If you can spend as much time on the photo, I would:
Create a new exposure module and put it below the original one (so that if you change the exposure, it does not change the mask).
Create parametric mask for the eye (red) colour and/or lightness. If there is more red, combine with drawn mask around eyes. Some blur and feathering might be needed.
Decrease the exposure in the new layer.
Apply monochrome with the raster mask from exposure module and blend opacity to your liking.
You can also use tone curve instead of exposure decrease, but I recommend creating the mask in the exposure module. Using tone curve might make for easier catchlight preservation. Catchlight can be considered also in parametric mask creation, but it depends on the image.
You can also use color balance instead of monochrome, I have a frogmouth red eyed photo, where I decreased global saturation by 89 %. But I think monochrome would be easier with nearly the same results.
It might also be possible to use the RGB curve, but I do not use it.
The issue in my topic is rather more complicated to deal with, generally I darken the eyes a lot, monochrome, add noise and add synthetic (frequently not so good) catchlight by another tonecurve in hard light mode. Often I need to use color zones to take care of the colour spill from the eyes. Yes, it is very tedious .
PS: If you have more photos under similar light and exposure, you can create a style from the modules above and then simply create the drawn mask for each (or possibly not even that, if the parametric mask allows).
Be sure the cat tab is set to bypass and then the CC should equal the channel mixer… I suspect you might also be able to leverage the new primaries module…
Hello, @arturoisilvia:
The first link you point me to is the one I refer to in my first post and is now obsolete due to deprecation.
I did search for red eye and did not find anything that I could use.
And I don’t want to use anysthing else than dt.
Thanks for posting anyway, as as usually I often come back on subjects and read again so maybe I find something that I could not understand at first.
-=o#=O=#o=-
However, I think I finally accidentally found how to achieve my goal, yesternight just after thanking everyone here and turning this computer off. Dt is installed on a linux laptop in my bedroom, and before shutting it off I gave it another try and accidentally did two things I had not done before:
I created a duplicate instance of the color calibration module instead of using the existing one, and
I decreased the R level in the R tab instead of trying to adjust chroma and tint in the CAT tab.
Well, the result is not as good as what @Thomas_Do shows just above, so I am going to examine his method. Just need to copy to an usb key and transfer to the dt laptop.
Yes, I would also advise using a second instance of the color calibration module to avoid interfering with the general white balance settings.
I just used the gray tab of color calibration with a drawn mask. The input for monochrome was limited to “blue” because this gave a darker result with the outer darker ring of the pupil less visible.
Yes. Your approach increasing grey
seems to give softer and more natural results than decreasing red,
but it may not be always so.
I am also trying to dig out from my archives the way I used to do long ago, in the times of Bibble (further Aftershot Pro).
Thanks for helping.
It is not really obsolete… CC module when in bypass and if you use only the R G B tabs is the channel mixer… it is just part of the CC module and if you create a second instance it will be created by default in bypass it not using the CAT space so that it can be used as such. Therefore any references from the past or in other software to the channel mixer are still relevant wrt using a "channel mixer " in DT
That is what I did not do first, and that led to awful results: I did not create another instance.
Now I do.
As a casual user, I only use very few modules that are defaulted with setting that (are to) suit my needs, and I never modify anything except exposure.
Thus I hadn’t understood how to transpose the use of the old channel mixer in the color calibration, or how to use cc as a cm. Now, I know.
Thanks.
It was a topic of discussion more than one time that the combined features might be better left separate but in the end it has stayed as is. I think wb using the CC module has been one area that has confounded or at least caused many users to pause and at times struggle a bit to settle in on how to use it…
Back when I was using a camera with on-board flash I implemented a red eye tool in rawproc. Don’t know if this can be translated into use of some existing dt tool but here’s a snip from the comments describing the operation on a single pixel:
//The essential algorithm is to calculate the red intensity proportionate to the averaged
//green and blue intensities; if greater than the threshold, the red value is replaced
//with the average of the green and blue. This means that things like light reflections
//(white highlights) and other "non-red" pixels are not touched, leaving a well-formed eye.
Hello,
After some more experiments, it seems Red slider is not always the best choice.
In fact, Blue slider sems to give better results in most cases.
Starting from a red eyed photo
In some cases, it can be useful to tune the Green slider.
I am still reading the posts related to the modules, but some refer to other posts, so it is not that easy.
Thank-you @kofa. Now I could understand some subtilities of masking, which I did not before despites watching many tutorials. And I know for sure of at least four methods to remove red eyes.
Rgrds,
I’m a little confused why you first desaturate red, and then select the for the saturated part of the Chroma (Cz) tab, when it’s the red areas we shall apply the mask to.
I appreciate it if you would explain that, @kofa.
.
