Project

General

Profile

Bug #11983

Panasonic TZ70 (aka TZ71) Automatic Lens Correction

Added by Chris Elston over 1 year ago. Updated about 1 year ago.

Status:
Closed: upstream
Priority:
Low
Assignee:
-
Category:
Darkroom
Target version:
Start date:
01/31/2018
Due date:
% Done:

0%

Estimated time:
Affected Version:
2.4.1
System:
other GNU/Linux
bitness:
64-bit
hardware architecture:
amd64/x86

Description

The Panasonic DMC-TZ70 (in Darktable referred to as TZ71) is a compact camera with a fixed lens.

Lens calibration settings are available in the lensfun project and can be selected manually from the Lens Correction module.

However, as the Lens (Leica - fixed lens) is not stored in the image's exif data, it is not matched automatically by the Lens Correction module.

As this is the only available lens for this camera, it should be automatically selected by the DT module and I can't find any way to make it do so.

I've tried creating a preset and a style but unfortunately both of these methods seem to retain other settings in the module in addition to the Camera/Lens settings (focal length/aperture) so won't work correctly when applied to differently exposed images. For example if I create the preset on an image shot at 22mm and f/6.3, those same settings will be applied when I select the preset/apply the style on a picture that's 6mm and f/8.

The only way is to manually select camera/lens each time I apply the module, which is quite tiresome to do on hundreds of images at a time.

P1010026.RW2 (12.7 MB) P1010026.RW2 Chris Elston, 01/31/2018 01:03 PM
P1010038.RW2 (12.6 MB) P1010038.RW2 Chris Elston, 01/31/2018 01:04 PM

History

#1 Updated by Roman Lebedev over 1 year ago

Sample?

#2 Updated by Roman Lebedev over 1 year ago

  • % Done changed from 0 to 20
  • Status changed from New to Incomplete

#3 Updated by Chris Elston over 1 year ago

Attached two samples with different aperture/focal length

#4 Updated by Roman Lebedev over 1 year ago

  • % Done changed from 20 to 0
  • Status changed from Incomplete to Closed: upstream

I see no "Panasonic TZ70" in lensfun library, only TZ71, therefore the autodetection fails.

#5 Updated by Chris Elston over 1 year ago

For some reason (see also Bug #11977) TZ70 is defined as alias of TZ71 in https://github.com/darktable-org/rawspeed/blob/develop/data/cameras.xml (line 7020). The denoise (profiled) module therefore detects the TZ70 as TZ71. I don't know of the reason for this but these modules should work consistently - it can't be treated as TZ71 in denoise (profiled) but TZ70 in lens correction.

The TZ71 data in the lensfun project is data for the TZ70.

#6 Updated by Roman Lebedev over 1 year ago

Chris Elston wrote:

For some reason (see also Bug #11977) TZ70 is defined as alias of TZ71 in https://github.com/darktable-org/rawspeed/blob/develop/data/cameras.xml (line 7020). The denoise (profiled) module therefore detects the TZ70 as TZ71. I don't know of the reason for this but these modules should work consistently - it can't be treated as TZ71 in denoise (profiled) but TZ70 in lens correction.

The TZ71 data in the lensfun project is data for the TZ70.

Please show me where in the lensfun there is data for TZ70.

#7 Updated by Chris Elston over 1 year ago

It's listed under model TZ71. As I said, other Darktable modules refer to the TZ70 model as TZ71, as does the lensfun project.

AFAIK the TZ70 and TZ71 are essentially identical cameras and have therefore been aliased within Darktable so that both models are referenced as TZ71. Denoise (profiled) matches my camera to stored profiles for TZ71 (that I submitted in Bug #11977).

Given that lensfun has stored correction data against the TZ71 model, and the corrections would be identical for the TZ70, this aliasing makes sense, but the model is handled differently within the lens correction and denoise (profiled) modules (matched in the latter but not in the former).

#8 Updated by Roman Lebedev over 1 year ago

Chris Elston wrote:

It's listed under model TZ71.

Link please.

#9 Updated by Chris Elston over 1 year ago

All I can tell you is what I was informed in response to my bug #11977: that TZ70 is defined as alias of TZ71 in https://github.com/darktable-org/rawspeed/blob/develop/data/cameras.xml (line 7020).

My noise profiles (submitted as TZ71) were accepted in this bug.

Someone on dpreview (https://www.dpreview.com/forums/post/56139467) has compared the specs for the cameras and not found a difference.

I've applied the lens correction profiles for the TZ71 to a selection of my pictures and they seem like a perfect match.

The question still remains, why does my camera match to TZ71 in denoise (profiled) but not in lens correction? That is inconsistent behaviour and therefore a bug with one of those two modules.

