18

Yesterday I took photos of an event at school.

The lighting was poor so I increased the iso to 3200, shutter speed was between 80 and 250 and the aperture between 2.8 and 4.0. Canon EOS500

about 10% of the photos got this weird horizontal black area. The position of the black area is not the same, Sometimes it is in the middle, sometimes it is in the upper or lower part of the frame

Is this an artifact from the lighting frequency? or can it be a shutter problem?

enter image description here

mattdm
  • 143,140
  • 52
  • 417
  • 741
Nir
  • 283
  • 2
  • 5

2 Answers2

35

The variability you are seeing in your photos is due to the way many types of lighting convert alternating current into light. Although they look constant and steady state to our eyes, in reality they are flickering with the oscillations in the alternating current supplying them with electricity.

When shooting under any kind of flickering lighting, including fluorescent, each frame can have different brightness and color as the lights get brighter and dimmer due to the alternating current powering them. They tend to be bluer and fuller spectrum when at the brightest peak and browner and much more limited spectrum when at the dimmest part of the cycle. If you are using a shutter time shorter than half of the period of the current powering the lights, the color and brightness will change from the top to the bottom of the frame as the slit between the curtains of your focal plane shutter transits across the imaging sensor. The shorter the shutter time, and thus the narrower the slit between the first and second shutter curtains, the more pronounced this effect will be from one side of the frame to the other. Even with an electronic shutter you will see the effect with CMOS sensors, which scan across a sensor sequentially.

There are a few ILCs (interchangeable lens cameras) now on the market that use the light meter to detect the timing of flickering lights and time the shutter to open when the lights are peaking. This allows the photo to capture the image as the light is at both its brightest and fullest spectrum. Since the shutter opens at the brightest point in the cycle, it allows shorter shutter times for the same ISO and aperture settings. This can be quite an advantage when shooting sports under flickering lights. It also allows more uniformity of color from one frame to the next which simplifies the post processing workload.

The accepted answer to Lots of noise in my hockey pictures. What am I doing wrong? gives a few tips on how to shoot, and process, sports taken under such flickering lighting.

It's not sports or indoors, but Should I use a 24-105 or a 70-200 lens for photographing a high school marching band? covers basically the same thing: shooting action under relatively dim flickering lighting. The accepted answer delves into a lot besides just lens selection.

For more related reading here at Photography SE, please see:

In my answer to When should I upgrade my camera body? I use flickering lights as a rather extensive case study to illustrate the process that applies to the main question.
How to edit photos shot in fluorescent light
Is there a low light picture quality difference between 50mm 1.8 G and 85mm 1.8 G Nikon? (The answers reveal the difference was in the inconsistent lighting, not between two lenses tested under flickering light in a camera store.)
This answer to Color matching Product

Michael C
  • 175,039
  • 10
  • 209
  • 561
  • 2
    Do fluorescent lighting and shutter speed create a problem with color cast? is a good example of this as well. – Philip Kendall Jun 19 '18 at 12:09
  • 2
    Not just DSLRs — Fujifilm just added flicker reduction to several of its cameras in the latest round of firmware updates. – mattdm Jun 19 '18 at 12:36
  • It's worth noting that in most countries electricity uses 50 Hz, so shutter speeds slower than 50 should be free of this problem - but of course less useful on children and not useful in sports. – Mołot Jun 19 '18 at 13:19
  • @Mołot It is true that most countries use 50Hz, but it is also true that a majority of ILCs are sold in countries that use 60Hz (Primarily North and Central America, the most populous countries in South America {Brazil + Columbia}, most of the Caribbean, Taiwan, and Japan). The U.S. alone accounts for almost half of all ILC sales. – Michael C Jun 20 '18 at 04:26
  • @mattdm For stills? Or video only? Nikon has had a video only form of flicker reduction for years, but not for shooting stills. (Although it is hard to determine that only from reading Nikon's marketing materials.) – Michael C Jun 20 '18 at 04:28
  • I haven't really tested it, but from the description, stills. "For enhancing the quality of indoor sports photography, the upgrade allows users to reduce flicker in pictures and the display when shooting under fluorescent lighting and other similar light sources." – mattdm Jun 20 '18 at 06:37
  • Cool. Now if they could just develop decent continuous AF tracking. – Michael C Jun 20 '18 at 06:39
  • But, if the current frequency is 50/60 Hz, shouldn't the lamps flicker with twice that (100/120 Hz)? The usual way to get direct current from alternating starts with a diode bridge that "flips" half of the incoming sine wave to generate a pulsing DC at double the frequency of the incoming current. – remco Jun 20 '18 at 07:22
  • @remco Please carefully read this line from the answer: "If you are using a shutter time shorter than half of the frequency of the current powering the lights..." A shutter time half the frequency of 50Hz is 1/100 seconds. A shutter time half the frequency of 60Hz is 1/125 seconds. – Michael C Jun 20 '18 at 10:20
  • @MichaelClark that probably ought to be more precisely worded as "if you are using a shutter time shorter than half of the period of the current powering the lights...". Times and periods are in seconds, frequencies are in units of 1/sec. – scottbb Jun 21 '18 at 20:15
  • @scottbb Point taken, but Hz are expressed in frequencies of 1/sec. Frequencies can be either. Very low frequency waves can have a period of 20 (as in one cycle per 20 seconds) or 20 Hz (as in one cycle per 1/20 second). – Michael C Jun 22 '18 at 01:33
  • @MichaelClark I don't generally do photography where focus tracking is import, but it's my understanding that firmware updates have substantially improved that as well. I don't know if it's up to pat with Canon / Nikon higher-end models, though. – mattdm Jun 22 '18 at 02:00
  • I've not seen anything about Fuji's ability to do subject tracking. Since the introduction of the 1D X and D3 (the first flagship FF "sports" cameras from Canon and Nikon, respectively), very few top level sports/action photogs shoot with anything other than FF. Some PJs for smaller publications, like small town newspapers, do. But most of the newspapers in my area are all-FF. The word on the FF Sony α9 is generally "much improved over previous Sony implementations, but not yet near the level of Canon/Nikon." Prior to the D5 Nikon was also a step behind Canon WRT continuous AF tracking. – Michael C Jun 22 '18 at 02:23
0

Any speed that is not slow enough to get a full cycle will get you into trouble.

Remember that a full cycle is the time that takes sin wave of the ac to go from peak to peak, and the frecuency is the number of times the sin wave ac current goes from it’s peak to it’s lowest value per second.

Easy it takes twice from peak to peak than from peak to it’s lowest value. So for 50hz (60hz) countries speeds over 1/100 (1/120) may get you into trouble when using faster speeds in led environments while not using strobes or studio lights.

The easy way get a couple of strobes.

abetancort
  • 438
  • 2
  • 7