sounds like you got it, but just adding a little template I made to
the thread. it depends on having a grid for the shot, and I know
that's a luxury not many people have, but it does seem to be becoming
more common.
it "zaps" CA by undistorting, bumping the offending channels and then
rewarping. It's just the thing for our stuff but I imagine CA varies
greatly between setups.
with the cam setup we use, it's been hard to profile lenses and pin
down a consistent formula - the lighting, focus, etc all play into it,
so I just gave up and decided I wanted grids every time. can't go
wrong.
if I recall, the Genesis setup that the last film I worked on was
using, there was a channel mismatch that wasn't necessarily CA - which
made things interesting! it wasn't a lens thing but a sensor thing
where the red channel just had to be scaled to 1.0045 or something
like that for everything to line up.
some bizarro stuff floating around out there that makes pulling keys
interesting, that's fo' sho'
set cut_paste_input [stack 0]
version 6.3 v3
NoOp {
inputs 0
name grid
selected true
xpos 1629
ypos 1506
}
LensDistortion1_0 {
serializeKnob ""
serialiseKnob "22 serialization::archive 4 0 0 0 0 0 0 0 0 0 0 0 0"
outputType Displacement
invertDistortion true
filter Rifman
"Card Parameters" 1
name LensDistortion
selected true
xpos 1629
ypos 1560
}
set N3a5efc10 [stack 0]
push $N3a5efc10
push $cut_paste_input
NoOp {
name plate
selected true
xpos 1738
ypos 1509
}
Group {
inputs 2
name ChromaticAbZapper
selected true
xpos 1738
ypos 1643
addUserKnob {20 User l NormanVision}
addUserKnob {4 hero l "Hero Channel" t "The channel that will remain
unaffected throughout processing" M {Red Green Blue}}
hero Green
addUserKnob {7 red l "Nudge Red" R -1 1}
addUserKnob {7 green l "Nudge Green" R -1 1}
addUserKnob {7 blue l "Nudge Blue" R -1 1}
addUserKnob {26 slug l "" +STARTLINE T "*requires LensDistortion node
set\nto displacement in displace input\n**meant to be done on a grid
and\nthen applied"}
}
Input {
inputs 0
name source
xpos 572
ypos 448
}
Dot {
name Dot3
xpos 603
ypos 488
}
set N896cfb30 [stack 0]
Shuffle {
red alpha
green alpha
blue alpha
name Shuffle4
label rgba.alpha
xpos 914
ypos 577
}
Dot {
name Dot2
xpos 945
ypos 822
}
Input {
inputs 0
name displace
xpos 469
ypos 517
number 1
}
Dot {
name Dot1
xpos 500
ypos 628
}
set N6c6eb060 [stack 0]
Grade {
channels forward
blackpoint {{parent.blue i}}
whitepoint {{1-parent.blue i}}
black_clamp false
name Grade4
xpos 750
ypos 651
disable {{hero==2 i}}
}
push $N896cfb30
Shuffle {
red blue
green blue
alpha black
name Shuffle1
label rgba.blue
xpos 792
ypos 578
}
STMap {
inputs 2
uv forward
name STMap3
xpos 792
ypos 715
}
push $N6c6eb060
Grade {
channels forward
blackpoint {{parent.red i}}
whitepoint {{1-parent.red i}}
black_clamp false
name Grade2
xpos 469
ypos 665
disable {{hero==0 i}}
}
push $N896cfb30
Shuffle {
green red
blue red
alpha black
name Shuffle2
label rgba.red
xpos 570
ypos 582
}
STMap {
inputs 2
uv forward
name STMap1
xpos 571
ypos 713
}
push $N6c6eb060
Grade {
channels forward
blackpoint {{parent.green i}}
whitepoint {{1-parent.green i}}
black_clamp false
name Grade3
xpos 635
ypos 654
disable {{hero==1 i}}
}
push $N896cfb30
Shuffle {
red green
blue green
alpha black
name Shuffle3
label rgba.green
xpos 682
ypos 578
}
STMap {
inputs 2
uv forward
name STMap2
xpos 682
ypos 712
}
Copy {
inputs 2
from0 rgba.red
to0 rgba.red
name Copy2
xpos 674
ypos 844
}
Copy {
inputs 2
from0 rgba.blue
to0 rgba.blue
name Copy3
xpos 674
ypos 930
}
Copy {
inputs 2
from0 rgba.alpha
to0 rgba.alpha
name Copy5
xpos 674
ypos 1007
}
NoOp {
name versionator
label "v1\n7-25-11\njbills\nborn"
xpos 674
ypos 1059
}
set N39b8f1e0 [stack 0]
Output {
name Output1
xpos 674
ypos 1176
}
push $N39b8f1e0
push $N39b8f1e0
Viewer {
inputs 2
input_process false
name Viewer1
xpos 1134
ypos 989
}
end_group
LensDistortion1_0 {
inputs 2
serializeKnob ""
serialiseKnob "22 serialization::archive 4 0 0 0 0 0 0 0 0 0 0 0 0"
distortion1 {{parent.LensDistortion.distortion1.L}}
distortion2 {{parent.LensDistortion.distortion2.L}}
distortionCenter {{parent.LensDistortion.distortionCenter.L.x}
{parent.LensDistortion.distortionCenter.L.y}}
filter Rifman
"Card Parameters" 1
cardScale {{parent.LensDistortion.cardScale.L.x x1 1}
{parent.LensDistortion.cardScale.L.y x1 1}
{parent.LensDistortion.cardScale.L.z x1 1}}
a {{parent.LensDistortion.a.L x1 0}}
b {{parent.LensDistortion.b.L x1 0}}
c {{parent.LensDistortion.c.L x1 0}}
name LensDistortion_clone
selected true
xpos 1738
ypos 1740
}
On Tue, Sep 27, 2011 at 3:51 AM, Darren Coombes
<[email protected]> wrote:
> Geez, that worked. So simple.
>
> Last 2 days have been loongg. So heads not to clear at moment.
>
> Thanks.
>
> Darren Coombes
>
> [email protected]
> 0418 631 079
>
> On 27/09/2011, at 8:37 PM, Eetu Martola <[email protected]> wrote:
>
>> On 27.9.2011 13:22, Darren Coombes wrote:
>>> Ok, so I hate to sound like a dumbass, but once I've shuffled the R, G, and
>>> B channels , what's the best way to merge them back together to form an rgb
>>> image?
>>
>> You surely do know this, but for quite a many operations you do not need to
>> shuffle things out and back in - you can e.g. transform the color channels
>> separately by just having the transform(masked) nodes in the main pipe and
>> tick the channels you want to scale.
>>
>> The Nuke way. :)
>>
>> eetu.
>> _______________________________________________
>> Nuke-users mailing list
>> [email protected], http://forums.thefoundry.co.uk/
>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
> _______________________________________________
> Nuke-users mailing list
> [email protected], http://forums.thefoundry.co.uk/
> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
>
>
_______________________________________________
Nuke-users mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
