Raw Therapee on a top tier smartphone camera - Sharp Aquos R6

The Sharp Aquos R6 has a one-inch BSI sensor. I found a .dng at

You can see how I processed it in RT at

and the .pp3 is attached.

Leica just announced its first smartphone, the Leitz 1, which is basically this phone.

SharpAquosR_sample1.pp3 (12.3 KB)

ART is great with smartphone pics…using Flat field correction it will apply the embedded gain map…
image

1 Like

Thank you - but when I re-opened the DNG, the Flat-Field dialog had no box for Embedded in metadata. Either I made a mistake or Sharp/Leica did things differently.

It must not be supported I have tried it on three phones and it worked …this is a newer phone so maybe not for this one??

Did you test with ART or just with RT?

With my pixel phone this only works in ART and not RT…ART also seems to be detecting some embedded info to make a lens correction…not sure if it is using different approach to reading metadata vs RT or not??

According to exiftool, DSC_0136.dng was taken with a 6.8mm focal length lens. This seems unlikely with a one-inch sensor.

This dng does contain a GainMap (lens shading map), and it does work with the Flat-Field - Embedded in metadata option in ART 1.9.3. It does affect the color, with default settings the image has a blue-green cast if it’s not applied.

That has been a great tip you shared on this one a while back…much better starting point for my pixel shots…

Ah, just now I see there is a fork of RT called ART. Might RT bring in this function of using a Flat field embedded in the image file? I can’t keep current on both RT and a fork of it.

Leica claims this for the Leitz 1. It might differ from the Sharp Aquos R6 in specific parameters, although based on it in general. However, I doubt that Leica put a different lens in there, since it apparently did the engineering for camera in the Sharp phone.
1” CMOS image sensor, 20.2 MP
Aperture: f/1.9
Focal length: 19 mm (35 mm format equivalent)
https://en.leica-camera.com/World-of-Leica/Leitz-Phone/Leitz-Phone-1

Actually easier than you think…nightly builds of art are available as zip files just unzip and go…delete the folder if you don’t like it or have no use…it is a very nice little streamlined version of RT…If you are on windows…here…there are others as well ART-W64NightlyBuilds/ – Keybase.pub

https://bitbucket.org/agriggio/art/wiki/Home

As a workaround… I have been experimenting with an alternate way of dealing with these smartphone dngs that contain a GainMap: a program that outputs another dng with the GainMap applied that can then be edited in RT, Darktable, or any other editor that does not have the feature built in. It’s not very user friendly but it seems to work. Details here.

The discussion has noted color problems such as an overall cast and, in paolod’s GainMap link, vignetting that varies by R,G,B. If you read the promotional material for the Leitz Phone 1, you notice emphasis on supposedly special processing that produces a Leica black and white look. That suggests that the some of the color problems are difficult to deal with. Until the hardware can avoid those problems, Leica seems to nudge you toward black and white.

You had offered that before but I never found enough instruction for me to compile it…

Is there a script or some simple syntax that you could share for building it??

Todd, you left out one additional step that MUST be done when you change to a new zipped version in a Windows installation: you must also delete the ART-dev folder “C:\Users\username\AppData\Local\ART-dev”. This folder will be recreated, and correctly for the newer version, when ART is started again.

1 Like

Good to know in the way I have run it I never noticed a problem but I guess I don’t rely on changed to the configuration file…

Sorry, I need to clean it up and add a readme… For now I think this should work:

git clone https://github.com/paolodepetrillo/raw2dng.git
cd raw2dng
git checkout dngpreprocess
mkdir build
cd build
cmake ..
make

If everything works this will create the executable at build/dngpreprocess/dngpreprocess. To run:

dngpreprocess IMG_9999.dng

This will output a file named IMG_9999_pp.dng with the GainMap applied.

Thank you. I will give it a try. I appreciate you taking the time to reply …

I gave it a try but I am not skilled enough to troubleshoot… I got this…looks like cmake is not set correctly ???

image