Recently I was shooting a kids' pinecar derby. I was indoors with no flash and crummy gym lighting, ISO was set at either 2500 or 3200, f/1.8 to 2.8 and a shutter speed of around 500, maybe a little more at times and I had it on continuous high. My problem is that a bunch of the photos came out yellowed, but not all. One or two out of every three, shots where I was shooting continuously and nothing changed, not any of the settings or camera position. At first I thought I'd accidentally left the camera on bracketing but I checked, nope. The weirdest thing is that some of the photos are only yellowish on the top half or bottom half, the other half is fine. I'm wondering it this is a hiccup with the continuous high mode or maybe a memory card issue? Anybody have any insight?
#3. "RE: Weird WB Issue (Maybe)" In response to Reply # 1
Sheri is most likely spot on here.
The problem is twofold. The 1/500th shutter speed with the typical older gym lighting systems is only allowing the camera to capture a small portion of the wave form controlling the lighting. Typically in such conditions anything over 1/120th will cause problems. Also shutter speeds over 1/250 will provide only a slit for which the light will pass through. If you are catching the wave form coming from, or approaching zero, you will see the banding type affect because one portion of the sensor is seeing a different amount of luminance than another. And I believe the difference in luminance is greater neare the zero reference than when the wave form is nearer the peaks.
#7. "RE: Weird WB Issue (Maybe)" In response to Reply # 6 Mon 05-Mar-12 02:44 PM by billD80
Even at 1/100th, with custom White Balance set, the problem will persist, as under fluorescent lighting, the pulsations affecting actual light temperature do not coincide with a given capture in camera.
You can see this easily by shooting a white sign in the gym with the high-speed (Ch) 6 FPS on your camera. Some shots will be spot-on, others purply, etc...
#8. "RE: Weird WB Issue (Maybe)" In response to Reply # 7
Good morning fellows! I find this problem very interesting, and at the same time quite intriguing. But first of all; I'm definitely no expert in fluorescent lighting in any way, but I think I at least have an idea of how they are supposed to work. However, I do get a bit confused about what you say here. I know that this kind of lights do fluctuate at the speed mentioned, at 60 or 50Hz (x2?) depending on which side of the pond you happen to be located, and I know that this flickering could be a real problem while trying to make a video for instance if the equipment is not calibrated to the specific light frequency you have to deal with. What I've been told is that the flickering is partly supposed to be reduced by the inside coating (the white powder) of the strip lights by absorbing the UV light generated by the electronic stream inside going through a gas (argon and some mercury), and then emit the light again to be spread out as visible light, at and in between the high amplitude frequency of the AC power shifting, or let's just say the emittance is a bit delayed to even out the actual on and off situation by the AC power. So what I don't understand is why the white balance problem happen, and hence suffer while shooting, even though it still is not a question even close to a solid light, despite the fact that the inside coating of fluorescents is supposed to even out that phenomena of AC flickering except that it also transforms the UV-beam for us to become visible. I can understand that the luminance is unstable in terms of the relatively short exposure times we are dealing with during photo shots, but what I really don't understand is how this could have an impact on colour temperature (white balance) as the electrons that illuminate the fluorescent inside cover should be the same no matter the actual amount of light emitted at any given time? Or maybe someone could explain if the colour temperature actually really changes by the amount of light emitted at fractions of time due to the AC fluctuations? Most of these strip lights are stated at a specific K (colour temperature) as I know it. Maybe it is just an average number?
I really do not question that there is a real and veritable problem you have experienced, I would just like to know if my understanding is incorrect in terms of how these light tubes really works?
Just as a side point: I have experienced in full manual of my D7000 that the white balance occasionally has seemed to change while doing a series of 6 frames/sek in Tungsten light. Maybe that's for another thread, I just felt I should mention it here BTW. Maybe someone has also observed the same with their D7K?
Sorry for a messy post written by a Swede trying to make a point in a second language, but I would be very happy if someone could explain this phenomena in a way I could understand!
#9. "RE: Weird WB Issue (Maybe)" In response to Reply # 8
Lake Stevens, US
I've encountered the same thing shooting basketball and volleyball at the high school gym. They use big argon gas lights for the lighting in the gym. If I'm shooting continuous high speed shots I get some "off-color" shots during the sequence. I have always attributed it to the lighting decay (flicker) of the 60hz cycle at the zero volts of the scale.
What happens to a filament that has the right voltage to it, it "burns" at the same brightness and color spectrum. Decrease the voltage, the light put out drops and the color shifts down the scale.
45hz is the lowest cyclic rate that most people can't detect. Back in the old days of film theaters, the projectors moved 23 frames worth of film in a second and showed each frame 2 times. This gave a 46hz signal to the eye, so we SEE no flicker. (Most people that is.)
Now, when the shutter is opening and closing 4 to 6 times a second it stands to reason the shutter will open in the zero volt state of the lighting. Or when the lights are nearing zero volts. The filaments aren't at their brightest and the color has a slight shift to it. We can't see this because our eyes aren't able to distinguish it, but don't think for a second that $1100 sensor you're holding to your eye can not only see it, but will capture it as it thinks you want it captured!
