Hot Pixels / Stuck Pixels (EOS 350D) - a cure?

nweibley said:

If you are using raw, forget it, jpeg only so software is not an
issue.

Click to expand...

Hang on here : did I understand this well : is this "trick" only working if you shoot in JPEG ? How comes ?

it worked for me. I found one white hot pixel with "x" and it appeared even at low ISO and medium short shutter speed (e.g. 1sec) . Then I decided to do the sensor cleaning method for 30 secs. I shot a dark frame at ISO 200 for 10 seconds. Then I tested using the dead pixel test software and it showed 0 hotpixel. Wow! Save my trip to Canon service centre.

nweibley said:

Not working for me at all

Followed directions exactly....
--
-----Nate

http://www.flickr.com/photos/nate-weibley/

Click to expand...

Has anyone tried this with Firmware version 1.01?
--
thezero

now I'm a believer. After owning my XT for a year, it finally developed
a stuck, blue pixel right in the middle of the frame. I figured, what the
heck, and did the sensor clean function. I shot a bunch of fresh files
and brought them up in PS, at 100%. The stuck pixel was gone. You know
the saying, "For those who believe, no explanation is necessary. For
those don't, no explanation will suffice."

Pix Man said:

I am now of the opinion that operating the sensor clean function
with a lens on and a lens cap on the lens, causes the camera to
either map out hot/stuck pixels, or to reset itself, neutralizing
them. Pure speculation of course, but it solved the problem
completely in 30 seconds. (N.B. I did not clean the sensor at any
point and have never done so.)

Click to expand...

--
'Don't sneak up on it - surround it'

Does the elimination also work on RAW files?

I can imagine some processing removing the pixels but RAW has got the unprocessed image / pixel.

Out of the several positive posts, can anyone confirm it has worked for RAW files as well?

After reading this "tutorial" about how to fix stuck pixels I was amazed that such a simple "trick" could fix all those problems people have around stuck pixels. Now while I am new to dSLRs and still learning the intricacy of how they function, I believe I have a simple explanation that does not revolve around "remapping" the pixels...

I dont believe the sensor gets remapped when following the OP's advice. Remapping is a software based process that requires the camera to either have this built into its firmware (ie Olympus) or have the manufacturer or authorized repair center re-program the camera. Just popping open the shutter should not remap any pixels.

Why then are people seeing a difference? I believe the answer lies in how stuck pixels arise...

Stuck or hot pixels are the result of overly charged pixels due to light. The longer the sensor is exposed to light and/or the higher the intensity of light, the more likely you will get a hot or stuck pixel. Some of you may see where I'm going with this already...

Covering the lens blocks light...no light, no charge...no charge no sensor activation....no sensor activation no pixels stimulated - no pixels stimulated no hot or stuck pixels.

One reason why covering the lens and taking photos works is that it provides the sensors and pixels time to cool down. By repeatedly taking photos with light you are constantly stimulating the pixels, thus preventing them from cooling down and becoming unstuck?

Why then didnt removing the battery for 3 days work? Because the pixels were still stuck. Cooling doesnt mean not using pixels, rather using them in a way that turns them off (ie stimulating them with no light). Even with the lens on and viewfinder covered, light is still entering the lens and hitting the sensor (unless you conduct this process in whats called "cave darkness"). The light that is present is enough to stimulate the sensor and pixels. However, it is not enough light to actually turn them on to a level where they will remain stuck. Its kind of like stimulating them to be a "blank" pixel capable of any color (hence the OP's reference to remapping).

You may think of this process as a non-software remapping process but I dont believe it is remotely similar to remapping. Instead, all you are doing by following the OP's guidelines is forcing those stuck pixels to become activated but not charged enough to remain stuck, thus "unsticking" them.

The best example would be flipping on a light switch and the light not going off. You then flip it on again, then off again - still nothing. You then give it one hard flip and the light goes off.

The real test would be conducting the same procedure but with the lens cap off. I wouldnt be suprised if that doesnt work due to too much light hitting the sensor.

Of course, I could be completely wrong. I'm still learning about how the sensor works

Regardless of why it works, if it does, then this was truly a great find!

I think what happens here is that the process I described cuts off all electrical supply to the CMOS chip, and in so doing, the pixels unstick - simple as that. I can't see any remapping going on, but if it works, it works and that's all that matters.

