Canon 5D III leaking light issue:

[arthurking83]

Just a warning for those that have or are getting the 5DIII.

It seems that the 5D III's top LCD screen has some kind of light leaking issue.

Canon has issued an advisory stating that a fix is coming.
It seems tho that the only way to fix this is either via a replacement, or an actual serivce.
This is not going to be a firmware fix(if it is, I'd be dubious about it!!)

5DIII ight leaking issue news (http://www.canonwatch.com/leakgate-eos-5d-mark-iii-light-leaking-issue-confirmed-by-canon/).

!!! No anti Canon fanboism here. Any posts to this effect will be removed !!!

This thread is only for the purpose of anyone that has been affected by this problem to discuss if they have the issue or not .... and just in case they haven't seen the news anywhere else.

[Chris G]

I've shared the link on my facebook page for anyone who had brought the 5Dmk III :th3:

[Steve Axford]

Doesn't sound like a major issue for most users. I've not noticed any problem, but I'll look for it.

[vk2gwk]

There is a whole thread about this issue on "Photography on the Net" (the Canon forum) (http://photography-on-the.net/forum/showthread.php?t=1170197) and some follow up threads - including the announcement as made above.
I have checked things out with my new 5D Mark III and cannot find anything wrong under normal shooting conditions.
When I shine a strong light onto the LCD top display nothing changes when I am focussing on a normal subject. Only when I do the same with the lens cap on or without a lens and the body cover on I see a change in the metering (put on Av at 800 ISO you can see the shuttertime change from say 30 sec down to 10 sec).
But who is going to shoot with the lens cap on... :)
May be astronomy photography or night time very long exposure shots may be affected by the issue. I'll just wait and see what Canon comes up with....

[flashc]

There is a whole thread about this issue on "Photography on the Net" (the Canon forum) (http://photography-on-the.net/forum/showthread.php?t=1170197) and some follow up threads - including the announcement as made above.
I have checked things out with my new 5D Mark III and cannot find anything wrong under normal shooting conditions.
When I shine a strong light onto the LCD top display nothing changes when I am focussing on a normal subject. Only when I do the same with the lens cap on or without a lens and the body cover on I see a change in the metering (put on Av at 800 ISO you can see the shuttertime change from say 30 sec down to 10 sec).
But who is going to shoot with the lens cap on... :)
May be astronomy photography or night time very long exposure shots may be affected by the issue. I'll just wait and see what Canon comes up with....

back in the 90's, a problem was discovered with the Intel Pentium processor mathematical calculations under certain conditions. As I recall, Intel said that this problem wouldn't affect most users but pressure was brought to bear on Intel to fix a flawed CPU. At the moment I can't decide if I want to buy a 5D mk II or 5d mk III to have alongside my 7D. This suspected issue of a physical light leak through the LCD says to me there is a design flaw. Canon always advises to use the eyepiece cover when shooting on a tripod to stop stray light entry into the body. If stray light is entering through the LCD but only under certain conditions, then I expect this to be fixed before I spend my $4000.00 on the off chance that I want to shoot under those conditions. 5D mk III - this is still the camera I've been waiting for that will shoot quality low light action without flash. Canon needs to recall all 5D mk III's if it's a build problem.

[ricktas]

Is this light leak through the internals of the camera, or is it that the large LCD and brightness of it, is causing the metering to see more light through the lens, and adjusting for it? From the Canon announcement it doesn't say which way is causing this.

Next question, can the brightness of the LCD be changed? On my Nikon's I can adjust the brightness of the LCD backlight output (from memory here cause I haven't played with that function for so long, but I recall reading I could adjust it, I think).

[etherial]

Have a look at the videos in this Petapixel post, they demonstrate the issues well...

http://www.petapixel.com/2012/04/13/canon-confirms-light-leak-issue-in-the-5d-mark-iii/

[arthurking83]

The LCD in question is the top plate info screen, not the review screen.

