Similar topic...
FWIW You can also use a similar technique to remove camera dirt etc, which is
behind the sensorClean tool (after a similar discussion a few years ago)
http://www.nukepedia.com/gizmos/gizmo-downloads/filter/sensorclean/
In this case you have a dirty held frame, a cleaned version of the same held
frame and the image to clean.
It divides the dirty and clean reference frames which is the case of sensor
dirt, gives a white frame with a superbright spot where the dirt was. the
result is then multiplied on the image to clean, and removes the dirt spots,
depending on how well you painted the clean reference. Not for all shots but
works on most (90% I'd say).
Howard
From: Fredrik Pihl fre...@gmail.com
To: Nuke user discussion nuke-users@support.thefoundry.co.uk
Sent: Sunday, 13 May 2012, 22:37
Subject: Re: Re: [Nuke-users] Analyze Brightness in a Region of Pixels to
drive Grade Node
lovely tip by mr Rowell :)
On Sun, May 13, 2012 at 9:57 PM, Paul Schoen dottore.pa...@gmail.com wrote:
Hey,
thanks for the useful input using the curve tool as well as the blurring the
rectangle-trick. I messed around with the curve tool with something like
this CurveTool1.intensitydata.r/CurveTool1.intensitydata.r(start_frame)(in
my case 1) and put in the gain of the grade node.
But because the shadows are wandering and do not darken the picture
uniformally, I had no luck with that. So I tried the tip by James Rowell
which works absolutely magnificent because you can adjust the crop according
to the area to be covered and minor local changes are reflected as well
depending on the strength of the blur applied to the crop. You can even mix
different area with aligning several crops together I found out.
Of course it still needed fine tuning because shadows do not only darken
certain areas but also make them more blue (=cooler)...
Thanks a lot!
Paul, Vienna
2012/5/12 Eetu Martola e...@undo.fi
From : James Rowell
Subject : Re: [Nuke-users] Analyze Brightness in a Region of Pixels to
drive Grade Node
Hi Paul,
There may be a tool that already does what I'm about to describe, but I've
done something like what (I think) you want in the past by cropping to the
small area that you want to analyse then put a sufficient blur on that
small area to smooth it all out, then reformat the small crop up to a full
size image. Then, take a frameHold on that image on some reference
frame, then divide the one by the other and you now have a full frame image
that you can use to multiply against anything else to CC it in a way that
will mimic the frame by frame difference of your little sampled area over
the course of the shot, so you can (for example) reproduce flickering, or
even remove flickering if you flip the order of the inputs on the divide.
Anyway, this may give you other ideas too.
Regards,
James Rowell
From: Paul Schoen dottore.pa...@gmail.com
To: Nuke-users@support.thefoundry.co.uk
Sent: Saturday, May 12, 2012 5:26 AM
Subject: [Nuke-users] Analyze Brightness in a Region of Pixels to drive
Grade Node
Hello,
this is my first post to the mailing list, so I hopefully get everything
right in explaining my particular problem:
I have a composite of a Background in which feet are moving over a
pavement. As they move they make shadows on the pavement. The foreground is
moving spider that was filmed on the same pavement later on. Now I want the
brightness of my foreground to change exactly like the shadows of the
moving feet change the brightness of the pavement in the background.
Long time ago, there was a node in Shake called AnalysePixel or something
like that which could read out a certain area of pixel over the whole
length of a shot. With this data one could drive the ColorCorrect-node to
automatically change the the brightness or any other value. I used it once
to match a non-flickering shot to a flickering one (NOT FLICKER REMOVAL!!!).
I found nothing similar in Nuke except the Spotmeter-function in the
Viewer. But I found no hint to read its values and use it in an expression.
I need something to constantly analyze the brightness over a given range of
frames in a certain region of pixels and to drive a Grade or
ColorCorrect-Node with this values via an expression.
Does anyone have any hints on this? I'm using Nuke 6.2v1.
For convenience I uploaded the comp layout:
http://dl.dropbox.com/u/1891745/dpadb_022_lay_v01.mov
Thanks,
Paul Schoen, Vienna
___
Nuke-users mailing list
Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
___
Nuke-users mailing list
Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/
