The Black Dot problem with the 5D Mark II had been brought up in the main 5Dmk2 thread. Unfortunately, with over 178 pages... it was barely discussed and eventually drowned by numerous other trivial discussions (memory cards, shutter sounds, availability announcements, etc). I thought the problem warranted more attention than that.

The issue has been discussed in this thread at DP Review. Clusters of black dots mysteriously appear on the right side of blown highlights from small light sources, visible even in normal prints. What's disturbing is that the black dots are not a JPEG artifact, but instead it's recorded in actual RAW data. This problem has been seen in many samples at DP Review, at Photo.net, and even at Vincent Laforet's website.

Why would these black dots only appear on the right side? Why are they even there at all? Does this look like something that could be fixed with a firmware update?

I'm hoping that someone with a technical background could offer some theories on the possible causes of this phenomenon.

maybe i'm being simplistic but this should definitely be something that will be handled in the firmware. even if its something hacky along the lines of the camera averaging the colour of the pixel from nearby ones

i doubt this will warrant any replacement programmes when software (cheap to distribute) could solve it

As to probable cause - it smells of some number overflowing and going from a very large number (white) to a 0 (black) but again this is just an ubstantiated hunch

Here is a damning sample that shows how bad it can get.

ouch.

This looks somewhat similar to the issue that was present in 1D II when it was released. There was, as far as I know, a firmware update that took care of it.

Yikes. Looks like something out of an old school video game. Hopefully there'll be a firmware fix released for this.

Raw files processed how?

Looks like it might be a Raw processing issue. Early versions of Raw converters?

So is this an issue for super blown highlights only?

I was out taking photos of plenty of neon-lighting fixtures last night where there was a lot of very sharp contrast between light and dark and, while I didn't blow any of the neon highlights, I didn't see any black dots anywhere either.

jcbenner wrote:
Looks like it might be a Raw processing issue.


It's not a RAW processing issue. As I wrote in the original message, the black dots appear in the RAW data itself. There's a link up there showing the black dots in the actual RAW file before any demosaic process is even done.

I remember Chuck Westfall saying that Canon altered the way signal readouts are done in the 5Dmk2 sensor (compared to the 1Dsmk3 sensor). Perhaps this black dot phenomenon is a unforeseen side effect of that. I really hope this is something that could be fixed in a simple firmware update, as some people have already implied. Otherwise if it's a hardware problem, then Canon will have another embarrassing 1Dmk3 fiasco in their hands. Two major blunders in two years may be enough to put the final nail in their coffin, as far as the Canon's quest for redemption is concerned.

Question is, why didn't Canon catch this before release?

It does sound like it could be an easy firmware fix as long as the idiots in canons management pull their heads out of the sands and don't do a mkIII over this. It's such a shame given how the chip in the 5D mkII is potentially the worlds best in any DSLR.

I checked my 1Ds3 files and it does not show anything like this at all. Glad I waited on the 5D2 this time.

Is there any RAW files exhibiting this issue available for download?

Rubber Soul wrote:
Here is a damning sample that shows how bad it can get.


That would have been such a great photo if it weren't for the black dots!

I guess this it why I didn't pick up one of the first ones...Thanks for posting about the issue. Looks like a quick firmware fix to me...

Dan Martin wrote:
That would have been such a great photo if it weren't for the black dots!

Is this a nightmare? Please, someone pinch my arm

Dan Martin wrote:
Rubber Soul wrote:
Here is a damning sample that shows how bad it can get.


That would have been such a great photo if it weren't for the black dots!

Any estimate on the actual number of people who have observed this with THEIR 5DII bodies?

I propose a "Chicken Little" forum where threads like this can go until the "issues" they refer to are confirmed, in some objective and convincing manner, to be real and widespread.

this one will be a little harder for canon to dispute if it is wide spread,, seemed the 1dIII af thing they tried to continue to dismiss as it was so hard to rule out user error,

I see the poll coming...

>>> Will you cancel/return your 5D II because of Black Dots ? <<<

Rob will then publish a multi part black dot expose, and on and on...

Well, it's been confirmed: the 5DmkII is a portal to the other side for UFOs and thus the black dots show up when the aliens are coming and going.

Sounds like a Canon conspiracy.

Hmmmm, could it be the new local micro contrast enhancement feature... , highly secret stuff

Wasn't long ago everyone wanted a blue dot,
then yellow was even better.

Black dots just get no respect.

In all seriousness,

This image shows them:
http://glubsch.files.wordpress.com/2008/11/_mg_0166.jpg

On the lower left side is where I looked. Could be else where too.

Almost looks like the same places where CA would show up.