The hot pixel issue

I have a 20D and 5D. I sent my 20D to Canon for cleaning and hot pixel remapping the weak before the warantee ran out. It came back clean.

It has since developed two very hot pixels (visible at ISO 400, 1/100th). My 5D had a couple too and, as reported elsewhere, running the sensor cleaning mode seemed to get rid of them.

So, for Chuck, if this feature is real (and it seems to be) why isn't it documented? Second, having to send in a camera to have these pixels mapped out is expensive in both time and money. I'd be happy if the 20D could do what the 5D seems to do or if another tool was released that could allow users to map the little buggers out themselves without the time and expense of sending cameras to Canon all the time for a simple table update.

So how about it, Chuck. When will owners be able to fix these little things either in-camera or with a tool like "camera window" or something like that?

Lee Jay

Lee, let's avoid addressing Chuck directly. He will respond to posts that he chooses, and I would rather not "pressure" him with named posts. That was the agreement at the original RG forum where Chuck posted the most. But thanks for the question and info.

Michael Tapes said:

Lee, let's avoid addressing Chuck directly. He will respond to posts that he chooses, and I would rather not "pressure" him with named posts. That was the agreement at the original RG forum where Chuck posted the most. But thanks for the question and info.

Click to expand...

No problem. Might I suggest you consider putting this useful tidbit in the registration "agree to" information so that fewer people make the same mistake I did.

Lee Jay

Lee Jay,

You can also message Chuck via this forum if you would like a personal discussion, be aware that he cannot answer everything. There is also his column at Digital journalist (a good read), you can find Chuck's "tech tips" at http://digitaljournalist.org/issue0606/tech-tips.html

$0.02

Re: hot pixel issue

Hi, Lee Jay:

I commented on this subject a while ago in RG's forums. The gist of it is that Canon Inc. does not confirm that the new firmware for the 5D has anything whatsoever to do with pixel remapping in the camera. My personal tests with a single camera before and after installing 1.0.5 indicate that there was no change at all in the number of hot pixels found through the usual torture test of shooting a blank frame with the camera's body cap on, then analyzing the images with the levels slammed to the edge. So to reiterate, we are not claiming this as a "feature."

Moreover, as an FYI, there is no such thing as a permanent fix for hot pixels. They can and will develop over time, no matter how many times the sensor is remapped. Yes, it can be annoying, but it's also unavoidable.

At the same time, Canon Inc. has indicated that it has no plans to release pixel mapping software to customers, so the only available solution for this problem is sending the camera to a Canon Factory Service Center. Don't shoot the messenger, I am just telling you the way it is, as opposed to the way some of us might like it to be.

Best Regards,

Chuck Westfall
Director/Media & Customer Relationship
Camera Marketing Group/Canon U.S.A., Inc.

By the way, Adobe Camera Raw effectively "maps out" hot pixels in the conversion. This is automatic, always applied, and does not require any noise reduction to be turned on. Thomas Knoll confirmed that this is an "official" undocumented feature of ACR.

Why not the same in DPP? (rhetorical question....I'm hoping Canon continues to develop their software).

Gary Jean said:

By the way, Adobe Camera Raw effectively "maps out" hot pixels in the conversion. This is automatic, always applied, and does not require any noise reduction to be turned on. Thomas Knoll confirmed that this is an "official" undocumented feature of ACR.

Why not the same in DPP? (rhetorical question....I'm hoping Canon continues to develop their software).

Click to expand...

Really? I use RSP and the hot-pixel pattern-noise slider doesn't seem to touch them. Maybe that's an improvement those folks could make that would solve my problem. I'd be thrilled with a way to manually map the little guys out in RSP in such a way that it would detect which camera the image belongs to and apply the appropriate correction.

How does ACR recognize the hot pixels without some dark frames or something?

Lee Jay

Lee Jay Fingersh said:

How does ACR recognize the hot pixels without some dark frames or something?

Click to expand...

Actually it's not a very difficult heuristic. It's a good idea to have it only as an option, though, because the same logic would simply remove stars from the night sky, for instance.

because the same logic would simply remove stars from the night sky, for instance

Click to expand...

If light from a star falls on a hot pixel, is it really a star ?

Stan Jirman said:

Actually it's not a very difficult heuristic. It's a good idea to have it only as an option, though, because the same logic would simply remove stars from the night sky, for instance.

Click to expand...

Indeed, not that difficult. One could apply a variation of a 'speckle' removal method such as the 'Crimmins Complementary Hulling' algorithm.

Bart

Lee Jay Fingersh said:

How does ACR recognize the hot pixels without some dark frames or something?

Lee Jay

Click to expand...

I don't know Lee Jay. But I asked on the Adobe Camera Raw forum if ACR mapped out hot pixels and under what circumstances. Mr. Knoll, who wrote the program, said "Yes. Always."

I have not seen any stars mapped out in ACR!