5D Sensor Clean mode eliminates hot pixels

[Its RKM]

Same effect as
thinking that a button, the sole function of which is to hold the
mirror up and hold the shutter open (after verifying that there is
enough battery power), will map out hot pixels. ;-)

And you, of course, are about to post the disassembled code or other evidence here to prove that is ALL that it does.

The fact is that you can't - you are guessing that is all that happens based on lack of evidence on your part.

Absence of evidence is not evidence of absence.

 

[Kabe Luna]

This urban legend keeps popping up.

Here's what Canon's Chuck Westfall has to say to a poster on the
Pro Photo Home forum:

"Let me put it to you this way, Michael: If Canon Inc. wanted to
acknowledge something like this, they would have done so publicly
in all of the promotional material about the 5D. The fact that they
did not ought to tell you something.

I'm glad that you are pleased with your 5D, but for what it's
worth, my informal testing shows that there is no difference in the
number of hot pixels in several before and after tests."

Canon will remap you hot pixels (free if under warranty) if they
bother you.

--
Gary
http://garyjean.zenfolio.com/

--
- -
Kabe Luna

http://www.flickr.com/photos/kabeluna/

 

[Gena]

No point of arguing with people who sighted a UFO landing or just talked to Elvis.

Unfortunately, the dead pixels cure by "Sensor Clean" has acquired the status of Lazarus rising from the dead. People who have done "tests" or had 'visions' don't have the ability to critically assess them nor enough practice and education to correctly conduct systematic and reproducible tests.

 

[blackhawk13]

--Urban legends and the flying web pigs. This whole notion is Looney Tunes brand insanity.

• Why would the camera do a self calibration of the sensor with bloody light falling on the sensor, even the smallest bit?

It wouldn't, if anything it would do it on the fly, and it does.

It's not the same thing as Canon programming out hot pixels though (which works quit well, my I had that done to my 20D). The sensor is self calibrating to a large extent as the white papers indicate.

What people are yacking about on the Rebel software is not hot pixel cleaning IN the camera; it's post processing on the computer after downloading the images.

-nothing beats a fast lense, except a fast girl-

 

[Robert Deutsch]

People who have done
"tests" or had 'visions' don't have the ability to critically
assess them nor enough practice and education to correctly conduct
systematic and reproducible tests.

I have not seen Elvis, nor UFOs, but I do have eyes: I can see the difference between the number of pixels before sensor cleaning and after, and I can read the pixel count before and after sensor cleaning as reported by the dead pixels test. I also have a Ph.D. in experimental psychology, so I have some inkling of the difference between an informal test such as I and others have done, and systematic, reproduceable research that is published in an academic journal. However, this is a Forum of opinions and observations made by individuals, and there are now enough of these observations that anyone curious about this effect might want to check it out for themselves. Neither I nor, to the best of my knowledge, anyone else who has reported positive findings has claimed anything mystical to explain the effect. The explanation that I find plausible is that this is an undocumented feature that works imperfectly, and thus Canon does not want people to try it and end up having to deal with a lot of complaints. There has been no categorical denial from Canon, just some negative findings from another informal test (not systematic, reproduceable research) by Chuck Westfall, and a suggestion that if Canon, Inc. wanted to acknowledge this effect they would have done so publicly. It does not require religion or mystical beliefs to consider the possibility that a major company like Canon would not want to endorse an undocumented feature that works imperfectly.

Bob

 

[saynomore]

Were you there for the original thread? In any case, I am not saying I believe it, I'm saying it's very plausible. Do you think religious myths and UFOs are plausible? Do you know what "testable", "repeatable" and "falsifiable" mean? If you confuse this issue with religious myth, you don't really know what being an atheist and skeptic is about. Such presumptuous attitude is what gives us the bad name.

