Canon Confirms Light Leak Issue in the 5D Mark III

[BrerFox]

I think you have hit the nail on the head, Sean, we are talking at cross-purposes.

You maintain that the issue at hand is small, insignificant to your application, and therefore irrelevant.

I maintain that Canon are following due process, which is the correct action.
When Canon produce a camera to a set of specifications, and they then find that the unit does not meet those specs, they are obliged to revise either the camera, or the specs to conform.
That is standard engineering practice, and should be admired.

Please also be very clear that this is not like a Spanish spelling mistake which has well-defined consequences. When data from a sensor is out of specification, it could affect many subsystems.
I investigated a problem just last week, where Eskom replaced a small, faulty toggle switch with one of exactly the same electrical specification, but was slightly longer.
The extra length caused it to touch something else under certain configurations, causing a melt-down.

Although you feel this Canon problem to be insignificantly small, you do not know if it adversely affects issues like noise and dynamic range. Neither do Canon at this stage, but are you
not comforted that they are taking this seriously enough to investigate the full ramifications?

[Forgiss - Sean Nel]

Although you feel this Canon problem to be insignificantly small, you do not know if it adversely affects issues like noise and dynamic range. Neither do Canon at this stage, but are you
not comforted that they are taking this seriously enough to investigate the full ramifications?

Yes off course!

And if it comes to light that the low light noise control or dynamic range can be better than it is now (which is already improved) then whoopee!

I think where the confusion comes in is that I am saying it's as insignificant as a spelling mistake and you are comparing it to a system meltdown I have showed you the spelling mistake, but nobody has been able to show me the molten sensor dripping over the shutter blades.

If I am wrong, I am wrong, my OPINION remains that it's a non issue

[BrerFox]

If I am wrong, I am wrong, my OPINION remains that it's a non issue

Which is precisely why so many people die of diseases that could have been treated if caught earlier.
For me, I would rather wait for the outcome of Canon's investigation. Again, we disagree.

[Grahame Lane]

Andre

I respect Sean as a photographer of vast ability and experience

He has given his opinion on the problem backed up by reseach which he performed and I accept his opinion, which is my right

Quite what your garden boy and the ANC have to do with this situation is beyond me

I will not comment on this matter any futher

[Sean Hearder]

HAHAHahahaha...

I see canon rumors reporting that Canon stopped shipping 5DmkIII's !?

Oh well... still can't get mine to do anything dodgy while you are actually taking a picture...

OK NO Problem Sean - I will take one current release at half price NO issues and no need to send it back

[Forgiss - Sean Nel]

Ok, I spoke to a Canon Rep this morning and was informed that the resolution will be a firmware upgrade. Cameras in transit went back to get the latest firmware and current owners should be able to get the firmware update within two weeks or so. Same time as we are expecting the new bodies...

Which still begs the question: was/is the issue a "Light Leak"? Maybe we will know more in two weeks...

[BrerFox]

Which still begs the question: was/is the issue a "Light Leak"? Maybe we will know more in two weeks...

It certainly does beg the question, doesn't it.
Why would they recall cameras for a firmware change, I wonder? They have never done that before, have they?

[Forgiss - Sean Nel]

It certainly does beg the question, doesn't it.
Why would they recall cameras for a firmware change, I wonder? They have never done that before, have they?

I don't know, can't say... Looking at the responses on the web, it may be a better option for them? As far as I know, the shipments were stopped and returned to the factory before they decided to fix with a firmware. Whether this is actually so, I doubt we will know...

Lets see where it goes...

Another person I spoke to today said that the total light recorded was less than 1/3rd of a stop over 30minutes of exposure. They seem to think that the fix will be a "shift" like what we used to do with film (reciprocity failure) as it's not light recorded on the sensor, but light measured by the light meter.

The rep this morning was quite emphatic that there will be no recall of delivered units.

[BrerFox]

I don't know, can't say... Looking at the responses on the web, it may be a better option for them? As far as I know, the shipments were stopped and returned to the factory before they decided to fix with a firmware. Whether this is actually so, I doubt we will know...

Lets see where it goes...

Another person I spoke to today said that the total light recorded was less than 1/3rd of a stop over 30minutes of exposure. They seem to think that the fix will be a "shift" like what we used to do with film (reciprocity failure) as it's not light recorded on the sensor, but light measured by the light meter.

The rep this morning was quite emphatic that there will be no recall of delivered units.

No - reciprocity failure is very easy to fix in software, assuming, of course that a CMOS sensor suffers from it. (The funny side is that we used to fix reciprocity failure on film with light bleed). Light bleed onto the sensor is more like light-blindness and very much harder to fix. But, the way I understand it is that this effect is limited to the metering, not the sensor. Perhaps an offset is appropriate in software, but it will certainly cost at least one bit of resolution. That is a 50% cut, which sounds a lot, but maybe it's not in real terms.
Pure speculation. If Canon is already working on a fix strategy, it seems as if their investigation is complete and they obviously fully understand the problem. It is a pity that they haven't told us yet - that means it must be with the marketing boys. Smoke and Mirrors. Maybe somebody influential will eventually squeeze a difinitive answer out of someone like Chuck Westfall. I, for one, appreciate honesty above a cover-up. But, the stakes are high, and misconceptions higher.

[Forgiss - Sean Nel]

I think u misunderstood me. I meant simply a compensation similar to what we used to have to do with long exposures on film because of the chemical reaction.

Andre, if I didn't know better I would swear that you are doing it intentionally.

[BrerFox]

No, Sean, I am not disagreeing with you intentionally.
Let me try and be clearer.
Some things are easy to correct in software - like reciprocity failure which follows a predictable graph.
Incorrect data from a sensor cannot be corrected in that manner. In the same way that night-blindness
reduces the sensitivity of your vision overall, flare onto a sensor produces the same uncorrectable result.
Turning the LCD off using firmware would work, as has been proposed earlier, but if the fix were that simple,
the camera would not need a recall.