#10 Updated by Roman Lebedev over 1 year ago

Chris Elston wrote:

All I can tell you is what I was informed in response to my bug #11977: that TZ70 is defined as alias of TZ71 in https://github.com/darktable-org/rawspeed/blob/develop/data/cameras.xml (line 7020).

My noise profiles (submitted as TZ71) were accepted in this bug.

Someone on dpreview (https://www.dpreview.com/forums/post/56139467) has compared the specs for the cameras and not found a difference.

I've applied the lens correction profiles for the TZ71 to a selection of my pictures and they seem like a perfect match.

The question still remains, why does my camera match to TZ71 in denoise (profiled) but not in lens correction? That is inconsistent behaviour and therefore a bug with one of those two modules.

In other words you are not reading what i'm writing, and you do not have the link i was asking you to show me

The TZ71 data in the lensfun project is data for the TZ70.

Please show me where in the lensfun there is data for TZ70.

It's listed under model TZ71.

Because lens correction module is using EXIF camera name, which is TZ70.

Lensfun only has data for TZ71, therefore either they need to add an alias from TZ71 to TZ70, or add the TZ70 calibrations (provided by you or some other owner). That is how it worked for all the other cameras in lensfun, always.
Report that ^ to lensfun.

#11 Updated by Chris Elston over 1 year ago

I take your point but I still maintain it would be better if the various darktable modules behaved consistently with regard to TZ70/71 models.

The denoise (profiled) module is not matching on exif camera name (but its alias), whereas the lens correction model is matching directly to the exif camera name.

This is inconsistent behaviour but since I now know that I can edit my own lens profiles to get the result I wanted I'm not going to press the point as you clearly don't see this as an issue with Darktable.

I'm merely trying to contribute to this project with the information I have to hand and I'm not sure why this needed to be such a terse confrontational conversation.

#12 Updated by Roman Lebedev over 1 year ago

Chris Elston wrote:

I take your point but I still maintain it would be better if the various darktable modules behaved consistently with regard to TZ70/71 models.

The denoise (profiled) module is not matching on exif camera name (but its alias), whereas the lens correction model is matching directly to the exif camera name.

Correct.

This is inconsistent behavior

True.

This camera deduplication you are seeing is a horrible mess. It should not have happened in the first place.
In "denoise profiled" case, all the data is internal to dt, so we kind-of can make it work.
But that is not true for lensfun. Practically everything else will query lensfun with exif camera names (like we do now), so that will likely to work for everyone, or not at all.

but since I now know that I can edit my own lens profiles to get the result I wanted I'm not going to press the point as you clearly don't see this as an issue with Darktable.

I'm merely trying to contribute to this project with the information I have to hand and I'm not sure why this needed to be such a terse confrontational conversation.

Thank you.

#13 Updated by Chris Elston over 1 year ago

FYI I have raised this upstream with the lensfun project.

#14 Updated by Roman Lebedev over 1 year ago

  • Target version set to 2.6.0

#15 Updated by Pär Lydmark about 1 year ago

Would this upcoming change destined fpr 2.6.0 also resolve similar issues with other models, or is this specific for TZ70/71? I have exactly the same issue with Panasonic DMC-TZ60, which in darktable is listed as DMC-TZ61 giving the same issues as with TZ70/71 in this thread. If that also could be resolved with this fix it would be great. Otherwise, is there anything I can do to make a TZ60/61 fix happen?

#16 Updated by Roman Lebedev about 1 year ago

Pär Lydmark wrote:

Would this upcoming change destined fpr 2.6.0 also resolve similar issues with other models, or is this specific for TZ70/71? I have exactly the same issue with Panasonic DMC-TZ60, which in darktable is listed as DMC-TZ61 giving the same issues as with TZ70/71 in this thread. If that also could be resolved with this fix it would be great. Otherwise, is there anything I can do to make a TZ60/61 fix happen?

I'm not sure about which issue specifically you are talking about,
but if it is related to the lens correction profile, then like it was said in this issue, please report that upstream, to lensfun.

#17 Updated by Chris Elston about 1 year ago

Pär Lydmark wrote:

Would this upcoming change destined fpr 2.6.0 also resolve similar issues with other models, or is this specific for TZ70/71? I have exactly the same issue with Panasonic DMC-TZ60, which in darktable is listed as DMC-TZ61 giving the same issues as with TZ70/71 in this thread. If that also could be resolved with this fix it would be great. Otherwise, is there anything I can do to make a TZ60/61 fix happen?

In the meantime it's pretty easy to make the change locally. Just run ‘lensfun-update-data’ to get a copy of the latest database installed (it'll tell you where it's put the new files), then you can edit the generated xml files directly - just rename the TZ61 entries to read TZ60 instead (or duplicate the TZ61 entry) within the compact-panasonic.xml file.

Also available in: Atom PDF

Go to top