My sharpening workflow for base ISO mages with the new features in RT (aka extreme pixel peeping)


(Ingo Weyrich) #61

I just added xtrans 2-pass (combination of xtrans 1-pass and xtrans fast)

(Gimbal Lock) #62

Since there is the contrast threshold adjuster for sharpening, I use much more aggressive settings :slight_smile:

But the threshold only works by masking away things that are soft. What if one could also mask away things that already are sharp enough? Would that be possible? Does it even make sense to do so?

(Sebastien Guyader) #63

Great idea about the x-trans 2-pass. The other day I was thinking that with this new compound demosaicing method, 3-pass sharpening for x-trans could be less useful (or at least, less often needed).

(Ingo Weyrich) #64

That would also be possible. Whether it makes sense, I don’t know atm…

(Ingo Weyrich) #65

Sébastien, currently the xtrans method names showed in gui are:

        "3-pass (best)",
        "1-pass (medium)",

which is not correct anymore as 3-pass is not the best imho. Can you, being an xtrans user, suggest better method names?

(Sebastien Guyader) #66

Info, maybe the 1-pass and 3-pass choices can be renamed to something like “1-pass (Markesteijn)” and “3-pass (Markesteijn)”. And 2-pass could be renamed “fast + 1-pass”, and 4-pass “fast + 3-pass”.

(Ingo Weyrich) #67

Sébastien, to keep things consistent with bayer demosaicers (where the flat demosaic algorithm is the second in the name), would you also agree with

1-pass + fast
3-pass + fast



How about family pass or season pass? :stuck_out_tongue:

(Ingo Weyrich) #69

@afre ?


Joking aside, for consistency, maybe:


If you like add (combo), (hybrid), (dual), or whatever next to it.

(Ingo Weyrich) #71


(Guillaume Castagnino) #72

Hi @heckflosse

I just opened a bug for you :
I get sometimes artifacts while playing with the contrast slider.

(Samuel Chia) #73

It could make sense if you are trying to even out the sharpness across the entire image. Usually the corners and edges are less sharp than the center. And most lenses are not radially symmetrical in sharpness either.

The other case for protecting edges that are already very sharp is to avoid the development of sharpening halos and/or clipping of the pixels along the edge.


I don’t have any Amazing and so on mode on Demosaicing tab:

O! I find that this mode is for Canon/Nikon, but I’m working with Fuji.

(Sebastien Guyader) #75

@Chawoosh the equivalent for xtrans are 1-pas+fast and 3-pass+fast


Yes, I select 3-pass + fast.

(Stefan Chirila) #77

This is a great tutorial btw. I wish there was a tagging system for all tutorials on the forum so I could just collect them all :smiley:

(Mica) #78

See this:

(Stefan Chirila) #79

@paperdigits hanks a bunch, I shall add tutorial as keyword on my previous tutorials

(niko) #80

Hy, im new here, im spanish and my english is poor, so ill try to explain as much as i can.
I´ve used RT since years, and i want to expose wich things are wrong and good to me as a all arround amateur photographer and if anyone who make rcd code could fix those things
For example, i ussually shot with my Pentax k50 and, while it has a directional OLPF, at high frequency lines or nyquist i see some horizontal bandings artifacts, even more visible when those bands are perpendicular with another one. I´ll try to upload images, from the test charts of imagingresource to compare both demosaicing methods