5D Mark III Horizontal Banding in stills?

[chase]

Does anybody know how to fix this problem, or even know if it's a sensor or lens problem? I was shooting a show last night and noticed these stripes about halfway through. I was shooting at 4000 ISO, 1/1600 and the stripes were extremely prominent, so I thought it was the shutter speed. I slowed down to 1/640 and got the same result. I was also using a 70-200L IS II USM. The first image was at 1/1600.. The second image was at 1/1000, and the third image was at 1/640. Any help is greatly appreciated. 

[RobertTheFat]

I believe the shutter on a 5D3 moves vertically, and at those speeds it moves as a slit, with the closing curtain trailing the opening curtain by less than the height of the frame. If the concert was using high-speed strobe lights of different colors, you'd see a band of whatever color light fired when the slit was in position to record it.

[chase]

Thanks, any tips to avoid this? And it's not a camera problem?

[kvbarkley]

Lower the shutter speed to the maximum flash sync speed which is basically the same problem. 

My T6S has flicker prevention, if your camera has it, you might try that, too. Though it is geared for 50/60 Hz flicker.

[chase]

That's only 1/200, and shooting a concert just won't do for that. I had no problem shooting a show at the house of blues about a month ago using higher shutter speeds and I assume they are using the same lights. I thought maybe it was a sensor damage problem, or a lens problem.

[RobertTheFat]

---

@chase wrote:
Thanks, any tips to avoid this? And it's not a camera problem?

---

Well, it's a camera problem in that it's a consequence of the way a focal plane shutter works. You can minimize the effect by using a slower shutter speed; but if the color of the ambient light is changing rapidly relative to the speed that the shutter curtains travel, there can still be a difference in color between the top and bottom of the frame.

[chase]

That makes sense if it's just a gradient color difference from top to bottom, but the fact that there's bands in between, makes me think something different. Color, black, color black. I don't think the strobes were changing that rapidly.

[RobertTheFat]

---

@chase wrote:
That makes sense if it's just a gradient color difference from top to bottom, but the fact that there's bands in between, makes me think something different. Color, black, color black. I don't think the strobes were changing that rapidly.

---

It's not just the rapidity of change; it's also the duration of the flash, which can easily be 1/1000 second or less. You may not notice how fast it is, because of the relatively slow recycle time of the human eye, but the camera notices. If the shutter slit reaches a given position and no light fires while it's there, you'll see a black bar.

[chase]

Thanks so much! Still a little difficult to grasp my head around, but I'll make do. Thanks again.

[Waddizzle · Accepted solution]

---

@RobertTheFat wrote:

I believe the shutter on a 5D3 moves vertically, and at those speeds it moves as a slit, with the closing curtain trailing the opening curtain by less than the height of the frame. If the concert was using high-speed strobe lights of different colors, you'd see a band of whatever color light fired when the slit was in position to record it.

---

I am inclined to agree with Bob's gut instincts.  Solid state lighting sources do not emit light at a continous frequency.  Instead, they emit light strobocopically, meaning the fixture will quickly flash different colors, to achieve an average that the human eye will see as a constant color.  For example, an LED can quickly strobe between red and green to yield what the eye percieves as yellow.  Unfortunately, digital DSLRs do not allow themselves to be as easily fooled as the human brain.