Problem w strange light

Discussion in 'Canon EOS' started by nick_carulli, Mar 6, 2007.

  1. I have acanon 1ds mark 11,16.7 mp. Recently while shooting indoor sports I
    noted 1/4-1/2 frame of random photos there was a yellow discoloration along
    the long axis of the image.( It was not a dirty sensor or software proble w
    photoshop.) it was not on each image. It also could not be from a different
    temperature of light since 3 frames shot at the same time and only 1 frame had
    this. It was sort of like when the flash sync was off on film camrea. I was
    shooting asa 1000, and it was most apparent on the 85mm 1.2. It was not always
    on the same side of the image, it was sometimes in the middle. Anyone else
    seen this?It does not seem to happen on outdoor subjects. The lenses are clean
    as well.
  2. Under fluorescent lighting, it's quite possible. The faster the shutter, the worse the problem. Try to stick to 1/60 or 1/120s
  3. The lights are flickering with 60 cycle current, and your faster shutter speed used with the f/1.2 lens caused it to show. This happens with many types of efficient lighting systems commonly used in indoor sports buildings.

    As Raphael said, you have to keep the shutter slower if you want to avoid it. But I think 1/120 is still too fast. Unfortunately, this will likely cause motion blur with many subjects, and you can only use so many panning shots.
  4. -- "The lights are flickering with 60 cycle"

    Actually, the light is flickering with twice the frecuency of the supplying AC. That is because the emitted power is the product of current and voltage.

    The voltage is U=Uo * sin(wt)

    The current is I=U/R or I = Uo * sin(wt) / R

    The power is P = U * I = Uo*Uo * sin(wt) * sin(wt) / R

    that is Uo*Uo * (0.5 - 0.5*cos(2wt)) / R

    as you see, the term in the cosine function now is 2wt (instead of wt in the sine function). Therefore, the frequency has doubled.

    This is also easy to visualise ... the emitteds power is the product of two sine functions. The peaks in emitted power are at the spots where the sine functions reach their positive maximum as well as where they reach their negative maximum. So, in one cycle of the sine function, the power function had already two cycles (aka double frequency).

    I wouldn't have mentioned this, because it seems to be a bit nitpicking ... nevertheless, the frequency is important to choose a correct shutterspeed.

  5. With fluorescent lights, I'm not sure that Rainier's analysis is strctly correct, since at lower voltages some of the phosphors are not emitting at all at higher wavelengths - which gives rise to the changing colour balance. He is correct that peak output happens on each positive and negative portion of the cycle, so the effective frequency is twice the mains frequency. There's a nice example posted in this thread:
  6. My analysis is only correct for the electrical power consumed, not for light being emited ... (because light emission doesn't go back to zero immediately when electrical power consumption reacheds zero) ... I was just commenting on the frequency.

    There are some interesting comments in the thread you mentioned ... which in afterthought bring me to the persuasion, that one should stay a bit below X-sync speed of the camera used, if tube/fluorescent-light is used as a main source of light. That way, variations in lightstrength and color can influence the complete frame. Using a very short/fast setting like 1/4000sec will not help, because the shuttercycle still takes about X-sync time to complete ... A faster/shorter time will only narrow the horizontal band. The trick is to use a slower/longer time, and thereby spread the band over the whole frame.

  7. Actually, a fluorescent tube has a resistance that falls with increasing voltage - hence why they are operated with a ballast that limit the current. Probably more than you want to know about fluorescent tubes:

    X sync isn't the key variable here - the important thing is to capture the output from an exact cycle or number of cycles of light output. If the shutter curtains move slowly across the frame so that only a strip of the sensor/film is exposed at the shutter speed, so long as each portion of the sensor/film is exposed for an exact cycle or multiple the lighting will appear even, although different parts of the sensor/film will start and end their exposure at different points of the cycle.
  8. Mark,

    on modern cameras X-sync is always around 1/100sec or faster (thereby
    X-sync is always near the timing of one power-cycle of the tube or X-sync is faster ... but X-sync is never slower.

    Shorter times than X-sync do expose each individual location of film/sensor shorter than X-sync-time (and thereby never for one or more
    power-cycles), the whole exposure takes a bit longer than X-sync-time.

    Therefore, not shutter times shorter than X-sync exist, where only a small strip of film/sensor is exposed for at least one (or more than one) cycles.

    So, only times of at least X-sync-time will help.

    An exception would be a camera with very slow X-sync of say 1/30 sec,
    where you probably get a good picture with 1/50sec (since your assumption now applies).

  9. maybe someone could give him some actual useful advice regarding his situation. any other
    options besides use a longer shutterspeed?
  10. I doubt it. If you shoot at high speed under fluorescent lights, this can happen.

    The only advice is to use a slower shutter speed or not to shoot under flickering fluorescent lights, neither of which helps a great deal.

    There are some things you just can't do.
  11. Well, there is always flash. Distances are probably great enough to require a pretty hefty unit though, and you would need to overpower the ambient lighting with it. A flash extender, like the type used for fill flash in bird photography, might help, but if using less than a 300mm lens you will probably get some vignetting.
  12. Re: Using a flash, I shoot basketball in our high school gym using a Canon 550ex in the middle of the stands and get good exposure at ISO 1600. I imagine the 580EX would be even better.
  13. The advice is to shoot at 1/125th or 1/60th assuming a 60Hz supply or give up, unless something can be done about the lighting, such as using a 3 phase supply, or fluorescents with high frequency electronic ballasts (discussed in links I provided above), or overpowering the ambient lighting with strobes.

    I think that Rainer has still missed that even if you use a Leica IIIf with an X sync shutter speed of 1/30th, if you set the shutter speed to 1/125th (essentially the same as 1/120th for this purpose), each section of the film will receive the light output from one complete cycle of the fluorescent light, albeit that the left side of the frame will be illuminated by a completely different cycle from the right side of the frame (the shutter travels horizontally). The period for which both curtains are open at max X sync need only be long enough to accommodate the flash - perhaps no more than 1/500th second with a modern shutter.
  14. Thanks for the input. i reshot today and did an experiment. When I used asa 1000 at 1/100 sec no problem, but as i increased shutter speed to 1/850 the yellow appeared on the edge, and in the middle.It seems to indeed have to do with the freq of the lights and the shutter speed. However I am still not sure if its only my camera, I need to have this shot with another 1 ds and see if the same thing happens.I have seen alot of shots done in a gym, and with gymnastics you need a fast shutter, but I dont recall seeing this problem befor. Please let me know if anyone did a similar test. i hate to imaging its just my camera. Thanks for the input, nick
  15. You will see exactly the same symptoms and causes in the photos linked from this thread:

Share This Page