling the preservation of colours during colour reduction.

The Optimized Palette allows the user to select any colour from the original image as a base for optimisation. The 'Colour range sensitivity to..' checkbox is available for selection when the Optimized palette is selected. Reduce the number of colours in your image until you reach a point where the image starts to lose quality. At this point check the 'Colour range sensitivity to..' checkbox. The Eyedropper tool to the right of this check box now becomes selectable and the colour swatch beside the eyedropper shows the current selected colour. Use the eyedropper to select a colour from the Original pane - choose a colour which has now been lost from the Result pane. Notice how the colours in the image in the Results pane changes. Select different colour pixels from the image in the Original pane and observe the results.




When you think you have a fairly good representation of the original image, click the Range Sensitivity tab on the Convert to Paletted (8 bit) dialog box. To use the slider controls in this dialog, both the Optimized palette and the 'Colour range sensitivity to..' checkbox must both be selected and activated. The selected colour (from the Options tab) is shown. Below this, slider controls for the Importance, Lightness and a (green-red) and b (blue-yellow) axes are available, plus Reset buttons to return settings to the defaults. Try shifting the Importance slider control first and note the result. Adjusting this slider can dramatically increase the time taken to render the image so be prepared for a wait during computations. Press Reset (for the Importance slider) then adjust the other controls and observe the results. One can spend quite some time refining an image this way but it provide the perfect gif image you require. Remember to view the resultant palette by clicking on the Processed Palette tab occasionally.


Checking and changing the actual colours within a variable palette.

The 'Processed Palette' tab in the Convert to Paletted (8 bit) dialog box provides a mechanism to manually change individual palette entries or to add or delete entries. To alter the palette manually, click the Edit button to bring up the Colour Table Dialog (also available from the Image/Color Table menu in normal edit mode). Any palette colour can be replaced by another by selecting a colour from the existing palette first (on the right) and then a colour from the colour selection area on the left and pressing the Replace button. Values for both colours are displayed under their respective selectors. Similarly colours can be added and deleted from the current palette. The Find Closest button finds the entry in the palette which is closest to that selected in the Colour selector region on the left. A more complete discussion on colour and colour selection methods in Corel PhotoPaint is best left for a future article due to the scope of such a topic. Meanwhile, try replacing, deleting and adding colours to your image's palette and note the results.