Pix Man said:

I think what happens here is that the process I described cuts off
all electrical supply to the CMOS chip, and in so doing, the pixels
unstick - simple as that. I can't see any remapping going on, but
if it works, it works and that's all that matters.

Click to expand...

Then you should have gotten similar results when you removed the battery as no charge was being supplied to the camera.

Unless the back up battery is supplying some power to the CMOS chip .... At the end of the day what does it matter? It worked for me and some of the other people who have commented on here. If it doesn't work, it didn't cost anything either .....

Why so defensive? I'm not bashing your advice or suggestions. If it does work, as I said earlier, than great! You've found an inexpensive and easy way to correct hot pixels that folks can do themselves. We could always use more people such as yourself who experiment and come up with creative and inventive ways of fixing troubling problems.

I'm just trying to offer a more realistic explanation as to what is happening. Some people may not care, others may. Just saying something works, though, is not enough for others. There are a few people here who have commented that your procedure did not work for them. The question, then, is why? Understanding the process better may help to explain why it works for some and not others and allow everyone to think about how to refine the procedure so it works for everyone.

And yes, you are correct in assuming that a backup battery (is there one? no mention in the manual that I could see) would provide a constant, albeit small, charge to the sensor. However, charging the sensor is not enough to stimulate the pixels. Light is the crucial element needed to stimulate the pixels. I proposed that with very little light, the pixels were stimulated but not to point where they would remain stuck. There was enough light, though, to stimulate them ("refresh" them) and when the power is turned off the pixels are "reset".

In the end, though, as you mentioned, it may not make a dfiference how it works, as long as it does work. I'm juts trying to figure out why it doesnt work for everyone...

thx Pixman!

I first took a test image (ISO 200, kit lens with cap on, 101 seconds in bulb mode.) Then I activated sensor clean for 45 seconds with lens and cap still on. Then another test pic (same conditions). An additional red stuck pixel showed up.
--
http://www.linelight.org/

strikeback03 said:

I first took a test image (ISO 200, kit lens with cap on, 101
seconds in bulb mode.) Then I activated sensor clean for 45
seconds with lens and cap still on. Then another test pic (same
conditions). An additional red stuck pixel showed up.
--
http://www.linelight.org/

Click to expand...

If my explanation holds, leaving the shutter open for 101 seconds, even with the lens cap on, is too long. While minimal light is hitting the sensor (low intensity), the time that the light hits the sensor and stimulates the pixels is too long (high frequency). This effectively cancels out any positive effect you may be able to achieve.

Let the sensor cool and repeat the process but this time dont leave the sensor clean on for more than 20-30sec. You want low intensity and low frequency - enough to stimulate the pixels but not "aggrevate" them to the point where they dont turn off.

yes I agree with the above post, you didn't follow the original instructions ....

that's why I asked if I had it straight - this thread is way too long to pick through for instructions. And I'm looking for few to no hot pixels on 15 minute exposures - it was fine when new, and now a little over a year later has 20-30 hot pixels.
--
http://www.linelight.org/

I posted the instructions at the start of the thread (on page 1). You will not get rid of hot pixels on very long exposures, it isn't a fault, it is a characteristic of dSLR cameras. You should be able to get rid of them though using Photoshop or other image editing software.

It worked for me like a charm. Good job!

Yes, I tried this technique before reading your post, and couldn't believe that a hot pixel in all images was gone. I thought it may have been dust, but when I tried sensor clean with lens on and then checked the image I thought it must be something to do with pixels, but still not sure as to how this helps get rid of hot pixels! Anyone have possible reasons to this?

toadlet said:

Yes, I tried this technique before reading your post, and couldn't
believe that a hot pixel in all images was gone. I thought it may
have been dust, but when I tried sensor clean with lens on and then
checked the image I thought it must be something to do with pixels,
but still not sure as to how this helps get rid of hot pixels!
Anyone have possible reasons to this?

Click to expand...

Try reading my post a little ways up entitled:
"Possible explanation - has nothing to do with remapping"

Great detective work, and, for me, this does work. I also have an EOS 350D Firmware Rev. 1.0.3. It is possible the sensor cleaning function does map these pixels out. Thanks for a very informative solution!