I would like to do more tests, but right now I don't have any hot pixels. Either way it goes, it's not worth just dismissing it with such zeal. People in the original thread were interested, and some of them made their own tests. Whatever their results, that alone is one step further than most of the "skeptics" have done. I can only assume you're not one of those. If that's the case, then we must find an explanation for all the others who did have hot pixels, which right after the sensor cleaning went away.

No point of arguing with people who sighted a UFO landing or just
talked to Elvis.

Unfortunately, the dead pixels cure by "Sensor Clean" has acquired
the status of Lazarus rising from the dead. People who have done
"tests" or had 'visions' don't have the ability to critically
assess them nor enough practice and education to correctly conduct
systematic and reproducible tests.

 

[saynomore]

Why are you so quicklly to dismiss it? It is easily testable and very plausible, and people who have tested it actually have gone one step further than you. In the original thread, many were skeptical (which is a perfectly reasonable, even healthy stance) just like you. Even the OP was skeptical, that's why he asked other people to try it. But then they went on and tested it themselves, many of them came back saying it worked.

Now I'm not gonna stand for any side right now, but I don't think your group is being fair. It is indeed plausible. In the original thread, it was pointed out that when you press the sensor clean button, the camera takes about two seconds before putting up the mirror. Not so on the 20D, check it yourself, you have both 5D and 20D. That's probably when it would take place.

In any case, as I said in the post above, I'd like to check it out right now, but I don't have any significant visible hot pixels right now. People were having some hot pixels visible even in correctly exposed ISO 800 shots in normal shutter speeds. Maybe the camera just targets those, not the ones that one sees on ISO 1600 at 1 second with the cap on. In any case, it's worth to check it out.

--Urban legends and the flying web pigs. This whole notion is
Looney Tunes brand insanity.

• Why would the camera do a self calibration of the sensor with

bloody light falling on the sensor, even the smallest bit?
It wouldn't, if anything it would do it on the fly, and it does.
It's not the same thing as Canon programming out hot pixels though
(which works quit well, my I had that done to my 20D). The sensor
is self calibrating to a large extent as the white papers indicate.

What people are yacking about on the Rebel software is not hot
pixel cleaning IN the camera; it's post processing on the computer
after downloading the images.

-nothing beats a fast lense, except a fast girl-

 

[saynomore]

As you can see, it didn't start with just the assumption. People were trying to come up with a theory for their observation that a prominent hot pixel cluster just went away after showing up always since they could remember.

And actually, a guy called MacGyver suggested the solution. His account is pretty believable too. It seems it works for the most prominent hot pixels, taken at a relatively low ISO and normal shutter speed. Reviewing some of my posts during that time, I did have a hot pixel showing up at low-ISO, and then it went away. I'd like to be able to test it now, but my pictures are pretty clean right now. You're not gonna find it works at ISO 1600, 1 second and the cap on.

http://forums.dpreview.com/forums/read.asp?forum=1032&message=16851724

 

[gium]

Unfortunately, the dead pixels cure by "Sensor Clean" has acquired
the status of Lazarus rising from the dead. People who have done
"tests" or had 'visions' don't have the ability to critically
assess them nor enough practice and education to correctly conduct
systematic and reproducible tests.

We are indeed delusional, and we have bad eyesight, or are just too plain dumb to count bright pixels in a dark photo.

one.. two.. four.. five.. seven.
You got me there!

And I wish I could create a reproducible test, but unfortunately my hot pixels are all gone now. That really sucks. I wish I could bring them back to do it all over again to comply to such systematic and reproducible tests you'd like to see.

oh enough with the sarcasm* .

Do you really think it's that hard to do such a test?

step1: you count the amount of hot pixels in photo x and take note of their positions

step2: do something that might get rid of the hot pixels (dance around, pray, call Canon, kiss a frog, whatever).

step3: you count the amount of hot pixels in photo x and take note of their positions
step 4: repeat test if necessary to make sure they're really gone.

If step 3 shows you a decrease in hot pixels, something done in step 2 probably worked. It's as easy as that. It's a matter of being able to "count the dots", it's not rocket science! Relating something as simple as that to mythical references, UFO sightings or even plain vague 'visions' tells me more about your rude attitude and your own ability to 'critically assess' what's going on..

