Thanks Frank! Will give it a shot. Is there a way so that I can isolate a pixel that has the same value in the R G and B Channel values. Eg. I have some "bad" pixels with values such as 0.5,0.5,0.5 0.765,0.765,0.765 0.32,0.32,0.32 etc....
Their neighbour pixels would always have different R G and B values. Usually lower because the bad pixels are always brighter than their neighbours. I'll try to post an image later. Dave On Wed, Jul 24, 2013 at 12:42 PM, Frank Rueter <[email protected]>wrote: > in that case just use something like > r>13, g>13, b>13 > > to check every channel against a threshold or to check the average instead: > (r+g+b)/3>13 > > etc. > > If you don't like the expression node you can also use Clamp, set the > minimum and maximum to whatever threshold and turn on minClampTo and > maxClampTo to map the clamped values to black and white. > > > > > > > On 24/07/13 16:36, David Yu wrote: > > Not NaNs. the hot pixel have RGB values do isnan does not work > > > On Wed, Jul 24, 2013 at 5:28 AM, Frank Rueter <[email protected]>wrote: > >> and if it's NaNs or infs you need to isolate, you are after you can use >> "isnan()" and "isinf()" in the Epxresison node to get a mask >> >> >> On 24/07/13 09:04, Frank Rueter wrote: >> >> I'd use this technique to generate the mask for the bad pixels, but >> instead of blurring heavily, I would stencil out the original frame with >> the mask, then blur it just enough for a subsequent unpremult to fill the >> gap. >> >> >> On 24/07/13 05:47, John Mangia wrote: >> >> Use an expression node with r > value ? 0 : 1 where value is the upper >> threshold. Use that matte and keymix in a heavily blurred or translated >> frame within the hot pixel matte. >> >> On Tue, Jul 23, 2013 at 1:43 PM, Ron Ganbar <[email protected]> wrote: >> >>> Define hot. >>> Do you mean NaNs and INFs? >>> >>> >>> >>> Ron Ganbar >>> email: [email protected] >>> tel: +44 (0)7968 007 309 <%2B44%20%280%297968%20007%20309> [UK] >>> +972 (0)54 255 9765 <%2B972%20%280%2954%20255%209765> [Israel] >>> url: http://ronganbar.wordpress.com/ >>> >>> >>> On Tue, Jul 23, 2013 at 8:41 PM, David Yu <[email protected]> wrote: >>> >>>> Hi. Anyone know a script or expression to remove hot pixel noise from >>>> images or 3d renders? >>>> >>>> I tried some crude 3x3 matrix tricks to enhance the pixels and tried >>>> to "pull" the pixels out and subtract from the main image. Then i used the >>>> result as a mask for a median filter set at 1. >>>> >>>> Looking at the RGB channels, the hot pixel is the same value on all >>>> 3 channels while the neighbour pixels are different per channel. I'm hoping >>>> for a more elegant way to detect hot pixel using an expression. >>>> >>>> Dave >>>> >>>> _______________________________________________ >>>> 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 >>> >> >> >> >> -- >> John Mangia >> >> 908.616.1796 >> [email protected] >> >> _______________________________________________ >> Nuke-users mailing [email protected], >> http://forums.thefoundry.co.uk/http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> >> >> >> >> _______________________________________________ >> Nuke-users mailing [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 [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
