HELP!!!! 10D Burnt Pixel on CCD!!?!!!!

[45R]

I really need some input guys. I took a few pictures of the eclipse tonight and I noticed a purple shot in the same spot in all my pictures. This spot is really notable when shooting on black but with other colors its not as bright. What is it!?!! I checked the CCD for spots. Its clean!

I really appreciate anything suggestions!

http://www.pbase.com/image/35608535.jpg

[FlipsidE]

How long was the exposure? I'm just curious. On exposures longer than 1/10, these things start showing up. It's no big deal. These are what is referred to as "hot pixels." Even the 20D's have 'em on long exposures. Just clone it out if it gets in the way.

FlipsidE

[defordphoto]

Just a side-comment: The 10D does not have a CCD. CMOS it does have. Hotspots happen. Use the heal tool and all is well.

[45R]

The pictures were over exposed at .6 seconds others that I noticed were exposed at 1/8 of a second.

[FlipsidE]

The pictures were over exposed at .6 seconds others that I noticed were exposed at 1/8 of a second.

- Sounds about right then. Nothing to worry about.

FlipsidE

[45R]

OKay so if this thing is a hot spot and I can use the heal tool, thats cool but will it show up at the same spot each time?

I am pretty new to this hot pixel thing.

[dschwartz69]

It is common amongst almost all digital cameras - you will see it at the exact same spot when it (or they) appear.

I freaked out a couple months ago when I first noticed mine on my, then new, 10D.

No worries...

[45R]

Thanks for the help guys. I was about to lose it :) Sending something for repair at Best Buy is a nightmare!

[Persian-Rice]

Burnt photosites will appear black. Hotspot, which basically means the photosites amplification settings have gone out of wack and has created much too much heat will give purple, orange, yellow, red, green or blue tones.

Some will appear all the time, known as stuck photosites or just random ones which is normal, especially at higher ISO's as amplification is magnified even more.

As said, just create an action that will auto repair this. One of my 10D's does this at 400-800+ ISO in one of the corners, especially if there is something dark in that area which will make it more obvious. Anyway, it's really easy to fix.

Believe it or not, if you operate the camera under cooler conditions, hotspots will be far less noticeable.

Cheers.

[Ed Rotberg]

You can also send the camera into Canon to have the hot and dead pixels remapped. It probably doesn't cost a whole lot. If you ever have a time when you aren't going to be needing your camera for a week or so, you might think about this.

Just about every CMOS/CCD sensor has these - even right off the line. That's why mapping out bad pixels isn't a big deal. It's almost certain that your camera came from the factory with a number of thse already mapped out. The yield of 100% perfect sensors would make the cost of a digital SLR prohibitive. On the bright side, mapping out a handful of pixels out of 6+ million is pretty much not noticable.

= Ed =

[Andy_T]

Ed,

that sure sounds interesting!
Do you have further information (e.g. sources or testimonials from users who had it done) on this how it is technically done and where at Canon I could send my camera?

Best regards,
Andy

[Ed Rotberg]

Ed,

that sure sounds interesting!
Do you have further information (e.g. sources or testimonials from users who had it done) on this how it is technically done and where at Canon I could send my camera?

Best regards,
Andy

Well,

I had it done to my 10D in June. With SLR camera I have purchased since about 1978 or so, just before the camera goes out of warranty I send it in for cleaning/adjustment, etc. Since th 10D was my first DSLR, I requested that the camera also be checked for dead/hot pixels and that they remapped them. IIRC, the service record came back indicating that the remapping had been done, and I know that I checked for the one bad pixel I had noticed and it appeared to be gone.

As for technically, I've read somewhere that the software in the camera keeps a table of known bad pixel locations. It then substitutes an average of the actual data from the surrounding pixels for those locations. I've been a professional graphics programmer for over 30 years, and this seems like the obvious approach, so when I read this, it made total sense. My guess is that they use a weighted average of the surrounding pixels, giving the more weight the the axially aligned adjacent pixels than the diagonally adjacent pixels, but that is pure specualtion. That's how I would do it though.

Just send the camera into any Canon Service Center (I sent mine into the Irvine, CA center) with a letter indicating the service you wish performed, and, if appropriate, the receipt indicating that the camera is still under warranty. You might as well ask them to do a general tune-up, sensor cleaning, etc. while it is there.

= Ed =

[Andy_T]

Ed,

that is very interesting and useful information, thank you 8)

I'll inquire whether Canon can also do that on my G2, which has several hot pixels by now. It might be slightly out of warranty, though :wink:

Best regards,
Andy

[timmyquest]

Ed,

that is very interesting and useful information, thank you 8)

I'll inquire whether Canon can also do that on my G2, which has several hot pixels by now. It might be slightly out of warranty, though :wink:

Best regards,
Andy

Indeed, i always thought of it as a hardware only problem. I never figured it could be managed through software

[Persian-Rice]

Ed,

that is very interesting and useful information, thank you 8)

I'll inquire whether Canon can also do that on my G2, which has several hot pixels by now. It might be slightly out of warranty, though :wink:

Best regards,
Andy

Indeed, i always thought of it as a hardware only problem. I never figured it could be managed through software

Amplification of the signal is all software controlled *eherm eherm* it's called ISO setting. Therefore the software can actually reduce amplification. I am assuming they are using some form of sensor system that records temperature readings and as Ed said, averages out using some mathematical formula.

Another reason that the "sensor tracking" system makes sense is. I really doubt Canon want to sit there and run hot or dead pixel tests and then start going through an image trying to segregate pixels and find the corresponding photosite on the sensor. I also think it would make sense to do that, since problematic pixels are not uncommon.

[45R]

I'm pretty bunked about the hot pixel but I guess I'll just send it over to Canon in Irvine to have it remapped. I'll have them clean the sensor also.

[Levitron]

I'm curious....so does this mean that if you reflash your camera's firmware, then the "software memory" of where those hotspots are, will be gone?

[Ed Rotberg]

I'm curious....so does this mean that if you reflash your camera's firmware, then the "software memory" of where those hotspots are, will be gone?

I seriously doubt that that memory is contiguous with the program address space. In short, it seems very unlikely. That would be absolutely horrible design.

= Ed =