What seems to happen, in low light, is that if you have the camera set up with a semi auto mode like Shutter priority or Program auto, and you switch the top LCD light on and off, the exposure setting indicated on the top LCD panel changes .. not by much, but enough.

I think in aperture priority mode used in one of the videos, the shutter changed from 1/6s to 1/10s simply by turning on the LCD panels internal light.

There is most likely a light leak between the LED light that illuminates the LCD panel, through to the prism or metering sensor. Maybe a small leak.
You should get a similar exposure difference when shooting in a semi auto mode and covering uncovering the viewfinder, if it doesn't have a shutter to keep it dark.
I use either my hand, or an old lens cleaning cloth to cover up the vf if I'm not peering through it.

[Bennymiata]

This is one reason I'll wait until June/July to buy one.
They should have fixed all the little problems by then, and Adobe will also have good RAW support for it by then too.

[Dylan & Marianne]

seriously , I know this is a flaw in the camera and it shouldn't happen

BUT

I'd like to ask a real world question : Who actually goes out night shooting and relies on the camera's metering to determine exposure?
Who can give a realistic example as to when they would have ever been affected by this issue?

My usual flow when I go out shooting stars at night would be - 30 seconds, iso1600 , F2.8 in dark conditions - see how it goes then adjust - I don't think I've ever relied on the exposure meter to help me with this? If I want star trails instead of star freeze, I do some mental arithmetic to work out what iso and aperture I should be at. Perhaps I'm doing something different to what most people do at night but I suspect some theorycrafting going on without asking the people actually taking photographs if it is affecting them one iota. I confirmed that this does happen with my 5dmkIII and yes you'd expect better from a 4000 camera but I'd like to hear some examples of where this would affect you in the way you shoot.

Benny, ps. Lightroom 4.1 now has RAW support for the 5dmk3

[flashc]

Hi..
How much difference in exposure does it cause the metering on your 5D mk III to vary? Less than 1 stop or more?

As you say, take a photo, check the exposure, adjust it if necessary and do it again. We couldn't do that with film.
I've taken shots on a boardwalk at the beach towards the ocean at night with backlighting from the council overhead lighting. (15s ISO 100 F8) Sometimes this lighting falls on the top of the camera. I've also switched on the LCD on my 7D to see the time exposure if I take an AV shot just to test exposure. If the exposure took 30 secs on auto and was overexposed by 2 stops, then I personally would switch to manual and use my newly calculated settings. However, not everyone who has an expensive DSLR works this way and would want to rely on the system to get it right.
I read that the light leak only seems to affect the exposure values and the leak doesn't fall directly on the sensor producing partial overexposure. Also, if you were using liveview, then the mirror would be up - I don't have a 5D III yet, but will the exposure also be affected when using liveview.
Whatever values the light leak gives at the instant the shutter is released, would these values be locked in and not be able to vary even if a bright light was shone on the LCD display during the full exposure length.

[I @ M]

but I'd like to hear some examples of where this would affect you in the way you shoot.

Amen,
there are plenty of keyboard worriers warriors that like to sprout off to spread FUD about a "fault" but I am yet to see an image from the 5D M111 that has been ruined due to this phenomenon, maybe that is because they are all taking these "photos" with a body cap on and without a lens. :Doh:

[Dylan & Marianne]

flash, but the point is, that this is a DSLR and not a film camera! and because you can review mistakes, my learning curve went up very quickly in 2007 compared to the preceding 10 years when I didn't dare these shots in case I had to spend money developing a reel of dud shots!
also, if you're having to adjust after a test shot in AV mode, you could do exactly the same thing from shooting in manual in the first place.
also, I 'tested' this while sitting in a room with just my monitor on (ie when I read about this issue) - pointed away from the screen, turned the LCD light on and it did not affect the exposure readings - I think we're talking very dark situations here - eg , Lens cap on or next to no available light.
I haven't tested with live view yet but then again, in the dark, all I do with live view is use it to help me focus (with a torch shining on an object) and not with exposure.