 

[PIXSurgeon]

No point of arguing with people who sighted a UFO landing or just
talked to Elvis.

...You just can't imagine how much FUN I have reading this whole thing...

People who have done
"tests" or had 'visions' don't have the ability to critically
assess them nor enough practice and education to correctly conduct
systematic and reproducible tests.

ALERT! OBSCURANTERS on in flight! You will know what I mean (check the main thread)...

This is SO FREAKING similar to the story of that mysterious "bug" that appeared in everyone's video-camera recordings, with that odd-and-strangely shaped body, all over the world...

Dead-on, Gena.

--

TIP: If you do not like this post, simply press the 'COMPLAINT' button. Mommy/Daddy are just one click away.

 

[saynomore]

and please don't go all CAPS crazy with me... I'm just asking.

No point of arguing with people who sighted a UFO landing or just
talked to Elvis.

...You just can't imagine how much FUN I have reading this whole
thing...

People who have done
"tests" or had 'visions' don't have the ability to critically
assess them nor enough practice and education to correctly conduct
systematic and reproducible tests.

ALERT! OBSCURANTERS on in flight! You will know what I mean (check
the main thread)...

This is SO FREAKING similar to the story of that mysterious "bug"
that appeared in everyone's video-camera recordings, with that
odd-and-strangely shaped body, all over the world...

Dead-on, Gena.

--
TIP: If you do not like this post, simply press the 'COMPLAINT'
button. Mommy/Daddy are just one click away.

 

[blackhawk13]

Now I'm not gonna stand for any side right now, but I don't think
your group is being fair. It is indeed plausible. In the original
thread, it was pointed out that when you press the sensor clean
button, the camera takes about two seconds before putting up the
mirror. Not so on the 20D, check it yourself, you have both 5D and
20D. That's probably when it would take place.

In any case, as I said in the post above, I'd like to check it out
right now, but I don't have any significant visible hot pixels
right now. People were having some hot pixels visible even in
correctly exposed ISO 800 shots in normal shutter speeds. Maybe the
camera just targets those, not the ones that one sees on ISO 1600
at 1 second with the cap on. In any case, it's worth to check it
out.

--Urban legends and the flying web pigs. This whole notion is
Looney Tunes brand insanity.

• Why would the camera do a self calibration of the sensor with

bloody light falling on the sensor, even the smallest bit?
It wouldn't, if anything it would do it on the fly, and it does.
It's not the same thing as Canon programming out hot pixels though
(which works quit well, my I had that done to my 20D). The sensor
is self calibrating to a large extent as the white papers indicate.

What people are yacking about on the Rebel software is not hot
pixel cleaning IN the camera; it's post processing on the computer
after downloading the images.

-nothing beats a fast lense, except a fast girl-

--I hate BS artists and trolls. Use your head. Why would a sensor be calibrated with an unknown level of light falling on it? It wouldn't be!

This isn't rocket science. Maybe the 5D does recalibrate on boot up or during operation. Canon probably wouldn't disclose this as it's normally done automaticly. Canon would say "do a sensor clean" to recalibrate the sensor especially in very cold or hot environment. They don't because that's not how that camera works, plain and simple.

-nothing beats a fast lense, except a fast girl-

 

[saynomore]

Now I'm not gonna stand for any side right now, but I don't think
your group is being fair. It is indeed plausible. In the original
thread, it was pointed out that when you press the sensor clean
button, the camera takes about two seconds before putting up the
mirror. Not so on the 20D, check it yourself, you have both 5D and
20D. That's probably when it would take place.

In any case, as I said in the post above, I'd like to check it out
right now, but I don't have any significant visible hot pixels
right now. People were having some hot pixels visible even in
correctly exposed ISO 800 shots in normal shutter speeds. Maybe the
camera just targets those, not the ones that one sees on ISO 1600
at 1 second with the cap on. In any case, it's worth to check it
out.

