[Cin] Cingg'divide mode, very divisive!

[Phyllis Smith]

Trying to learn how to use layers in GIMP right now -- probably will take
all day.
Simple one line change in cinelerra/overlayframe.h around line 72:
     OLD:    (Dc > ZERO ? (Sc * mx) / Dc : ZERO))
     NEW:   (Sc > ZERO ? (Dc * mx) / Sc : ONE))
This new line is what I have been experimenting with.  Switching from ZERO
to ONE shows white instead of black.  Switching order of Dc and Sc changes
application order of Source over Destination.

On Wed, Nov 9, 2022 at 5:14 PM Andrew Randrianasulu <randrianasulu at gmail.com>
wrote:

>
>
> чт, 10 нояб. 2022 г., 03:04 Phyllis Smith <phylsmith2017 at gmail.com>:
>
>> A small mod can be applied to improve the Divide overlay *as long as no
>> one objects*. (most like Gimp). Old projects can use a previous AppImage
>> to get the previous results. The mod seems to fix the minor imperfections
>> you can see in the following screenshots using the Overlay plugin.
>>
>>      hhttps://cinelerra-gg.org/download/testing/divide_flowers.png
>> (Top first layer order)
>> In both Cinelerra 7 and Current CinGG, look at imperfection on upper
>> right hand corner about 1/3 of the way down where there are 2 small red
>> rectangles. And there are small dark blue squares scattered around in the
>> light blue areas.  These look wrong to me.
>>
>
>
> can those be amplified compression artefacts? If you load same images into
> gimp 2.10.xx do you see them?
>
>
> And what exactly you suggest to change in code? :)
>
>>
>>      https://cinelerra-gg.org/download/testing/divide_waterfall.png
>> (Bottom first layer order)
>> Entire left hand side just looks ugly for both Cinelerra 7 and current
>> CinGG.  I do not think anyone would want this output.
>>
>>    https://cinelerra-gg.org/download/testing/divide_squares.png    (Top
>> first layer order)
>> This test is interesting because all 3 look the same with no obvious
>> imperfections.
>>
>>
>>
>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cinelerra-gg.org/pipermail/cin/attachments/20221110/d09cb720/attachment-0001.htm>