I'm not saying canon's mistake is excusable - it isn't. I'm just struggling with why this is being made into what sounds like a 'gamebreaker' when it isn't . NOw if I find that leakage of water is an issue - that WILL be a gamebreaker for me especially when shooting in harsh environments

[arthurking83]

I agree that the issue probably won't affect shots in the real world, but it could.

)from what I've read)The issue is not just restricted to using the LCD backlight, it can be affected by direct sunlight on the LCD as well.

if this is true, then if you set yourself up for a landscape shot, and the sun peeks out for a moment over your shoulder and affects exposure from what you set it up to be, with viewfinder covered and so on, it may become an annoyance sometimes.

I suppose any effect on exposure up to approx 0.7Ev from expectation, wouldn't be too annoying.
I've got lenses that differ this much between each other and I account for those differences.
eg. the Nikon 18-105VR underexposes by -0.7 without fail relative to all other lenses given the same exposure settings.
Knowing this, as a baseline setting to account for, is easy to do, and easily remembered.

Problem is when this exposure variation is unpredictable from moment to moment.

if Canon offers a service fix or replacement camera, you would be best off to take up the offer(and any other offers of compensation).

[etherial]

That's right Arthur, direct sunlight is an issue as well, check out the video on the Petapixel link I posted above. It shows up to 1 stop variation due to sunlight on the top LCD screen.

[etherial]

Another video posted by Petapixel today, again a 1 stop difference with the internal LCD backlight (not a big issue to me) and 1/3 using an LED light.

http://www.petapixel.com/2012/04/17/canon-5d-mark-iii-light-leak-issue-alters-exposure-by-about-13-of-a-stop

[Tony B]

Will not stop me from purchasing one. I still have a 1DmkIII that is flawless & we all know how that was trashed. Much due to some photographers not admitting to setting it up incorrectly a la 7D.

[Art Vandelay]

Mountains & molehills. :)

Have had mine about 3 weeks & initially followed this 'problem' around the net.

Considering all the hullabaloo, I'm also yet to see a real world photo affected by it.

It hasn't affected any of mine either, but then I don't have any pressing need to shoot with the lens cap on while turning on the LCD or shining a torch into it. I guess some find that a major part of honing their photography skills. :rolleyes:

Canon have posted an update here (http://www.usa.canon.com/cusa/consumer/products/cameras/slr_cameras/eos_5d_mark_iii?pageKeyCode=prdAdvDetail&docId=0901e02480538fc7)



To Users of the Canon EOS 5D Mark III Digital SLR Camera
Thank you for using Canon products.
Concerning the EOS 5D Mark III digital SLR camera, when the LCD panel illuminates in extremely dark environments, the displayed exposure value may change. Canon has concluded the investigation of this phenomenon, and this announcement informs you of our findings as described below.
Phenomenon
In extremely dark environments, if the LCD panel illuminates, the displayed exposure value may change. However, based on the results of extensive testing this change in exposure value will not noticeably affect the captured image.
Affected Product
Canon EOS 5D Mark III Digital SLR Camera
*Products whose sixth digit in the serial number is 1 or 2 are affected.
　　　For example,　“xxxxx1xxxxxx”　or “xxxxx2xxxxxx” ([x] represents any optional number.)
Support
Under almost all shooting conditions (including dark environments) this phenomenon will not affect your captured images. However, if you would like Canon to inspect your camera, we will provide this service free of charge upon request beginning in mid-May. Please contact Canon using the information below to request service.
This information is for residents of the United States and Puerto Rico only. If you do not reside in the USA or Puerto Rico, please contact the Canon Customer Support Center in your region.
Please register the EOS 5D Mark III. By registering, we will be able to notify you via email when service updates are available. If you already registered, please ensure you are opted-in to receive the notification.

[KeeFy]

Picking mine up today :). If shooting in RAW (which i do) 1/3 of a stop is not a big deal.