--Urban legends and the flying web pigs. This whole notion is
Looney Tunes brand insanity.

• Why would the camera do a self calibration of the sensor with

bloody light falling on the sensor, even the smallest bit?
It wouldn't, if anything it would do it on the fly, and it does.
It's not the same thing as Canon programming out hot pixels though
(which works quit well, my I had that done to my 20D). The sensor
is self calibrating to a large extent as the white papers indicate.

What people are yacking about on the Rebel software is not hot
pixel cleaning IN the camera; it's post processing on the computer
after downloading the images.

-nothing beats a fast lense, except a fast girl-

--I hate BS artists and trolls. Use your head. Why would a sensor
be calibrated with an unknown level of light falling on it? It
wouldn't be!

Did you read anything? The mirror does not lock up until after a second or two on the 5D. This doesn't happen on earlier cameras like the 20D and Rebels. If you are willing to admit that the 5D is capable to map out on boot, then why can't you admit at least this is possible ? That's the only thing I'm saying. It's possible, and plausible, in my opinion. Read the original thread I linked below for further reference.

What would anyone gain by convincing people this is true? IF this is true, there can be a number of reasons that Canon doesn't list it for, but that they don't list it is not an argument. There is no reason to be so hostile towards people who have seen this and tested it. By now I am assuming that you, Gena and Pixsurgeon have done very scientific tests to so bluntly and mightily put honest people down.

This isn't rocket science. Maybe the 5D does recalibrate on boot
up or during operation. Canon probably wouldn't disclose this as
it's normally done automaticly. Canon would say "do a sensor
clean" to recalibrate the sensor especially in very cold or hot
environment. They don't because that's not how that camera works,
plain and simple.

-nothing beats a fast lense, except a fast girl-

 

[Gary Jean]

And actually, a guy called MacGyver suggested the solution. His
account is pretty believable too.

Oh my feakin' word! :-O
--
Gary
http://garyjean.zenfolio.com/

 

[Steven Noyes]

And then we will the option to do it ourselves. Focus adjustments anyone???

Steven

--
---
2006 White Sands and Bisti Workshop
http://www.pbase.com/snoyes/white_sands_and_bisti

Winter 2007:
http://www.pbase.com/snoyes/images_winter_2007

 

[saynomore]

His name does not matter. I know who McGyver was. By the way, he didn't use any rubber bands and paper clips.

And actually, a guy called MacGyver suggested the solution. His
account is pretty believable too.

Oh my feakin' word! :-O
--
Gary
http://garyjean.zenfolio.com/

 

[Benedict Slotte]

I can tell you that I have "archived" my 5D's hot pixels since I got it in December 2005. Some 2-3 times a year I have taken a few pictures with the lens cap on in reduced light, and archived the test pictures as both RAW and JPG. In spite of the occasional sensor cleaning, the number of hot pixels is rising linearly with time. Very few of them have gone away, but many new have appeared. As far as I understand, hot pixels may sometimes also disappear spontaneously without any sensor cleaning mapping built-in. Another thing to factor in is the effect of the temperature during the test (which in turn will most likely depend on how many tens of seconds it is since the last picture was taken). If sensor cleaning were effective at mapping out hot pixels, not even nearly as many new ones would appear (or rather, the long-term average number of hot pixels would remain constant - some new and some gone).

So from the evidence I have, I would say with 90-95% certainty that "the myth is busted".

See also a previous thread by me here (unfortunately the gallery picture is no longer available for view):
http://forums.dpreview.com/forums/read.asp?forum=1032&message=18844913

--
B. Slotte
Turku, Finland
http://bslotte.smugmug.com
See profile for equipment

 

[blackhawk13]

Now I'm not gonna stand for any side right now, but I don't think
your group is being fair. It is indeed plausible. In the original
thread, it was pointed out that when you press the sensor clean
button, the camera takes about two seconds before putting up the
mirror. Not so on the 20D, check it yourself, you have both 5D and
20D. That's probably when it would take place.

