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

Reply via email to