Canon software (digital photo professional 4.14) produces really good results here with noise reduction as iso 12800, I cannot approach even close with colours and tonality,
rather default processing in canon software (picture style auto, saturation +2)
I’m trying to improve this with RT, but my results are worse,
I am aware of the difference between raw and jpg, and I’m not expecting to have exact results. But still … the canon results in the area of noise reduction are here pretty good.
What I hope to learn is the method using RT to create better high iso noise reduction results with such saturated colors in the foreground and such noise-free in the blurred background.
If there is such an algorithm in the RT - I want to learn how to use it, I will be gratefull for providing a link (currently tried clicking noise reduction and bringing the luminocity slider to the right).
If there is no corresponding algorithm in RT with such effectiveness - maybe it is also interesting to know. Or maybe it is just my subjective opinion that DPP result is more appealing here .
Hi @nullnull,
In many cases it is a matter of taste how to “develop” an image,
and how much to tweak it. Here is one interpretation, based on your cr2:
RT5.8 dev
Setting proper noise reduction values was a challenge, and the blue channel always wants to clip
After several iterations I came up with this:
@adlatus what did you do to make the flowers ping not blue ?
Could you explain which steps should I take ? (I tried moving in colour balance, but then the leaves changes colour too )
GENERAL QUESTION: I have one more general question. Why it is with this file, that auto matched curve is not equal to the camera ? Normally with my 6d ii, the auto matched curve produces results which are near the camera. In this raw file I have to tune the contrast way up to 68 ( ! ). What is the reason ?
Adlatus used both tone curves to do this, especially the second one made it more saturated (“pop”) without messing with the blues.
If you start playing with these curves you’ll notice that they can significantly influence the colour that’s shown.
I used the RGB curves module to do this. This does not influence the saturation all that much, which I prefer.
I use the RGB Parade option in the histogram to get an idea what needs to change. In this case the blue is to dominant. The red and blue parts need to be roughly on the same level (top part as well as the bottom part). You do need to keep an eye on the actual image though! In this case the mid tones need to be protected, so the midpoint of both red and blue curves need to be put back on the dotted line.
Here’s a rgb curves only profile that subtly changes the colours of your IMG_9807_rt-2.jpg.out.pp3 edit. It should give you an idea how to approach this.
About your general question: I really wouldn’t know. I never use the auto curves at all. If I have a look I notice that there isn’t much of an S curve at all, which has a big influence. I also noticed that RawTherapee has a hard time doing its demosaicing and capture sharpening magic with this image. In both cases I do not know why that is the case, I suspect a high ISO combined with an overall evenness of the RAW.
BTW: You might want to consider using LMMSE or IGV as your demosaicing method, those two are specifically for high ISO RAWs.
.
IMG_9807.CR2 (37.3 MB) 
IMG_9807_rt-1.jpg.out.pp3 (14.4 KB)
?