I have shot high school gyms a lot over the last few years and one thing I think my camera has problems with is when gyms are lit with mercury vapor lights. These are the ones that flicker with the 60Hz cycle and change color all the way from a reddish tint to a pasty bright blue-green color.

Now flash photography is not allowed for colorguard shows I shoot in these gyms. So I am stuck with available lighting.

With folks who have had to shoot available light, wide open, f2, in these gyms, does anyone else notice problems with lights that really cycle bad and cause the focus to drop out all the time?

Well, I've certainly noticed the color cycling, but it's never affected my focusing, even at f/1.8 which I use occasionally.

What makes you think the lights cycling is negatively impacting your AF performance?

The fact that when I go through the pics, I have a lot more throw-aways and it seems like it is in sync with the cycling. I notice it especially when shifting to the reddish, flat light.

crfrey71 wrote:
The fact that when I go through the pics, I have a lot more throw-aways and it seems like it is in sync with the cycling. I notice it especially when shifting to the reddish, flat light.

OK, so you have the effect (decreased AF performance), and have inferred the cause (lights cycling).

I'm not sure that light changing wavelength (color) will impact AF - light changing intensity, now I can see that.

But, I may be wrong.

well, they are definitely not as bright when they are the reddish-tint. The bright blue-greenish tint are almost a stop higher than the reddish tint ones.

Play with Custom Functions:
13, 17, 20 and even use the "Ring of Focus" - depending on your subject contrast availability relative to your background.

I too shoot in those conditions and can keep up quite well with variances in those CF settings.

Hammy.

crfrey71 wrote:
I have shot high school gyms a lot over the last few years and one thing I think my camera has problems with is when gyms are lit with mercury vapor lights. These are the ones that flicker with the 60Hz cycle and change color all the way from a reddish tint to a pasty bright blue-green color.

Now flash photography is not allowed for colorguard shows I shoot in these gyms. So I am stuck with available lighting.

With folks who have had to shoot available light, wide open, f2, in these gyms, does anyone else notice problems with lights that really cycle bad and cause the focus to drop out all the time?



I don't think the cyclic light causes focus problem. I use the old Canon 1D to minimize color shift causes by the cyclic light. The old 1D is less prone to color shift from cyclic light than the newer 1D Mark II and up.

I don't think the cyclic light causes focus problem. I use the old Canon 1D to minimize color shift causes by the cyclic light. The old 1D is less prone to color shift from cyclic light than the newer 1D Mark II and up.

You've piqued my curiosity - how can one camera be "less prone to color shift" from the lights cycling than another? Do you mean it's simply less sensitive to light in general?

I gotta admit - this entire thread has me baffled.

dmwierz wrote:
I don't think the cyclic light causes focus problem. I use the old Canon 1D to minimize color shift causes by the cyclic light. The old 1D is less prone to color shift from cyclic light than the newer 1D Mark II and up.

You've piqued my curiosity - how can one camera be "less prone to color shift" from the lights cycling than another? Do you mean it's simply less sensitive to light in general?

I gotta admit - this entire thread has me baffled.


CCD is less prone to color shift. You will still see it but less apparent. Another alternative is to shoot at 1/60s, which is not quite practical in sports unless you are panning.

10D - I hear what you are saying, but I still don't understand it. Just saying it doesn't make it so.

Color "shift" of the lights only means a change in the color due to the AC power and the sinusoidal cycles this power goes through. The lights emit different wavelength (color) light as the voltage goes through these sine waves, and the corresponding increases and decreases in voltage powering the lights.

The camera doesn't care about any of this, and has no way of knowing what color "reality" is - it only tries to reproduce what it senses. For all it "knows", the light is supposed to be red (or green), not white or that the light is supposed to change from red to white to green.

It seems to me that, if a CCD sensor was less prone to color shift, this would mean it would be less prone to picking up the colors to which the light is shifting, and therefore would have huge holes in its color response curves. This would not be good since sometimes you really WANT the sensor to pick up red (or green) light, as when light is reflected off something red (or green).

I still am at a loss to figure out any of this. Maybe it's me...

dmwierz wrote:
It seems to me that, if a CCD sensor was less prone to color shift, this would mean it would be less prone to picking up the colors to which the light is shifting, and therefore would have huge holes in its color response curves. This would not be good since sometimes you really WANT the sensor to pick up red (or green) light, as when light is reflected off something red (or green).

I still am at a loss to figure out any of this. Maybe it's me...

*** Deleted comment ***

See Hammy's post below instead.


Edited by Canon 10D on Mar 25, 2008 at 04:39 AM GMT (Reason: deleted comment)

10D,

The 1D does not experience color shift because the CCD executes the exposure and not the shutter leafs as all CMOS and film bodies do.

