lensfun / lensfun Goto Github PK
View Code? Open in Web Editor NEWThe Lensfun project Git repository
Home Page: http://lensfun.github.io
License: GNU Lesser General Public License v3.0
The Lensfun project Git repository
Home Page: http://lensfun.github.io
License: GNU Lesser General Public License v3.0
Hi guys,
lensfun 0.3.1 fails to build for me:
https://travis-ci.org/neothemachine/lensfunpy/jobs/62770917#L173
What's wrong here?
Cheers
Maik
Reported by: neothemachine
I used hugin lens calibrator.
Reported by: milan_knizek
Attached is my calibration data for the Sony Cyber-shot DSC-RX100 compact camera. It includes:
Reported by: tgessler
Reported by: sillyxone
Distortion-only calibration created using Hugin method here: http://wilson.bronger.org/lens_calibration_tutorial/
Shots taken about 4 meters away due to the ultra-wide nature of the lens. Photos were also uploaded to http://wilson.bronger.org/calibration.
Reported by: rwh86
Hi.
I have a Sony A6000 and a Sigma 19mm 2.8 EX DN lens (the OG model).
According to the LensFun coverage page:
http://wilson.bronger.org/lensfun_coverage.html
... support is present for the lens.
Using darktable 1.6.2 and liblensfun 0.2.8, the camera/lens combo was not recognized at all from the RAW metadata.
I manually updated to liblensfun 0.3 and camera/lens auto detect would not work. I could manually select the camera/lens combo, though.
I noticed that the RAW EXIF data shows:
"E 19mm F2.8"
... as the lens model, but the liblensfun xml file shows:
"Sigma 19mm f/2.8 EX DN"
... as the lens model.
Relevant snippet:
<lens>
<maker>Sigma</maker>
<model>Sigma 19mm f/2.8 EX DN</model>
<mount>Sony E</mount>
<mount>Micro 4/3 System</mount>
<cropfactor>1.534</cropfactor>
<calibration>
<!-- Taken with Sony A6000 -->
<distortion model="poly3" focal="19" k1="-0.00748" />
<tca model="poly3" focal="19" br="-0.0000466" vr="1.0003817" bb="0.0000346" vb="1.0000194" />
</calibration>
</lens>
Anyways, I opened a ticket on darktable redmine, lots more detail there:
http://www.darktable.org/redmine/issues/10331
They pretty much told me it's not a darktable bug, so I'm trying my luck here.
Basically, how can I force detection of the lens model?
Or is there nothing else to do on LensFun's side?
Thanks for your time.
Regards.
Reported by: rdeleon
I would like to add vignetting data for two lenses:
Olympus M.Zuiko Digital 17mm f/1.8
Lumix G Vario 14-45mm f/3.5-5.6 Asph. Mega OIS
Their distortion and tca data is taken from the current git. Only the vignetting data in the attached file is new.
Reported by: scheurer
This small patch adds the D750 and D810 to the Nikon SLR database.
Reported by: horsepunchkid
At least in one case the auto-scale that LF reports is wrong: If you want to remap a fisheye image to equirectangular. It still leaves dark corners.
Currently, Lensfun scans only the positions of the corders and the edge centres in order to find the proper scaling. This is too few points. I must scan at least 10 points or so per edge.
Alternatively, it scans 5 points per edge, finds the point for which the biggest scaling is necessary, takes the neighbour with the second biggest scaling, and uses a real maximisation algorithm between them to find the optimal scaling.
By the way, the routine should also take into account subpixel distortion. Otherwise, only the autoscaling for the green channel is found. However, the red or blue channel may be smaller.
Reported by: bronger
I'm trying to compile with MSVC on Windows 7 64 bit and have the following remarks/problems:
The README says "You will need Python (native Unix python will work fine if cross-compiling), glib binary and devel packages for Windows (download it from gtk homepage: http://www.gtk.org/download-windows.html) and the MSVC compiler." -> You also need GNU Make which can be found at http://gnuwin32.sourceforge.net/packages/make.htm.
It's not really clear what to do with the downloaded glib binaries/dev packages. According to the configure script it seems like both archives should be extracted into "./glib-2.0" but that didn't satisfy the configure script ("Checking for library glib-2.0 >= 2.0 ...NOT FOUND").
After commenting out the glib checks, the configure script ran through but then make failed:
C:\..\lensfun-0.2.8>make libs
mkdir out/windows.x86/release
Syntaxfehler.
make: *** [out/windows.x86/release] Fehler 1
...which makes sense as it should be "mkdir out\windows.x86\release".
When running configure with --target=windows.x86_64 (instead of x86) then "make libs" says: "build/tibs/rules.mak:208: *** Only the X86 architecture is currently supported by the MSVC toolkit!." Does this mean I cannot compile for 64 bit? Why not? I am on a 64 bit Windows so it should work.
