Canon acknowledges 'light leak' in the EOS 5D Mark III

"Canon, Delighting you always", is that where the light leak comes from???

Spoke with Canon USA and Canon Japan. According to the representives, the answer is NO. No modifications will be made to the Camera body both internal and externally for now nor in the future. Current stocks that are in the pipelines to be sent out to the retail stores will continue to have xxxxx1, xxxxx2, xxxxx3, and so on with no modications since this "PROBLEM" is not really a problem to begin with. And as Canon stated on their Website, There is no RECALL on this "issue" to be fixed.
But Canon will be more than happy to take a look at those customers who are stil paranoid beginning Mid-May.

Is this issue resolved on the new cameras they are going to ship? Or is it going to have same issue? I am planning to get one.

I placed an order for it today... No idea when Canon will start shipping the new lot.

Wish me good luck, I need one!

Canon has issued their findings and conclusion on this issue:

http://www.usa.canon.com/cusa/consumer/products/cameras/slr_cameras/eos_5d_mark_iii?pageKeyCode=prdAdvDetail&docId=0901e02480538fc7

Essentially stating that it is not a real issue.

The light Leak DOES affect the images. The faulty metering causes some images to be underexposed. Imagine a scene on the limit between two exposure levels. The light leak will then cause the scene to be exposed at the lower level. This will happen at ALL LIGHT LEVELS! Furthermore, in some situations, the light leaking onto the meter through the LCD screen will cause the image to be underexposed several steps. For example, I recently used a 10 stop ND grad to blur the water in a waterfall on a sunny day. In that situation, the sunlight leaking in through the screen is relatively large compared to the scene (as seen through the filter). The result was an image underexposed by many steps, and the images came out completely black. So the statement from Canon indicating that the leak does not affect the captured images is NOT TRUE..

There are more then 250 comments posted on this light leak issue, so obviously there is a problem. All my lenses and camera's are Canon and I just bought the Mark III. What I find rediculous about it all is the way a reputable manufacturer like Canon handles an engineering fault. If Canon where a car manufacturer they would recall all related models and fix the problem. I didn't pay $3000 + for a piece of equipment to have to put duct tape on the inside of an LCD screen, as suggested by some. Canon is responsible for their mistakes and should fix it, simple.

I also read in the manual book under specification, exposure control on page 374 , it is clearly stated that the metering range from ev1 - ev20, at ev1 there is no effect whatsoever on the metering value when you turn on the lcd back light. Thus, its pretty clear that this camera is not build to meter anything below ev1 accurately, so I dont think its a defect...

@Kinematic Digit and others, thanks for providing the results of all your tests. Based on your findings I think we can summarize the issue as this:

(1) - There is a very minor light leak issue through the top LCD, *but* it is only evident with the lens cap on. To date, no one has been able to replicate the problem with the lens cap off... yay! :)

(2) - Daylight exposures are not affected in any way at all.

(3) - Night time exposures are also not affected. Even when exposing for a very dark subject, shining a light on the top LCD or engaging the back-light has no effect on exposure.

If my understanding is flawed, please correct me and provide details about your test.

I can't replicate the light leak issue.
http://youtu.be/LYwALLww_60

I dont know if this is a valid argument, but if and only if let say there is a light leaking, then if you capture the image and you see the result there will be a 1/3 stop or 2/3 difference in the final image output? I dont see there is a difference between the 2 images with the light on and with the light off, at the same EV value, isnt it the image should be 2/3 stop brighter at the same ev value????? if you turn on the light, as the lcd indicates? Where is the metering sensor located at in relation to the top LCD? and where is the cmos sensor located at, on 5d mark III?

It is not an issue at all in normal photography work situation, the object and the surrounding has to be dark for it to affect the ae. Even at f4, at 2 seconds iso 800 does not affect the ae metering when u turn on the lcd top light! with the lens cap off and view finder being covered. Is this really an issue??? at ISO 800, f4, at 2 seconds no change! I am surprised that people are really hyped abt this. the object you are metering and the surrounding has to be really dark for it to affect the ae metering!!!. My goodness and people are really getting so fussy abt this!! and who shoots with the lens cap ON?????????