CMOS camera:
- depress shutter
- mirrors move out of the way
- CMOS turns on
- first shutter leaf start traveling...starting exposure on one side of the sensor
- second shutter leaf follows...ending exposure on starting side
- measurement of gap to expose sensor a direct function of shutter speed
- first shutter leaf travel complete at other end of sensor
- once second shutter closes the gap all the way, exposure is complete - but did so in a matter of time gradient across the sensor - therefore susceptible to color shifting of lights
- CMOS off
- Mirror slap down
- shutter leafs return to home position.
- repeat as needed

CCD camera operation
- Press shutter button
- mirror moves out of the way
- shutter leafs open all the way
- CCD controls the shutter speed with ON/OFF
From Canon (http://www.usa.canon.com/EOS-1D/specs.html)
"Electronic shutter with the CCD sensor and vertical-travel, local-plane shutter with all speeds electronically-controlled"
- shutter leafs close completely
- mirror returns
- shutter leafs reset.
By having the entire CCD 'exposed', the light shifting is still prevalent, but not a gradient. The whole CCD will be subject to the shift - wherever in the phase it may be and go through - but it will be a solid color cast.

The CCD can pick up colors just fine at the cycle speed of lights. You're correct that our eyes don't sense anything above around 18 frames per second, much less 60. Video camera don't seem to experience it because they are operating at both a higher fps as well as different shutter speed.
http://www.100fps.com/how_many_frames_can_humans_see.htm
Shooting at 1/60th or less will nix the gradient completely because the frame is allowed to see the whole cycle - giving it the proper color. Even up to 1/120 technically can get you half a cycle - unless its the complete lower half. Also shutter speeds above 1/640th will defeat the gradient because the exposure is so fast. This will net image like the 1D (CCD) with a fixed color cast - wherever it may be.

Otherwise,
there is an "averaging" occurs in CCD or you can say that CCD is less accurate in reproducing color that changes at 60 Hz compared with CMOS
sounds interesting! (enough to confuse Dennis anyway )

Hammy.

Hammy wrote:
Otherwise,
there is an "averaging" occurs in CCD or you can say that CCD is less accurate in reproducing color that changes at 60 Hz compared with CMOS
sounds interesting! (enough to confuse Dennis anyway )
Hammy.

Sorry for the confusion, I meant to say that CCD is less prone (not less accurate) to reproduce the color shift. I had too much wine for dinner. My bad

Hammy,

Now THAT's what I call an answer, and no, it didn't even confuse me.

Hammy, you rock, man!

See, I knew there was a pony in here somewhere. Thanks!

Dennis

Dennis,
I figure the averaging didn't confuse you - I know that takes alot! But I figure it had you wandering down a different path of curiosity.

I have this issue at the Valley Forum Hockey rink in Haverhill, MA, not at any other rink with these lights. I have actually given up trying to take pictures there.

There are two issues going on, the first is the voltage/current going into the lights, the VF turns the voltage down running the lights at half power to save money which causes a really bad red shift (a CWB is useless btw), the second is they are using 12 of these lights for the rink which is not adequate at all which means I have a 2 stop variation in light across the ice. Shooting in RAW can help but I still get color shifting and light fall off. To fix the problem they would need to double the number of lights on the ceiling.

During playoffs I am tempted to slip the manager a 20 to get him to turn the power all the way up so I can get some descent shots.

Wow! This topic got off what I was wanting, but, it's cool.

gerry has hit it on the head. I understand all the color shifts. I merely described the color shifts, so you knew what I was referring to, as the conditions I was facing.

The mercury vapor lights are the biggest pain. Yes, they have ugly color shifts, but the exposure variances like gerry describes are very true. I can experience an almost 2 stop variance. The shift to red is underexposed and really flat lighting, which is where I lose a lot of my focus. Then it almost overexposes as it shifts to the bright green/blue tint. The sodium vapor lamps, I usually don't have much issue with, only the mercury vapor lights. With florescent bulbs, I have no issues whatsoever. Therefore, I know it is not me.

Now, I shoot really tight at 200mm and at f1.8. Plus, there are sabers, rifles, and twirling flags that always can throw off the focus. So, I have a lot to deal with and I usually set my custom function to moderately slow tracking sensitivity. That way, it shouldn't right away grab the saber, rifle, or flag so quick. It is hard to keep focus on in good conditions, since sometimes the shiny weapon of choice might be up by there face, or down by their waist. The camera loves to grab that shiny weapon and lock on it. I try using none expansion points, 7, or even 13. I usualy stick with 7, because it is very difficulyt to stay locked on with no expansion points.

I am thinking with the shooting conditions and the way I like to be very tight on these type of shots, it leaves very little room for error and very thin focal plane. So, I think that these mercury vapor lights might cause me more issues than with folks shooting basketball, gymnastics, etc...

I just wanted to see who else thought or had experience with these mercury vapor lamps, with no flash, that might have had similar issues.

Charles,

I know you said you can't use flash, but would they still be averse to strobes bounced indirectly off the ceiling?