Reported by: neothemachine
Please apply to git master
Reported by: edgomez
Please add the following lens (Exact string!) to your database:
Nikon AF-S DX VR Zoom-Nikkor 18-200mm f/3.5-5.6G IF-ED
There is already this lens but here is the story:
There exist two versions of this lens the one with the string I passed here and the one with the string already contained in your database. It's exactly the same lens. The one you entered in your database is the string registered in image tags of the 2009 (and forward) versions that simply have a lock button to avoid zoom creep. On wikipedia you can confirm what I'm talking about here.
http://en.wikipedia.org/wiki/Nikon_AF-S_DX_VR_Zoom-Nikkor_18-200mm_f/3.5-5.6G_IF-ED
Thanks a lot.
P.S.: I Already added the string copying all correction parameters from the original one and worked like a charm :)
Reported by: zazy
Currently, everything printed to stderr and stdout by lensfun is not marked as such. This should be added so that people see that the message is not generated by the lensfun-using program but by lensfun itself.
Reported by: bronger
I'm having trouble compiling the latest lensfun version 0.3 under MSVC 2008. According to the site https://connect.microsoft.com/VisualStudio/feedback/details/783412/isnan-and-isfinite-support-does-not-exist-in-math-h-or-cmath it exists only since the 2013 version.
However there exists the equivalent function _finite: http://msdn.microsoft.com/en-us/library/sb8es7a8%28v=vs.100%29.aspx So maybe this could be fixed by conditionally defining #define isfinite(a) (_finite(a))
Reported by: neothemachine
Hi, I noticed that the Nikon lens "Nikkor AF-S 50mm f/1.8G" has an incorrect crop factor of 1.523 - this not a "DX" lens, i.e. crop factor should be 1.0.
Thanks.
Reported by: wolfgangoertl
Hi.
I tried profiling my Nikon Coolpix S3300. But please verify the attached XML agains the photos carefully -> I completely relied on Hugin's automatic calibration as I couldn't find the manual feature on my Hugin release :-(
Find the original Photos I used for calibration on http://www.andreas-weller.de/download/Nikon-S3300.tar.bz2 (43 MB)
XML attached!
Thank you and regards,
Andreas Weller
Reported by: wellera
Hi,
I have added the code below to the file "mil-sony.xml" to add support for the camera "Sony A7M2 " and the lens: "Sony FE 55mm F1.8 ZA". Please consider using the code. Thanks.
Sony ILCE-7M2 Alpha 7M2 Sony E 1....................
Sony FE 55mm F1.8 ZA Sony E 1Reported by: kwokyinc
Hello,
current lensfun doesn't compile properly with Clang 3.4.2 on Gentoo due to symbol visibility problems.
Clang implements visibility modifiers identical to GCC.
I have created a patch to allow for clang compilation by triggering the same #defines used for GCC.
The patch works correctly, lensfun compiles. Since there is no testsuite, I compared the symbols exported from a clang-compiled lib and a GCC-compiled one, and they match.
Please include the patch in the next lensfun release.
This is the original Gentoo bug: https://bugs.gentoo.org/show_bug.cgi?id=519006
Thank you,
Fabio Scaccabarozzi
Reported by: fsvm88
Distortion & vignetting information.
No photozone review yet, so here's lenstip:
http://www.lenstip.com/401.1-Lens_review-Olympus_M.Zuiko_Digital_25_mm_f_1.8_Introduction.html
images available upon request!
Reported by: cchiappa
Calibration data for another cheap lens.
Camera 5D, vignetting with a diffusor
Reported by: rbrnhrdt
I have the LUMIX G VARIO PZ 14-42/F3.5-5.6, a Micro Four Thirds lens, which I have used with the Panasonic GF2 and GX7. I use ufraw, which incorporates Lensfun. About a year ago, I built ufraw from source, because I was having trouble with the version packaged with Fedora. When I got the GX7, I found that the now-old version of ufraw wasn't working right with it, and I needed to upgrade. When I did, I started seeing a weird bug: an image which is mostly good, but has many pixels set to bright blue, red, or magenta. By trial and error, I managed to establish that the bug showed up when I used the VARIO PZ lens with the GX7 or the GF2, but not when I used a different lens with either camera, or when I used the VARIO PZ lens with --lensfun=none. I formed the hypothesis that the bug was in Lensfun, rather than ufraw proper; I tried updating to Lensfun 0.3.0, but the problem persisted. Instead, I tried dropping back to Lensfun 0.2.7--and the problem went away...but the output of --lensfun=none is identical to that of --lensfun=auto. Looking at the data directory, I find no entry for the VARIO PZ lens.
So...some kind of bug in Lensfun or its data for this lens. I'm attaching the RAW file and (in a separate comment) the PNG output of 0.2.8.