I have tested canon eos 5 d mark III, its not a problem that people should really worry so much, because with lens cap off the metering doesnt shift under a dim light condition, but it will shift around 2/3 of a stop under extremely dim condition, and we also encounter the same problems with at least 2 canon 5d mark II as well, and the question is why people start complaining now? we also try to conduct the same test on Nikon d300s because thats the only nikon camera we have available, Nikon didnt even bother to start metering with the lens cap on, and with its view finder being covered.

I've tried this on my 5DMk3 and in my opinion it's a small hiccup.

The way I tested it was:
1. turned the 5DMk3 on, set to M (ISO 12800, f4, forgot what was the shutter speed)
2. removed the eye-hood, then put black duct tape on viewfinder to prevent light going into sensor.
3. used 24-105mm lens with lens cap on.
4. turned off the lights in the room.
5. using iPhone4 LED, shone across 5DMk3 top-LCD.

Findings: the exposure reading altered by +1/3 to 2/3 ONLY when I have the LED 1-inch away from the LCD with it shining on the centre of it. When the LED was shining on other parts of the LCD, there was no impact to the EV.

I'm not too concerned about this light leak.

I'm more concerned about amp glow (pink-purplish) glow in the bottom right corner as I crank up the ISO up to 25,600 and beyond (51200, and 102400). Does anyone else notice something similar on their 5DMk3? It's easily seen in dark/pitch-black images.

some people including myself encounter the same problem on canon eos 5d mark II, we conducted the same experiment,and we encounter the same problem. But why nobody had been bitching about it????? is this a marketing war between Nikon and Canon? canon mark III is already out in the market and ready stock in most stores and in most countries, and Nikon is still struggling with their d800 inventory??

That's why I'm on the look out for a cherry, seldom used, yuppy owned low mileage 5mk2, it is already hitting around U$1200 in my country, God I remembered how it used to cost U$3000 when it first came out.

I was about to place my order... its all confusing now: should I wait for the issue to be rectified or totally go for another FF camera? Ofcourse, I am locked on to Canon as I have couple of premium lens... Any suggestion? :-(

Ok, i just watched the videos frm peta...so no need to explain. Tq

Software can be easily written to compensate for the light leak i think. No need to send in for physical servicing..

If the noon sun is shining down on it, will it have any effect?? I dont own one..so dunno..how about strong halogen lighting shining down on it while doing indoor shooting? Wil anything happen? Tq.

for those who have the early runs of the 5DMkIII and still rely on AE metering sensor and the top lcd lit at the same time, a firmware update would be make sense as follows:

Assuming you can even view the top lcd (in cases where the camera is mounted too high above eye-level, LV can be used instead.
1) actuate the lcd light (deliberately or by accident):
2) 'holds' the exposure-iso-sensitivity values (but not locking them)
3) thus as soon as one deliberately makes any exposure-iso-sensitivity adjustment, they change as needed anytime, zero delays.
4) if the light goes out after the normal illumination, and one wants to review the values again, whether changed or not, repeat step 1).
5) one can 'tap' the lcd light button again and turn the deliberately light off.

If one does not make adjustments at all, one shouldn't have any problems

If one is looking over at the top lcd then light going into the OVF will also alter the AE reading given your eyes are not covering the OVF anyway...

Although I think that even with "issues", Canon gear yield excellent results, I suggest you check the following link; they say that maybe even with daylight there would be light leak (sorry but it´s on French. "plein jour" means more or less "daylight"...). Sure, when I purchase my 5D3 all would be fixed.... :)

http://www.focus-numerique.com/canon-confirme-probleme-fuite-lumiere-5d-mkiii-news-3228.html

Canon.... are this still in the camera business? What on earth for?

ah... the pitfalls of relying on AE metering sensors (AE modes).

all the more why Full M/M-ISO w/ ExpSim LV (relies only on the image sensor alone) is ultimately the 'easiest' aspect of digital photography, and that is accurately dealing with exposure determination (in M no less!)

:D