In any case, as I said in the post above, I'd like to check it out
right now, but I don't have any significant visible hot pixels
right now. People were having some hot pixels visible even in
correctly exposed ISO 800 shots in normal shutter speeds. Maybe the
camera just targets those, not the ones that one sees on ISO 1600
at 1 second with the cap on. In any case, it's worth to check it
out.

--Urban legends and the flying web pigs. This whole notion is
Looney Tunes brand insanity.

• Why would the camera do a self calibration of the sensor with

bloody light falling on the sensor, even the smallest bit?
It wouldn't, if anything it would do it on the fly, and it does.
It's not the same thing as Canon programming out hot pixels though
(which works quit well, my I had that done to my 20D). The sensor
is self calibrating to a large extent as the white papers indicate.

What people are yacking about on the Rebel software is not hot
pixel cleaning IN the camera; it's post processing on the computer
after downloading the images.

-nothing beats a fast lense, except a fast girl-

--I hate BS artists and trolls. Use your head. Why would a sensor
be calibrated with an unknown level of light falling on it? It
wouldn't be!

This isn't rocket science. Maybe the 5D does recalibrate on boot
up or during operation. Canon probably wouldn't disclose this as
it's normally done automaticly. Canon would say "do a sensor
clean" to recalibrate the sensor especially in very cold or hot
environment. They don't because that's not how that camera works,
plain and simple.

-nothing beats a fast lense, except a fast girl-

Did you read anything? The mirror does not lock up until after a
second or two on the 5D. This doesn't happen on earlier cameras
like the 20D and Rebels. If you are willing to admit that the 5D is
capable to map out on boot, then why can't you admit at least this
is possible ? That's the only thing I'm saying. It's possible, and
plausible, in my opinion. Read the original thread I linked below
for further reference.

What would anyone gain by convincing people this is true? IF this
is true, there can be a number of reasons that Canon doesn't list
it for, but that they don't list it is not an argument. There is no
reason to be so hostile towards people who have seen this and
tested it. By now I am assuming that you, Gena and Pixsurgeon have
done very scientific tests to so bluntly and mightily put honest
people down.

--Allow me to retort.

Sensor cleaning is the worst time to calibrate; the camera is less shielded against EM.

Just because there's a delay, people assume any suggestion made. You ever think the delay is because the camera is doing a current check on the battery? On a almost depleted battery voltage/ma can spike to a near normal level and then rapidly fall off. Thus opening the shutter instantly could cause a false reading ending with bad result for the users.

That makes more sense, and is probably the answer to this behavior. There did I maybe solve this little perpexing riddle for you?

-nothing beats a fast lense, except a fast girl-

 

[saynomore]

Sensor cleaning is the worst time to calibrate; the camera is less
shielded against EM.

So, as I was saying, there is no difference during that second or so before the mirror locks up. Now...

Just because there's a delay, people assume any suggestion made.
You ever think the delay is because the camera is doing a current
check on the battery? On a almost depleted battery voltage/ma can
spike to a near normal level and then rapidly fall off. Thus
opening the shutter instantly could cause a false reading ending
with bad result for the users.

Well, have you ever tried to do a sensor clean on the 5D with a low battery? it won't even let you into the sensor clean menu. That's one step before you actually press the button to do a sensor clean.

That makes more sense, and is probably the answer to this behavior.
There did I maybe solve this little perpexing riddle for you?

So it doesn't really make more sense, since the camera with a low battery won't even let you to the point where you can press the clean sensor option.

And besides, that's not the point. There's then the whole lot of people who have indeed found that their clusters of very prominent hot pixels went away after doing this.

Come on, you should admit it's at least possible. I'll go as far as to say it's very probable. And in any case, this whole issue didn't deserve the demeaning responses it has gotten.

 

[Octane]

--
-------David-------
http://flickr.com/photos/childish/