More detail of my trial and error can be found at
https://sourceforge.net/p/ufraw/bugs/393/
which is the bug I originally filed against ufraw, until I realized it wasn't their fault at all.
Reported by: johnstracke2
Please create an entry for:
Sigma 70-300mm f/4-5.6 DG OS
Reported by: linuxcav
Hi,
I'm owning the new NEX, Sony Alpha 6000, I'm using Darktable which use your project.
Sony Alpha 6000 (codename ILCE-6000) seems don't be in your database.
How can I help you to add it ?
Thanks
Reported by: atlas95
Hi,
here is the calibration file for my Cosina 19-35 1:3.5-4.5 MC.
This lens is also labeled as Soligor,Vivitar and Voigländer.
Pictures were taken with a Canon 5d.
Reported by: rbrnhrdt
Sony RX100M3 lens is not supported yet.
I own the camera and if there is anything I can do, I'm willing to help.
Reported by: farhadi
I'm still having trouble compiling 0.3 (with isfinite fix) with MSVC:
lensfun\modifier.cpp(90) : error C2065: 'M_PI' : undeclared identifier
lensfun\modifier.cpp(126) : error C2065: 'NAN' : undeclared identifier
lensfun\modifier.cpp(225) : error C2065: 'NAN' : undeclared identifier
The following pages refer to these issues:
http://tdistler.com/2011/03/24/how-to-define-nan-not-a-number-on-windows
I use VC 2008 but the issues apply also to 2010. I think 2010 should still be supported.
Reported by: neothemachine
Hi,
This is the calibration data for two Nikon lenses on crop factor 1 (Nikon D600 body):
The 50mm was not in the library.
For the 20mm lens, there's an entry for "Nikon 20mm f/2.8 D AF" (both are optically equivalent) but only for 1.5 crop factor.
I calibrated the lens distortion manually with Hugin, and tca with tca_correct.
Cheers!
Nicolas
Reported by: nlann
On my distro, we install arch-specific files into a different prefix (i.e. /usr/x86_64-pc-linux-gnu/) than data files (/usr/share). That requires the ability to set the latter to an absolute path.
The attached patch accounts for this according to
http://www.cmake.org/cmake/help/v3.0/module/GNUInstallDirs.html
Reported by: berniyh
I just calibrate vignetting on my Nikon AF Nikkor 50mm f/1.4D with this tutorial: http://wilson.bronger.org/lens_calibration_tutorial/#id3
So My calibration data in attachment. Please add it to database.
Reported by: sghpunk
This patch adds support for the FUJIFILM X100T camera model. The X100T has the same lens as the X100 and X100S.
Best regards,
Niels Kristian Bech Jensen
Reported by: nkbj
In my distribution (Exherbo) rst2man is installed as rst2man.py which effectively hides it from the build system. The attached patch changes that and instructs cmake to also look for slightly different names.
Reported by: heirecka
I have a Nikon 55-200 lens which is identified as Nikkor 55-200mm f/4-5.6 G ED AF-S DX in lensfun. However in the EXIF data from my images the lens is identified as Nikon AF-S DX VR Zoom-Nikkor 55-200mm f/4-5.6G IF-ED.
Reported by: khilas
Hi !
I've created a lensfun profile for my PENTAX FA 31mm.
The xml file contains distortion profile, vignetting profile for all apertures, and tca profile. I've attached the xml file.
Could you load this profile into lensfun ??
Thank you !
PS: I own PENTAX FA 77mm lens too, so I could make another one for this lens.
Reported by: nikkoblackbear
Fujifilm XC 16-50mm f/3.5-5.6 OIS Vignetting not working on latest GIT-Version, but its listed on http://wilson.bronger.org/lensfun_coverage.html
Reported by: guentir
This patch adds support for four new camera models:
Olympus E-M5 Mark II
Panasonic DMC-GF7
Nikon D7200
Fujifilm X-A2
Best regards,
Niels Kristian Bech Jensen
Reported by: nkbj
Suggested corrections for the second version of LUMIX G 14mm
Distortions are actually extracted from comparison with the camera corrected jpg (and are rather close to those from the 1st version of the lens). TCA are calibrated from a photo (and are quite different from version I in the database).
Vignetting is not calibrated at the moment.
Reported by: fedxa
Hello,
I think the current 'latest' digital Ricoh GR (2013) is not correctly listed in 'compact-ricoh.xml'. I would provide all the information you need, just let me know.
A 12MB DNG from the can be found here:
https://drive.google.com/folderview?id=0B_T1hFpr9CPffkpGOHhUaHdiaFN5b3NNcXUwTUhESGhmWGpqZk5wbE84bGl5Rk15RWtsNGM&usp=sharing
The output of exiv2 -pt *.DNG | grep -i -E 'lens|model|make' looks like:
0023025.DNG Exif.Photo.MakerNote Undefined 63428 (Binary value suppressed)
R0023025.DNG Exif.Image.UniqueCameraModel Ascii 28 GR
R0023025.DNG Exif.Image.LensInfo Rational 4 1830/100 1830/100 28/10 28/10
R0023025.DNG Exif.Image.MakerNoteSafety Short 1 0
R0023026.DNG Exif.Image.Make Ascii 30 RICOH IMAGING COMPANY, LTD.
