Gaussian blur bias

[snibgo]

According to the documentation http://www.imagemagick.org/script/comma ... ssian-blur , this can take a "bias" argument.

-gaussian-blur radiusxsigma+bias

What does bias do, or what should it do? I've tried it with various numbers, with or without "%", and it seems to make no difference.

(I ask because I'm building an interface so nip2 can drive IM. If bias doesn't do anything, I'll leave it out.)

Bias isn't mentioned on Anthony's http://www.imagemagick.org/Usage/blur/ page.

The same question applies to motion blur: "-motion-blur radiusxsigma+angle+bias"

[fmw42]

I don't read code that well, but as far as I can see there is no bias argument for motion-blur or gaussian-blur. If it ever existed, it was probably removed when the code was modified to use morphologic equivalents when morphology was introduced. You can look at effects.c in the magick folder of the software download.

[snibgo]

Thanks, Fred. I've just seen that the "-sharpen" documentation also mentions bias, which also seems to have no effect.

Perhaps these should be removed from the documentation.

[fmw42]

Thanks, Fred. I've just seen that the "-sharpen" documentation also mentions bias, which also seems to have no effect.

Perhaps these should be removed from the documentation.

I will check with Magick and if not used, remove them.

[fmw42]

Magick says they were removed when switching to the morphology equivalent code. I think they were just adding a bias value to the output such as with -bias. But I tested and none of the functions seem to be affected by -bias or -define convolve:bias=. I had never used them and do not think they are that important or useful.

So I will just remove the bias arguments from the options page for those functions later today. Probably won't show until tomorrow or at least a few hours.