Hey @donyagithub I have checked the images and I need these to have more overexposed areas. Compare with my photo here, that is accepted by the noise profiler. As you can see, the overexposed area is approximately three times bigger than yours.
Also, here is the frame I used to get the successful results.
Make a new set of photos and I am happy to try again!
Hi @lomo okay, as soon as I can I’ll take new photos by increasing the pure white area. Thank you
here you can download the images
Actually, the overexposed area is still too small but this time I could extend it with a retouch and export the images as tiff files and do the noise profiling. Not perfect but should do.
I found some raw photos online from the camera and the denoise looks pretty good to me.
noiseprofiles.zip (478.0 KB)
If you are on Windows, find your DT installation, make a backup of your existing noiseprofiles.json and replace it with the file attached.
If you are on Linux, you will find the json file in the /usr/share/darktable catalog.
Then start DT as usual, open a photo from your camera and the Denoise Profiled module should find a noise profile for your camera.
I will send the profile for approval to get it into next release.
perhaps, at this point, it would be better if I redo the photos, so as to improve the final result. If you can wait a few days, I should be able to do them.
Absolutely, it may make a (small) difference so let’s try.
The generated profile is tested on raw files and they look good though.
here we are! Let’s see if we can improve it. Here you can download the photos. Thank you
Nope sorry, none of these were accepted by the profiler.
These images are one stop darker compared to the previous batch and even though the white area is larger, these photos aren’t over exposed. The previous batch really was clipped and overexposed. The histogram must be pushed all the way to the right side and then some. If you open one of these in DT and disable the exposure module, you will see that there is quite a gap in the high end of the histogram, and we don’t want that gap.
ok, I understood the problem, I will take new photos in a few days
This time they should be sufficiently overexposed. I shooted two group of images, with two different level of overexposed. I hope that all will ok. Thank you
The fourth batch worked without complaints. A quick comparison to the previous batch, this one leaves a bit graininess but details looks a bit better. Here is a new noiseprofiles.json for you. Like I wrote in an earlier comment:
If you are on Windows, find your DT installation, make a backup of your existing
noiseprofiles.jsonand replace it with the file attached.If you are on Linux, you will find the json file in the
/usr/share/darktablecatalog.Then start DT as usual, open a photo from your camera and the Denoise Profiled module should find a noise profile for your camera.
I have sent this batch to the devs as a replacement for the earlier approved profiling.
noiseprofiles.zip (478.3 KB)
PS. The fifth batch worked too, and the outcome was very close so need to dig further.
EDIT: Already updated in the repo for release 5.0. 
Could this be okay, or is there still something we can do to improve it?
This is exactly what should be and it is already approved as a denoise profile for next major release.
Very well! …being able to cooperate is a great Thing 
Thank you