Hi Larry,
Indeed HDRWaveletTransform is very addictive
Wait to see the new tutorial (due tomorrow), where we describe a detailed, step-by-step processing example that uses this tool, among others. The results are outstanding.
Basically, multiple HDRWT iterations can be used to cause a stronger effect. But be aware that it's easy to overprocess an image in this way, so use multiple iterations with care.
With more than one iteration, you can also check the
inverted option. Inverted iterations means that the process alternates between applying the HDRWT algorithm to dark and bright image structures. This is useful to prevent excessive darkening in the shadows. By default (with one iteration or no inversion), only bright structures are affected.
The term "flat" here means that the HDRWT algorithm tends to uniformize the distribution of brightness in the image. Imagine an image of M42. If we plot brightness versus position on a straight line that crosses the image, the extremely bright core of M42 would be represented as a tremendous "mountain". After applying HDRWT, the mountain would be much easier to climb, and the valleys would be higher. After several HDRWT iterations, we'd end with a sort of plateau.
What makes HDRWT so powerful is its
multiscale nature: it flattens the image up to a given scale, which corresponds to the
number of layers parameter. The more layers, the larger structures will be flattened. Of course, each image is different and requires different parameters.
An interesting option is to apply HDRWT several times at different scales. For example, you can try with a first application at 6 layers and a second one at 4 layers. This would apply the algorithm to a wider range of image structures.
