Vignetting correction

Dear list,
I have captured some images when the aperture is f/11. And I used the Canon EOS 6D with a SIGMA 8mm f/3.5 EX DG fisheye lens. However, the experiment to get vignetting curve for the lens is rigorous and complicated. I would really appreciate it if anybody can share the vignetting correction function of the same lens. Any suggestion is OK. Thank you !

Dear Lee,

Although the vignetting curves for two lenses of a similar brand and model are believed to be interchangeable, I would still recommend checking it for a few angles at least, and the aperture of interest (in your case f/11). Using the rotation method as described in Pierson et al, 2019 is not so complicated and does not require very specific equipment.

As I have used the same lens as you, I can share the vignetting curve that I got for the f/11 aperture:!AmPLn6dHOfNI8fEiSKpVjw0ApgwGnw?e=j2Luii

Best of luck,


Hi Clotilde ,
Thank you ! Could you please provide more details about the paper ? I think the complete name is better as I haven’t find it until now.

Hi again Lee,

I am talking about this paper:
Pierson C, Cauwerts C, Bodart M, Wienold J. Tutorial: luminance maps for daylighting studies from high dynamic range photography. LEUKOS 2020. DOI: 10.1080/15502724.2019.1684319.
I think you will find answers to most of your recent questions in there.

Also, regarding my answer for the vignetting process, Jan made me notice that you are using a Canon EOS 6D (and I was using a 5D Mark II). I don’t think this will cause a change in the vignetting curve that I derived since the lens is the most important here, but I haven’t tested that and cannot be 100% sure. So I would recommend you derive your own vignetting curve (and if you can compare with the one I provided and comment on whether they are similar, this would be great!).

Best of luck!

Hi Clotilde ,
Thanks a lot ! Actually I have learned a lot from this paper.

I have several relative questions about the experiment:

  1. Performing the experiment in the daylight room or electrical light one, which one is better? Cause that paper suggested keeping a stable lighting environment.
  2. How to find the coordinates of the center of the fisheye view? Cause the center of the image is probably not the true center.
    someone has noticed this question:

Hi Lee,

have a read through my presentation at the 2012 Radiance workshop. It should answer your question 2.

you should carry out such calibration exercises under artificial lighting. However, please be aware that many LED lights flicker. This is also touched upon in my presentation.

As Axel mentioned, you should not do the vignetting calibration under daylight. In the paper, one option is to do it under an artificial sky (which is a lab environment where we try to simulate an overcast sky through electric lighting), the other is in a dark room with an electric light source (and this light source, as Axel mentioned as well, should not have flicker that could impact your measurements).

Thanks a lot! :heart:
I am going to make preparations for the experiment.

Hi Axel,
I have read your presentation. It is almost impossible to find some common LED lights without flicker.
Will flicker cause significant error if I choose the lamp with the Mod% < 0.1%? How to solve this problem or correct its influence?

LEDs are normally driven with PWM (pulse width modulation) drivers to allow for dimming. What you want is a DC driver. They are very uncommon in luminaires. However, you can improvise with a bench top power supply. They start at around $150.

With LEDs, you need to limit the current. The required mA (milli Amps) are printed on the driver that the LED fitting comes with. This is what you need to set the bench top power supply to.

Hi Axel,
Thanks for your suggestion. I have an idea that using the power bank and USB lamps to avoid flicker. Is this method OK? If this way can’t work, I plan to purchase a transformer of 220V(AC) to 12V(DC) and lamps of 12V(DC) .How about the second way?

I guess you’ll just have to try it out and see if it works. The USB option would work, but probably not give you much output. Yes, you can also use 12V lamps (I guess you mean old-fashioned tungsten lamps), but you would need some kind of housing. Also I don’t know how clean the DC output from cheap transformers is. It is entirely possible that you still get some form of AC ripple.

So just try it out. It would be good if you can ask an electronics person to cobble together a flicker meter that you can use to test your kit. You’d need a digital oscilloscope and some kind of light sensor, which can as simple as an LED. You can operate LEDs in PV mode, so that the don’t emit light, but rather act as light sensors.

Hi Axel,
Thanks for your reply. I have purchased some DC lamp tubes to link with the transformer. I am waiting for some applications to carry out my experiment.