9

I updated my firmware both C and L on my D810 and D610 today and notices that photos taken right after each other are now not exposing the same. I checked I do not have bracketing on, I added more light to the seen and that does help. Does anyone know what could make this happen?

Update: I am noticing now if I move the camera to somewhere else, like pan then the images turn out exposed correctly but if just fire off a bunch in one spot you get the uneven lighting as shown below in the photos. I also got out my D7000 to see if it is just the lighting or something environmental and it does not have the problem.

Photo settings f/1.4, 1/400 sec, ISO 2500

Lighting Light is LED ceiling light. (Can be seen in tv reflection.)

enter image description here enter image description here enter image description here enter image description here

2
  • Your welcome, thanks for asking them, is very important info that I left out. Nov 18, 2017 at 2:57
  • @scottbb With your questions, it made me think more about it and it seems that it is shutter speed. If I lower things to ISO 800 at 1/125 sec the problems goes a way. Nov 18, 2017 at 3:01

2 Answers 2

18

The problem is your shutter speed is too fast for the type of lighting you have. Most LED lights flicker (and fluorescents, in a slightly different manner and degree) at either 50/60 Hz, or 100/120 Hz, depending on the mains frequency where you live.

At a shutter speed of ¹⁄₄₀₀ second, it's a bit of a crapshoot whether your shutter will be timed with the peak of the light's output. For instance, in your 3rd and 4th photos, look at the TV's shadow on the wall. In the 3rd photo, the light was increasing in intensity as the shutter traveled across the image sensor (remember, the image enters the camera from the lens upside-down, so the bottom of the photo was the first to be exposed). The top of the photo is clearly brighter — therefore as the shutter traveled from the beginning of the exposure to the end of the exposure, the lighting became brighter.

In the 4th photo, the lighting was decreasing in its flicker pattern. Again, the bottom of the photo is brighter, which is counter-intuitive for lighting coming from the ceiling. The shutter exposed the top of the sensor (corresponding to the bottom of the image) first, and as the shutter traveled, the ambient lighting was falling.

Solution: slow down your shutter speed to something like ¹⁄₂₅ or ¹⁄₃₀ of a second (¹⁄₂₅ if you have 50 Hz AC mains frequency, ¹⁄₃₀ if your AC mains are 60 Hz). This will allow for two or four complete flicker cycles, smoothing out the lighting. Of course, you'll have to compensate by either decreasing your aperture or decreasing your ISO by 4 stops, or some combination of both to account for the +4 stops of shutter exposure.

You could expose at ¹⁄₅₀ / ¹⁄₆₀, which would allow for either one or two full cycles (again, depending on the specific type of LED light you have). And of course, your total aperture/ISO compensation would only be by 3 stops (relative to your settings when you shot at ¹⁄₄₀₀).

You could also do some testing to determine if your flicker is at mains frequency (50/60 Hz), or at twice mains (100/120 Hz), by taking lots of shots at ¹⁄₁₀₀ / ¹⁄₁₂₀. If you never see any variable lighting when shooting at that rate, your LEDs probably flicker at 100/120 Hz. If you do see some darker photos, then they probably flicker at 50/60 Hz, so you should drop your shutter speed to allow for at least one full cycle of the lighting.

At least 2 full cycles is probably better, to ensure much more consistent and predictable results. The camera's shutter timing isn't necessarily so precise as the mains frequency.


Most of the following questions pertain to fluorescent lighting, but many of the answers cover much of the same territory, and and ultimately boil down to the same reason — flicker due to AC line frequency.

Related questions:

Also at electronics.stackexchange.com:

8
  • 2
    WOW! that makes perfect sense now, and you gave a great response, thank you for taking the time to help me out. I think I was just getting way to nervous about the firmware update, never have done that before and well these two bodies are not cheap and just so happened that I update the firmware on the two cameras today and figure that was it! Thanks again. Nov 18, 2017 at 3:10
  • I'm not convinced that shooting at 1/30s would alleviate the running shutter problem. I would use fill-flash instead, replacing the existing light almost completely. This would be the preferred solution anyway since the TV screen would otherwise be too bright compared to the room walls. And with a double flash, the metering would be on the spot regardless of the flicker changing between the metering and the main exposure itself.
    – Zdenek
    Nov 18, 2017 at 14:42
  • 2
    @Zdenek I’m pretty sure the point of the question is not to achieve the actual shot of the TV/wall. It’s merely a demonstration of a suspected problem. Flash merely masks the underlying issue, which is having a shutter speed too fast for the duty cycle of non-constant illumination. Assuming OP is in the US or Canada, then shooting at 1/30 or slower will guarantee that the LED light will have completed at least 2 (perhaps 4, depending on rectification or alternating LED substring powering) full AC waveforms will have powered the LEDs. Thus, over that time, the RMS (i.e., “average”) LED...
    – scottbb
    Nov 18, 2017 at 15:49
  • @Zdenek ... illumination, just as our eyes see it, will have been captured by the camera. As is always the case in sampling and integration, the longer the period of integration (i.e., capturing the light), the better the average will be. But of course, as is always the case in photography, we might not have the luxury of keeping the shutter open for the “ideal” amount of time. There are always trade offs.
    – scottbb
    Nov 18, 2017 at 15:53
  • 1
    @Zdenek If one replaces the ambient light almost completely, then that isn't really fill flash. The problem of inconsistent exposure would still remain because TTL (if that is what you mean by 'double flash?') usually uses a very low power pulse for metering so the variation in the ambient lights would still have influence over initial ambient metering and subsequent flash metering.
    – Michael C
    Nov 19, 2017 at 0:51
3

(While this is not the most likely explanation for the examples posted in the question, which scottbb covered very well, someone else who finds this question and has a similar problem - changes in exposure after a firmware update - may be shooting under more constant lighting and find the following to be applicable.)

When you update firmware with most cameras all of the options are returned to factory defaults.

This includes various metering options. Some metering modes may place more emphasis on certain areas in the frame, while other metering modes use a library based model to decide how to weight different areas of the scene. If you had the metering and exposure options set to something other than the factory defaults before your firmware updates, those options were almost certainly reset to the factory default which could explain some of the difference you are seeing. This could affect how your camera is metering the scene.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.