Sigmoidization: stick to the "safer" contrast or even less

[NicolasRobidoux]

General comment regarding image resampling and enlargement using sigmoidization: I am starting to be of the opinion that even the "safer" value is a touch too large. The "less safe" value is almost certainly much too large unless you value perceptual sharpness at the expense of just about anything else, or your image is at least a little blurry (and not too noisy).
I'm still not sure, but it looks like the best "all purpose" contrast value may be closer to 7 than 10. So, the value used by Anthony Thyssen in his ImageMagick Examples discussion of sigmoidization http://www.imagemagick.org/Usage/resize ... _sigmoidal is probably pretty close to just right.
(In a month or so I'm hoping to have results from a quantitative test suite, notwithstanding the fact that their results should be taken with more than a postmodern grain of salt.)