R0023026.DNG Exif.Image.Model Ascii 64 GR
A Ricoh GR Digital != Ricoh GR (2013).
http://en.wikipedia.org/wiki/Ricoh_GR_digital_cameras
Thank you!
Reported by: dantableau
Proposal from the mailing list:
https://sourceforge.net/p/lensfun/mailman/message/33186260/
We could move the aspect ratio and crop factor into the calibration tag and allow several calibration sections per lens. This could solve the problems of multiple lens entries for different crop factors or aspect ratios and even for RAW/JPEG.
For example:
<lens>
...
<calibration aspect-ratio="4:3" cropfactor="1.536" RAW="true">
<distortion ...>
...
</calibration>
<calibration aspect-ratio="4:3" cropfactor="1.0" RAW="true">
<distortion ...>
...
</calibration>
</lens>
Reported by: seebk
Distortion, tca, and vignetting
Reported by: boeleman1
Here are a few new camera model entries.
Please do carefully verify before committing.
Not all entries were verified against actual cameras, as some aren't yet released.
Reported by: pmjdebruijn
The lookup for the lens "Fotasy M3517 35mm f/1.7" doesn't work. Nothing is found. I suspect the "M3517" is interpreted as a focal length during the search. The source code contains a test for this bug which currently fails.
Reported by: bronger
I have a build warning with gcc 4.8.2.
lensfun-code/tests/test_modifier.cpp:46:147: warning: deprecated conversion from string constant to ‘char*’ [-Wwrite-strings]
char* geom_names [] = {"rectilinear", "panoramic", "equirect", "fisheye", "fisheye-equisolid", "fisheye-orthographic", "fisheye-thoby", NULL};
lensfun-code/tests/test_modifier.cpp:46:147: warning: deprecated conversion from string constant to ‘char*’ [-Wwrite-strings]
Shouldn't string literals be an array of N const char* ?
Find attached a patch, it fixes this and deprecated doxyfile tags.
Reported by: rconstruct
I tried to upload test shots but I guess your server timed out.
Reported by: jamiekitson
Hi,
Since there was no vignetting correction available for the Nikkor AF-S 55-300mm f/4.5-5.6G DX ED VR lens (only distortion and chromatic abberation), I have measured vignetting using a high quality diffusor and my Nikon D90.
I have tested the calibration on various photos and it works nicely.
Please add the vignette correction (see attachment) to the existing calibration of the lens.
Best regards,
Jan
Reported by: madscientist23
Just a reminder, so we won't forget... We have to update the database documentation. The following is missing or not up to date:
Reported by: seebk
Hello,
please add calibration file for Canon Canon EF 70-210mm f/3.5-4.5 USM.
Camera 5D, vignetting with a diffusor
Reported by: rbrnhrdt
Please add configuration for « Fujifilm IS Pro » camera into slr-nikon.xml file (the same than Fujifilm S5 Pro)
Thanks
Reported by: grondin85
XML contains data both from the Hugin calibrate_lens_gui as well as from manual control point selection (1 image each for each focal length in any case). Spot checking, I think the tool is doing a better job here than I have with the manual selection, so I've commented out the manual data.
Photozone: http://www.photozone.de/m43/845-panasonic45150f456
Reported by: cchiappa
The Python scripts for updating the database and adding mount compatibilities should be made usable on Windows and OSX, too.
Reported by: bronger
The Olympus Stylus1's firmware version 2.0 has changed the camera's model name to: "STYLUS1,1s". The camera is currently not part of the lensfun database. I have modified my local db (/usr/share/lensfun/compact-olympus.xml) by copying the existing Stylus1 entry & updating the model entries.
exiv2 -pt P1310584.JPG | grep -Ei 'lens|model|make'
Warning: Directory OlympusCs, entry 0x0101: Strip 0 is outside of the data area; ignored.
Exif.Image.Make Ascii 24 OLYMPUS IMAGING CORP.
Exif.Image.Model Ascii 17 STYLUS1,1s
Exif.Photo.MakerNote Undefined 12984 (Binary value suppressed)
Exif.MakerNote.Offset Long 1 3620
Exif.MakerNote.ByteOrder Ascii 3 II
Exif.OlympusEq.ConversionLens Ascii 32
Exif.OlympusEq.FlashModel Short 1 None
Exif.Photo.LensSpecification Rational 4 0/0 0/0 0/0 0/0
Exif.Photo.LensModel Ascii 32
Reported by: olive22
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.