Canon 5D Mark III Light Leak?

Status
Not open for further replies.
smithy said:
I've tried this test on my 40D and 1V, and neither has the 'issue'.

kareldonk said:
The LCD backlight doesn't affect the shutter speed/exposure when on. but why is the exposure fluctuating that much?
You need to take the lens off and just put the cover on the camera body. Lens caps are designed to protect the lens, not block out 100% of light.

I did, same tests only in this case with body cap on (no lens attached). same fluctuations. why?
 
Upvote 0
Since this doesn't affect every situation, would this be a suitable temp fix?

If you in a situation where light, either moving from sun to shade, or anything that triggers this issue would applying a piece of gaffers tape over the top screen solve the issue?

For example the poster who was in the beach in and out of the palm trees and direct sun, would this issue be non existant if the screen was covered?

I know people are covering the viewfinder and body without lens to see if the issue occurs, in normal operation I was wondering if for certain situations simply blacking out the top screen would avoid any possible issues?

sorry if the solution sounds stupid, i dont think i fully understand the issue, i didn't have a chance to watch the video or test it myself.

One thing is for sure, canon should not have waited so long to release this camera, being the overwhelming wait people are putting this thing through the ringer.
 
Upvote 0
smithy said:
You need to take the lens off and just put the cover on the camera body.

If that's an absolute requirement to demonstrate this problem, then it's not a real problem at all - at least, it's not for those of us that prefer to take pictures with lenses mounted. :P
 
Upvote 0
This reminds me of the amazon review, how the reviewer was much happier when he finally attached a lens to the camera, and suggested canon come up with a solution to the camera not taking great pictures out of the box.

neuroanatomist said:
smithy said:
You need to take the lens off and just put the cover on the camera body.

If that's an absolute requirement to demonstrate this problem, then it's not a real problem at all - at least, it's not for those of us that prefer to take pictures with lenses mounted. :P
 
Upvote 0
neuroanatomist said:
smithy said:
You need to take the lens off and just put the cover on the camera body.

If that's an absolute requirement to demonstrate this problem, then it's not a real problem at all - at least, it's not for those of us that prefer to take pictures with lenses mounted. :P

Makes you wonder how this issue was found in the first place! ;)
 
Upvote 0
prestonpalmer said:
Fleetie said:
Remember, in real life, taking a picture, a lens would be on the camera, and wide open. There would be no visible or significant effect on the exposure in that case; the light coming in through the lens totally dominates, and that coming through the top LCD becomes insignificant.

You are wrong. We are seeing the change from sunlight to shadow change the exposure by as much as 1/2 stop when the shooter moves from daylight to shadow in real shooting conditions. I can confirm this with a bright sunlit beach photo shoot yesterday.

I tried that just now, no change in exposure from bright sun to shade as long as my viewfinder was covered. Uncover the viewfinder, or hold it to your eye so even a little light leaks by, and there is a large change from very bright sun to shade (when focused on the same subject.)

Were you shooting in your tests with the viewfinder blocked so there was no influence from light leaking in thru the viewfinder??
 
Upvote 0
I'm constantly amazed at the reactions.

The problem clearly isn't understood totally yet... and still people already demanding product recalls and claiming this is a huge problem for Canon's reputation.

No product is perfect! EVER

No product is ever fully tested before it is released - it is impossible to test every situation in the lab.

I'm not saying this shouldn't be investigated and potentially improved, but let's not over-react
 
Upvote 0
Geeeeez!
Now what is new??... Can't use the 70-200, can't use EOS Utility, now a light leak!! Canon has really shot themselves in the foot with this one. They need a major recall or offer a refund to all buyers. I sure am glad I didn't pre-order one of these lemons. $1,000 over the 5DII for two stops increase ISO sharpness...bend over here comes your cannon!
 
Upvote 0
planetMitch said:
I'm constantly amazed at the reactions.

The problem clearly isn't understood totally yet... and still people already demanding product recalls and claiming this is a huge problem for Canon's reputation.

No product is perfect! EVER

No product is ever fully tested before it is released - it is impossible to test every situation in the lab.

I'm not saying this shouldn't be investigated and potentially improved, but let's not over-react

+1

Even with my eyepiece blocked, I could get a lot more light into the camera with a bright light than I could thru the lcd.

I tried several ways to get the leakage to change a real world shot to no avail. When holding the camera to your eye, far more light gets thru the viewfinderand does change the exposure when you go from shade to sun (wearing glasses as I do). Without glasses, my eye might get a better seal to block more light. I need to try that.
 
Upvote 0
OK, I have done a few tests on the two bodies that I have and they both fail this test. But, as long as I don't shot with the LED light on there is no problem that I can see. With a lens mounted in very low light or darkened room I can not get the meter to change when I put a flashlight to the LED screen, so it seems ambient light is not effecting the meter. I have tried this test inside and out under different lighting conditions and can not get the meter to change. That seems to be good news and explains why I have not seen any strange metering effects. I can see how this would be an issue if you are shooting in low light and using your top LED screen a lot, but the rear 'info' LED screen gives you the same information and for me much easier to read. This will be interesting to follow, and I look forward to hearing more on the issue. 8)
 
Upvote 0
I dont think there is a light leak.

1. If I switch to P mode (never even knew I had a P mode, I am always in M)
2. With the lens cap on and I turn on the top lcd backlight I do see the change talked about.
3. If I take the lens cap off I dont seem to be having the same issue.
4. I shined a mag light at the top lcd screen and saw no change.
5. If you switch to A+ mode and I turn on the top lcd light even with the lens cap on I do not see any change in the evaluation from the camera.
6. I think that the fact that the case 1 and 5 are different suggest a software bug or something. Since my mag light does not seem to affect the exposure. I dont think there is a real leak at least from outside light. Anyhow who uses P mode? If you do why do you have such an expensive camera?

Ryan
 
Upvote 0
Just tested my 5d3, it has the same problem, most noticeable under very dark environments with the top LCD light ON. The exposure changed dramatically, a few stops.

I think this is related to a hardware design issue. I'm not sure how Canon can really "fix" this for the people who own the first batch of 5d3s by not doing recalls.

As a user I can try not to use the LCD light feature as much as possible when making an exposure, especially in dark environments. I can live with this so I'm not too worried, but still, who doesn't like a perfect product? Especially the illumination of the top LCD seems like such an important feature because Canon even made a button specifically for it at the most comfortable and reachable place. lol.
 
Upvote 0
Mine does it. It even changes a full second exposure if I tilt the camera towards the only light in my apartment that's on.. not much hitting the LCD...

I went out and did a shoot today, and I did notice some really weird metering issues. Coming from the 7D, I am already fairly used to the metering changes that Canon has been making, but some of these were telling me to expose drastically wrong in spot meter mode. I'm wondering if this really is a "non issue" as some are saying, or if this is a bigger issue. Regardless, we need to hear from Canon on the matter.

On another note, I also got an Error 80 today. That was fun.

Not all bright and fuzzy feelings anymore...
 
Upvote 0
I am seeing a 1/3-2/3 stop exposure change with a lens on both of my 5d3 bodies. This happens both when I have the LCD on and when I shine a small pen light onto the LCD. Will this affect me in the real world? Not really. I'll just have to remember not to turn the LCD on if I'm ever doing any night timelapse or other low light work in an automatic metering mode. The issue is definitely there for me, but it's not a concern.
 
Upvote 0
Status
Not open for further replies.