Regarding the AF locking onto sabres, flags, rifles, etc., I have faced the same thing while covering band/color guards/flag teams and have found that if I focus lock on the subject (via the back button focus) and providing the subject isn't moving towards or away from me, I can concentrate on timing the shot when the rifle isn't blocking the face (or take a burst). However, even if the AF does lock onto the shiny weapon, is your DOF that small that this throws the subject OOF? If so, an alternative is to shoot a longer lens from farther away which will open up your DOF somewhat (though if you're already at f/2.0, I presume you're shooting a 200mm lens already, and if you need f/2.0 in order to get your exposure values correct, 200 may be as long as you can go).

I have never really noticed focus problems - but color shifts in a fast sequence can be quite different. When I do my white balance, I drop my shutter speed to 1/20 of a second and the white balance gets both ends of the color shift in the same frame and all images have the same white balance ( an average between the all the color shifts).

Gerald

I shoot winterguard also. This year I went to a gym in Austin Texas and really got to experience that color shift that you are talking about. I use manual exposure and you can literally see the color shift from 1 second to another. I have have even had half a picture get caught between changes.

Now I do notice that my focus does seem less sharp when the lights change and the picture starts to become underexposed. I shoot with a Mark III and in those situations I am already shooting at iso 64k and what I think I am noticing is the 'more noise' in the picture due to the slight underexposure. When I take a picture at the light's brightest moment it is perfect but when they start shifting to a lower light spectrum I notice a significant noise increase.

As far as bouncing strobes goes I have tried that. We have a couple of issues though. Unlike basketball where the focus is usually on just two areas of the court we have to take pictures all over the entire gym floor. That means a bunch of strobes to make sure the entire floor is lit evenly.

The other big hurdle i have is convincing the circuit directors to let me even run strobes to begin with. I went to a rehearsal and ran them one time and the kids did not even notice at all. I think it is much more of a mental thing we are going to have to work on with colorguard people. Until Winter Guard International lets photogs at their events run strobes I don't see the local circuits letting us run them.

Right now I do not own any strobes and they can be pretty pricey because I would have to get at least 4 strobes, stands, extension cords..... it can add up pretty quick... I am fortunate in that 90% of my gyms do not have lights that cycle really bad and until I can get a guarantee that I can even run strobes I am not going to invest in them ... at least for this sport

dmwierz wrote:
Charles,

I know you said you can't use flash, but would they still be averse to strobes bounced indirectly off the ceiling?

Regarding the AF locking onto sabres, flags, rifles, etc., I have faced the same thing while covering band/color guards/flag teams and have found that if I focus lock on the subject (via the back button focus) and providing the subject isn't moving towards or away from me, I can concentrate on timing the shot when the rifle isn't blocking the face (or take a burst). However, even if the AF does lock onto the shiny weapon, is your DOF that small that this throws the subject OOF? If so, an alternative is to shoot a longer lens from farther away which will open up your DOF somewhat (though if you're already at f/2.0, I presume you're shooting a 200mm lens already, and if you need f/2.0 in order to get your exposure values correct, 200 may be as long as you can go).

Winterguard and colorguard during band contests are quite different. Shooting colorguard at band contests, no problem whatsoever, except if a band has 300 members and you are trying to zoom your way back to the other performers. The colorguard routines at band contests are pretty fluid, and in line formations, etc... Winterguard shows are much more challenging and stop-n-go'ish. If they are standing and just twirling, I can usually pin-point the focus no problem. The focal plane is that narrow. Therefore, I try and shoot on a 45-60 degree angle. That way, if the focus locks on the rifle on that angle, it is most likely in the same focal plane as the face. Shooting straight on, if the focus locks on the rifle, the face will be soft and OOF. Like I said before, I really shoot tight and I want that. That of course means little or no room for minimum focus errors. I accept that, especially when i get some really nice shots. I shoot close because I want that hand-eye-coordination kind of shots. I also like the look of the 45-60 degree angle shots. there is more of a 3D look to them and the weapon and body have a better separation.

wiens51 wrote:
I have never really noticed focus problems - but color shifts in a fast sequence can be quite different. When I do my white balance, I drop my shutter speed to 1/20 of a second and the white balance gets both ends of the color shift in the same frame and all images have the same white balance ( an average between the all the color shifts).

Gerald

gerald,

I set my color balance the same. That way I try and minimize the color shift and average it the best. I do a color balance at 1/60 sec with my expodisc. I also use a Gretag color checker and take a few shots at different ISOs, mostly to get a nice color profile so I can set a make a nice noise profilefor ISO's I might use at that venue. It also provides another nice reference.

Yeah, I have tried unsuccessfully to convince circuit directors that strobes would work fine and not distract the performers. I told them the flash duration would be somewhere around 1/1000-1/1200 sec, so fast that our human eye wouldn't have time to notice it. I try and tell them by referencing to basketball games. I said you watch the games on TV and asked them, ever really notice the storbes firing away? Anyways, one of these days, maybe.

I think gymnastics doesn't allow strobes also. Am I correct?

As far as I know, the "NO STROBES/FLASHES" applies also to gymnastics and cheer.

John