Modulate needs a complete overhaul.
Posted: 2008-07-02T16:55:41-07:00
Actually I have never thought -modulate ever worked quite right, and I am right.
Applying modulate to a gradient image such as
I determined that a hue 'percentage' of 100 does nothing
That is -modulate 100,100,100
a value of 100,100,50 rolls red ->green -> blue -> red
and 100,100,150 the oppisite direction red -> blue -> green -> red
While values of 0 and 200 rotated the hue 180 degrees
Okay its weird but reasonable.
I would think '0' for no rottaion made more sense.
However switching to a 'colorwheel'
Then brigthness of 100 does nothing (fine)
and 50 looks like a 'half' brightness, and 0 pure black
but a brightenss of 200 is only 'half' white. That does not seem right!
Even a brightness of 1000 still has some 'color' and is not totally white!
That is brightness for some reason is using some try of non-linear scale!
Saturation also appears to be non-linear
100 no change 50, is greyish 0 is totally gray
however while 200 is very bright color wise 500 is even more so
and 1000 more still. Again it appears to be non-linear!!!
The -modulate is unfortunately one operator I never properly looked at in depth , when I first wrote the draft section on 'colors' in IM examples. The -level and -gamma were others operators which also were not looked at, but that had chnaged in the recent updates in the 'Color Modifications' page of IM Examples.
Now is the time to get -modulate FIXED now, in preparation for that re-write, in the user usage guides.
The question for developments then is how should it be fixed?
Applying modulate to a gradient image such as
I determined that a hue 'percentage' of 100 does nothing
That is -modulate 100,100,100
a value of 100,100,50 rolls red ->green -> blue -> red
and 100,100,150 the oppisite direction red -> blue -> green -> red
While values of 0 and 200 rotated the hue 180 degrees
Okay its weird but reasonable.
I would think '0' for no rottaion made more sense.
However switching to a 'colorwheel'
Then brigthness of 100 does nothing (fine)
and 50 looks like a 'half' brightness, and 0 pure black
but a brightenss of 200 is only 'half' white. That does not seem right!
Even a brightness of 1000 still has some 'color' and is not totally white!
That is brightness for some reason is using some try of non-linear scale!
Saturation also appears to be non-linear
100 no change 50, is greyish 0 is totally gray
however while 200 is very bright color wise 500 is even more so
and 1000 more still. Again it appears to be non-linear!!!
The -modulate is unfortunately one operator I never properly looked at in depth , when I first wrote the draft section on 'colors' in IM examples. The -level and -gamma were others operators which also were not looked at, but that had chnaged in the recent updates in the 'Color Modifications' page of IM Examples.
Now is the time to get -modulate FIXED now, in preparation for that re-write, in the user usage guides.
The question for developments then is how should it be fixed?