Unexpected results when editing / saving / exporting metadata ?

Software ART
1

I’ve noticed something and again I don’t know if it’s a bug, expected behavior or if I’m just an idiot…

I have a manual Ricoh 50mm lens so of course it doesn’t provide metadata. In ART I typically edit the Lens Model, FNumber and Focal Length tags to add the missing data, e.g.:

image

Today I took a series of images, one at each f/stop the lens offers: 1.7, 2.8, 4., 5.6, 8, 11 & 16. In ART I manually entered those values in the FNumber field for each image. However, ART changed what I entered for f/8, 11 & 16:

Image I enter ART shows Export JPG
IMG_0964.CR3       F1.7     F1.7     1.7
IMG_0965.CR3       F2.8     F2.8     2.8
IMG_0966.CR3       F4.0     F4     4.0
IMG_0967.CR3       F5.6     F5.6     5.6
IMG_0968.CR3       F8.0     F6     6.0
IMG_0969.CR3       F11     F6.9     6.9
IMG_0970.CR3       F16     F8     8.0

The sidecar tag values weren’t consistent…

C:\data\images\Our-photos\20231230_Red_River_Park>grep FNumber= *.CR3.arp
IMG_0964.CR3.arp:FNumber=17/10
IMG_0965.CR3.arp:FNumber=28/10
IMG_0966.CR3.arp:FNumber=F4
IMG_0967.CR3.arp:FNumber=56/10
IMG_0968.CR3.arp:FNumber=F8
IMG_0969.CR3.arp:FNumber=F11
IMG_0970.CR3.arp:FNumber=F16

…so I manually edited the FNumber field in the sidecars to make them all “n/10” format but it didn’t help.

Interestingly, when I disabled the FNumber field checkbox in ART for the last three images, they exported rotated 90 degrees counter-clockwise (at least as seen in XnView MP). I rechecked them and they’re back upright again.

Bug(s) or am I once again just missing something? :slight_smile: Raws and sidecars can be downloaded (zipped) here: https://drive.google.com/file/d/1Rz7gADSSS-9AoNW0wb15LNgZvelsKhU1/view?usp=sharing

ART info

Version: 1.21-17-g20159fbe0
Branch: master
Commit: 20159fbe0
Commit date: 2023-12-22
Compiler: cc 13.2.0
Processor: Znver2
System: Windows
Bit depth: 64
Gtkmm: 3.24.8
Lensfun: 0.3.95.custom
Exiv2: 0.28.1.custom
LCMS2: 2.16.custom-fastfloat
LibRaw: 0.21.1.custom
OpenColorIO: 2.3.0
CTL interpreter: version 1
Build type: release
Build flags:   -std=c++11 -ffp-contract=off -march=znver2 -Werror=unused-label -fno-math-errno -Wl,--stack,4194304 -Wall -Wuninitialized -Wno-deprecated-declarations -Wno-unused-result -fopenmp -Werror=unknown-pragmas -Wno-aggressive-loop-optimizations -DNDEBUG  -O3 -ftree-vectorize
Link flags:   -march=znver2 -s -O3
OpenMP support: ON
Mi-malloc: V2.0
Build OS: Windows
Build date: 2023-12-22T14:13:01Z

Thanks.

2

Odd indeed… I’ll take a look, thanks for the report!

3

Hi again,

It’s a glitch of the editor, which doesn’t perform the necessary validation. I suppose you literally entered “F11” in the text box, but this is an invalid value for the FNumber, which must be actually a number :slight_smile:
So, for the time being, try with simply “11” and it should work. In the meantime I’ll see if I can make the editor a bit more robust and at least report an error instead of silently failing…

4

I thought I had tried entering just the bare number but since you said that works, apparently I didn’t because I never made it work correctly.

Thanks.

5

I just pushed some enhancements to the input validators, it should work better with the current master

1 Like