Those off color shots can become a huge part of the sequence you shot or might be a rarity, depending on the power systems you have in your country, the fps you're shooting and when you mashed the shutter button in that cyclic rate on the overheads. If you find you are getting alot of off-color at 6fps in the US change the camera to a 5fps Max, or even 4fps. You might miss some action, but your on color hits should improve.
Post on those off-color shots is an at your own risk, from my mileage.
#11. "RE: Weird WB Issue (Maybe)" In response to Reply # 10
if you shoot in raw, you will have a decent chance at adjusting the WB in post.Especially if you have a grey object to use as a adjustment point.In ViewNX2 if I set the WB for Floresent and play with the sliders a little I can get passable results.The 'banded' shots might be harder, depending if the coloration has an effect on your main subject .And with better PP software too ( I only have View)
#12. "RE: Weird WB Issue (Maybe)" In response to Reply # 8
>What I've been told is that the >flickering is partly supposed to be reduced by the inside >coating (the white powder) of the strip lights by absorbing >the UV light generated by the electronic stream inside going >through a gas (argon and some mercury), and then emit the >light again to be spread out as visible light, at and in >between the high amplitude frequency of the AC power shifting, >or let's just say the emittance is a bit delayed to even out >the actual on and off situation by the AC power.
Who ever told you that is incorrect. The phosphor coating when bombarded with the UV light created by the arc fluoresces which produces the visible light. Once the arc stops, the phosphors for the most part go immediately dark (watch a fluorescent light when you flip the switch). It’s the latency of our visual system that makes us believe the blackout does not occur,but it actually does.
In fact many moons ago when as a teen I worked at a TV and radio repair shop, I often used a strobe disc for calibrating the speed of phonograph turntables that took advantage of this flicker. And while the latency of our visual system allows us to believe there is no flicker, we still do detect it. And while it’s controversial, it is purported by some to cause some people to experience health problems because of it.
In another five to ten years this problem for photographers will for the most part be a nostalgic one (at least here in the USA).
#13. "RE: Weird WB Issue (Maybe)" In response to Reply # 9
>I have always attributed it to the >lighting decay (flicker) of the 60hz cycle at the zero volts >of the scale. > >What happens to a filament that has the right voltage to it, >it "burns" at the same brightness and color >spectrum. Decrease the voltage, the light put out drops and >the color shifts down the scale. >
This is true of incandescent lighting which is also subject to flicker from the same sine wave power source, but the residual energy retained by the filament (heat) helps mitigate the effects.
But fluorescent and mercury vapor lighting produce visible light by gas discharge arc, not a glowing filament. Fluorescent tubes do contain a filament. But its primary function is to act as a cathode to emit the electrons for the gas discharge arc, not producing the visible light.
Incandescent lights need a specific voltage applied before they begin to glow, but will continue to glow even at the zero reference due to residual energy when the frequency of the power source is high enough. You can test this out using a dimmer switch on an incandescent bulb. If you turn the dimmer all the way down before you switch on the light the filament will be dark. Begin to slowly raise the dimmer, and you will notice the light stays dark to a certain point, after which it then begins to glow. Then after the glow is achieved, turn the dimmer down and you will notice the bulb stays illuminated well past the point it needed to start to glow. Once the incandescent filament begins to glow, the residual energy left when the power source nears or achieves zero reference mitigates the fluctuation and prevents the flicker (the light still dims a wee bit, but never goes dark allowing the color temperature to remain fairly constant).
Arc discharge lamps on the other hand need a certain amount of voltage to maintain the arc. And that needed amount does not occur for a period of time before and after the zero reference. Unlike the incandescent filament, there is no residual energy left once the arc stops. Once the arc stops, the lamp goes dark. So I believe it is the rise and fall time from the blackout state that creates the problem from color temperature and luminance shift.
Today’s electronic ballasts appear flicker free, for they operate at frequencies of around 10,000 to 20,000 hertz, whereas the older magnetic ballast resonated at the line frequency of 50 to 60 hertz. So while they still flicker, a camera with faster shutter speeds would have less of a problem.
#14. "RE: Weird WB Issue (Maybe)" In response to Reply # 11
I do shoot in RAW and had pretty good luck with adjusting the white balance. Yes, the split ones were a true pain. I had pretty decent luck with using a graduated filter (I use Lightroom) to brighten the warmer half of the photo after doing my best to get the white balance as close to correct as possible. Of the 200 photos I edited luckily only a few were split, a little more than half of the rest were fine and a little less than half were yellowed and had to be fixed. Thank God for Lightroom's lovely sync feature.
#15. "RE: Weird WB Issue (Maybe)" In response to Reply # 10 Sat 10-Mar-12 12:54 AM by kuzzy
A FPS adjustment will not help you with what you most likely encountered. As explained throughout these posts the lights are going to cycle. At a fast enough shutter speed you are capturing only a portion of the lights cycle so the color shift shows up. Whether you take one shot or 6 FPS it really does not matter simply because our eyes cannot see the shift and you have no way of knowing where in the cycle you are when you press the shutter.
