Redcine NR

[Anders Holck]

Just did a small test of the build-in NR in RedCine 1.0.3.
All images is cropped 1:1 from a 4k R3D file at sharpness:100 and Detail:max.

This is the blue channel:

RGB:

And a color chart showing the color bleed:
http://www.andersholck.com/red/noise3.mov

More testing is needed, but it Looks promising at defeating those dreaded artefacts....

 

[R Fogg]

I agree, the noise reduction does wonders for the blue channel noise. I wonder if having channel specific NR would be worthwhile.

 

[Graeme Nattress]

Anders, can you let me have the R3D that you used for your colour bleed test? I think what you see should be gone in the next iteration, but I want to be sure.

Graeme

 

[Anders Holck]

The Chart is jBeale's from his site http://www.bealecorner.org/red/test-080108/R3D/4k-ChartA-T56-1-3.R3D
He also offers this one: http://www.bealecorner.org/red/test-080108/R3D/4k-ChartB-T8-1-12.R3D which should stress that routines to the edge...

 

[Graeme Nattress]

Thanks Anders!

Graeme

 

[Peter McCully]

Anders
Thanks but where in Redcine do I find the NR tool?

 

[Anders Holck]

It's a "per shot" setting in the color pane:

 

[Peter McCully]

Yes, of course. I was looking for it in the previous version! My stupid.
I have been having a close look at the noise levels in low-light and there is more work to be done in future to improve this. I have noticed the noise diference in the 'half way line' - where there is far more noise in the lower half of the frame compared to the top half. There are also really ugly spots of random noise on otherwise smooth parts of the picture. That, and the clipping at the top end are problems I'm sure the Red team are continually working on.....I hope anyway.

 

[jbeale]

This is all great stuff Anders, I am happy to hear the test chart files are seeing some use! And thanks Graeme, for continuing to make an already good thing better.

By the way, that color chart at the center of http://www.bealecorner.org/red/test-080108/index.html#Test_2 has accurately known colors, see
http://www.bealecorner.org/red/test-080108/R060101.txt
(listed for a D50 illuminant, which is not far from what I actually used). So I believe that you could actually build a color profile for the camera from that image. I don't know how close the camera-to-camera matching is, though, if you wanted to use it for some different Red.