Running Linux, 5th Edition - Matthias Kalle Dalheimer [172]
Figure 9-26. Channels dialog
Figure 9-27. Layers dialog
Figure 9-28. Layers and mask
Filters. The final major aspect of GIMP we cover here is its filters. Filters are effects that can be applied to an entire image or a selection. GIMP is shipped with a large number of different filters, and it is possible to plug in new filters to extend the capabilities of GIMP. Filters are located in the right mouse button Filters menu. The Channel Mixer is an example of such a filter. We discuss two useful filters, Gaussian Blur and Unsharp Mask, and apply them to the image from the previous example.
Gaussian Blur
This filter provides a nice smooth blurring effect. Try it with different blur radius settings. The IIR-type Gaussian blur seems to look better than RLE with most images.
For our example we are not going to blur the actual image. Instead, we are going to smooth out the transition between the high- and low-contrast layers. Do this by selecting the layer mask in the sky layer and applying Gaussian Blur. A radius of 8 seems to work well here. Zoom in on the border between the trees and the sky, and don't be afraid to experiment — you can always press Ctrl-Z to undo and try again. Figure 9-31 shows a closeup of before and after applying Gaussian Blur to the mask. The effect is subtle, but important for making the two layers blend seamlessly.
Unsharp Mask
Despite its name, Unsharp Mask is a filter for enhancing the perceived sharpness of images. It offers more control and often provides more pleasing results than the simple Sharpen filter. Unsharp Mask works like this: first it makes an internal copy of your image and applies a Gaussian blur to it. Then it calculates the difference between the original and the blurred image for each pixel, multiplies that difference by a factor, and finally adds it to the original image. The idea is that blurring affects sharp edges much more than even surfaces, so the difference is large close to the sharp edges in the image. Adding the difference back further emphasizes those sharp edges. The Radius setting for Unsharp Mask is the radius for the Gaussian blur step, the Amount is the factor that the differences are multipled by, and the Threshold setting is for ignoring differences smaller than the chosen value. Setting a higher threshold can help when working with images with digital noise in them so we don't sharpen the noise.
Figure 9-29. Two layers
Figure 9-30. Two layers
Looking at our example with the sky and car again, we notice that the high-contrast part lost some details in the shadows when we pulled up the contrast. This can also be remedied with Unsharp Mask. To do this, we apply Unsharp Mask with a high radius and low amount. This technique is called "local contrast enhancement." Start out by making a copy of the whole image by pressing Ctrl-D and merging all layers in the copy. This is done by choosing Image → Flatten Image from the context menu. Then we want to scale the image for screen viewing. Open the scaling dialog by choosing Image → Scale Image from the context menu and choosing a suitable size and the bicubic (best) scaling algorithm. Now we are ready to apply Unsharp Mask for local contrast enhancement. A radius of 25, an amount of 0.15, and threshold of 0 seems to look good.
Figure 9-31. Blurring the mask — before and after
Finally, we want to sharpen up the edges a bit. To do this, we apply Unsharp Mask with a small radius (0.5) and higher amount (0.5) and with a threshold of 6. Figure 9-32 shows the unsharpened image on the left, the image with local contrast enhancement applied in the middle, and the image with the final sharpening pass applied on the right.
Recording Tools
If you want to create your own MP3 files, you will need