Marc There are always two people in every picture: the photographer and the viewer.-Ansel Adams
#16. "RE: Weird WB Issue (Maybe)" In response to Reply # 12
>>What I've been told is that the >>flickering is partly supposed to be reduced by the inside >>coating (the white powder) of the strip lights by >absorbing >>the UV light generated by the electronic stream inside >going >>through a gas (argon and some mercury), and then emit the >>light again to be spread out as visible light, at and in >>between the high amplitude frequency of the AC power >shifting, >>or let's just say the emittance is a bit delayed to even >out >>the actual on and off situation by the AC power.
>Who ever told you that is incorrect. The phosphor coating when >bombarded with the UV light created by the arc fluoresces >which produces the visible light. Once the arc stops, the >phosphors for the most part go immediately dark (watch a >fluorescent light when you flip the switch). It’s the latency >of our visual system that makes us believe the blackout does >not occur,but it actually does.
Thanks Pete for the clarification! That sounds very reasonable to me.
However, maybe there are some different versions out there, as I have noticed that some tubes really keeps glowing a while after they have been turned off, but as the light that is visible after turn off is very week, I suppose it's not even close to be a factor on the subject of this thread.
#18. "RE: Weird WB Issue (Maybe)" In response to Reply # 17
>Sorry, forgot to ask just out of curiosity, are you thinking >about LED-lights here? >
No, I was referring to the policies put in place by the US Energy Policy and Conservation Act.
While not outright banning the manufacture of magnetic ballasts, put such a high energy efficacy factor on them it makes their manufacture impractical due to the cost and size that would be needed.
As such any new installation or retrofit will be done with more energy efficient electronic ballast that operates at much higher frequencies.
There was a time period where replacement ballasts would still be made and sold in the US to allow time for everyone to prepare for the cost of the upgrade, but I believe that now the manufacture and importation of magnetic ballast has stopped, and the stock pile of those who hoarded them to prepare for this day is dwindling.
So in another 5 years or so people will be forced to switch over if they have not already.
#19. "RE: Weird WB Issue (Maybe)" In response to Reply # 18
>No, I was referring to the policies put in place by the US >Energy Policy and Conservation Act.
Aha, ok I understand. Thank you for sharing this info. Might be a bit OT, but as I live in Europe, all incandescent lamps are being phased out in an energy saving decision by the EU, which will force fluorescent lights to be the main light source from home to industry level. That will definitely become a problem when you have to "retrofit" even in theatres, concert halls etc. The solution for these kind of venues are that they are now being forced to use LED's instead (a great light source when it comes to energy savings, which is a good thing of course), for moving lights, and spots. For the photographer that will be a nightmare when LED-lights are programmed (and supposed) to produce something near white light, as the LED's at the moment is very narrow in spectra (lot of spectral peaks and dips) compared to a more "fullrange" incandescent light. Looks horrible, and everyone looks like a ghost under these lamps. How the cameras will handle this regarding WB in the future, I have no clue, but when I see my pictures taken in these conditions with Auto WB, I just want to cry...
They say the LED is still under progress and that they will be better in the future, but I have my doubts it will ever get close to a satisfactory level for the eye at least.
#20. "RE: Weird WB Issue (Maybe)" In response to Reply # 19
>Might be a bit OT, but as I live in Europe, all incandescent >lamps are being phased out in an energy saving decision by the >EU, which will force fluorescent lights to be the main light >source from home to industry level.
A bit OT, but the same thing is occurring here across the pond. Presently phase out of 100 watt incandescent happen this year, 75 watt next year, and 60 watt the year after. But let’s make it less OT and relevant. After all we’re going to be faced with this issue and must learn to deal with it.
When viewing the spectrum of LED, CFL and tungsten lighting, you’ll notice LED and tungsten are fuller spectrum. LED is heavy on the blue end where tungsten is heavy on the red end. From a human vision point, the LED will look whiter (blue and yellow combined I believe) where the tungsten will be warmer looking. From a photography standpoint they both would work well, they just need a difference in WB tonal adjustment.
But discharge lighting, found in sport arenas and from fluorescent lighting, as you noted are lacking in specific areas of spectrum creating spikes. A problem not discussed with such light sources is one of metamerism. I participate in another photography related forum where there was a person who asked why the team jerseys of a hockey team looked navy blue in the photos, but was actually purple. But the rest of the photo looked fine, and they wanted to know what they did wrong. My reply was that they most likely did nothing wrong, and it was an issue with metamerism. The pigment in the dye that colored the shirt depended upon certain parts of the spectrum to be in the light source which when combined is perceived as purple. With one spectrum missing, it causes a shift in the reflected light. Our brains can adapt to this deficiency our eyes capture (there is a debate about why, but one theory is ‘discounting the illuminant’), but the camera sensor only records what is there, not what should be there. One blaring example for people is often with auto-body repair. A number of body shops will use third party paints instead of OEM because they are cheaper. In daylight the repair area looks great, under sodium vapor street lights on the other hand, the repaired are is often blaringly apparent.
For the most part it will not be an issue for the average enthusiast, but for commercial product and fashion photography, it’s a potential disaster.