Dust or Glitch?

[matthew lies]

<p>This picture is from the K20 dust alert function. It's pretty shocking, I tried the dust removal function a number of times, checking afterwards to see if it was gone, only to find that the dust moved and made interesting patterns...I put the mirror up so I could see if there was in fact any dust on the sensor....could not see any, tell me what you think...<br>

<a href="http://s47.photobucket.com/albums/f172/liesmattjon/?action=view&current=DUST.jpg" target="_blank"><img src="http://i47.photobucket.com/albums/f172/liesmattjon/DUST.jpg" border="0" alt="Photobucket" /></a></p>

[miserere_mei]

<p>Have you used a blower to try to clean the sensor?</p>

[matthew lies]

<p>Another pic to show what I'm talking about...very disturbing.<br>

<a href="http://s47.photobucket.com/albums/f172/liesmattjon/?action=view&current=DUST2.jpg" target="_blank"><img src="http://i47.photobucket.com/albums/f172/liesmattjon/DUST2.jpg" border="0" alt="Photobucket" /></a></p>

[joe_jackson4]

<p>I think you've found "art" mode... :) Cool stuff.</p>

<p>Not really sure what's going on there, but I wanna see more... :)</p>

[michael_elenko]

<p>Have you tried the typical diagnostic for dust issues by shooting at f22 into a blue sky?<br>

Enlarge the results and then see if anything shows up,</p>

<p>ME</p>

[matthew lies]

<p>Mis: Have not tried a blower, I'll give it a go later.</p>

<p>Paul: I could make a ton of these and they would all look different, all I have to do is get it so people will take an interest, could be the next modern art haha :)</p>

<p>Michael: Hopefully I'll be able to try that this weekend, it's been pretty overcast lately :(</p>

<p>Thanks for the tips!</p>

[a few images]

<p>reminds of the guy who dropped an early digital camera into the water. He dried it out and started to use it again but each photo was crazily modified by the cameras own faults caused by the water intrusion - at the time it was so mysterious, partly because I think folks didn't understand digital cameras in the first place. If I recall he had made 100 or so prints and each sold for some big money.. </p>

[michael_elenko]

<p>Matthew,<br>

Another, an indoor and very effective method, is to create a white screen in say Photoshop. Clean your monitor screen and camera lens. Then set the camera on a tripod in front of the screen, set a fairly close focusing lens (I use a 50mm) to f22, manually focus, and take the shot.<br>

Examine the results in software that lets you adjust contrast to help see blobs etc.</p>

<p>ME</p>

[shots worth sharing]

<p>What's not clear is whether the alert is for real or a false alarm (I'm not sure which is preferable.)<br>

You don't need a blue sky--a blank wall will do. Just set it to MF and shoot with flash (aperture doesn't matter and neither does focus) then look at the result on your computer. Does the image look like the alert?</p>

[pj_vesterback]

<p>Or just a shot with the white body cap (no lens) on should show any dust also.</p>

[rnt]

<p>I don't believe that a white body cap would work. You're looking for something with a very small aperture so the dust is nice and sharp. You need a point-ish source of light to do this because the dust isn't on the sensor- it's on the IR filter a small but significant distance in front of the sensor. I use a body cap with a pinhole when I go dust hunting... It's always a depressing experience.</p>

[shots worth sharing]

<p>Ok, I get it: you're right of course: the smaller aperture gives a clearer image. </p>

[pj_vesterback]

<p>Actually the body cap did work for me. I only did it once but the dirt was extreme. It looked like someone used a crayon on the filter. I have since bought and used a sensor loupe. That can be depressing.</p>

[s_arena]

<p>Matthew,<br />Don't panic. The first time I used this feature of the k20d, I had the same problem. This happens if you do the sensor test in relatively poor lighting. It is most likely NOT dust. Remember the camera is making an exposure for this test. It is an "undocumented feature" that this dust identification routine needs good lighting. Try this: Just go outside and point the camera at a clear sky in the daylight (not at the sun) when you run the test and I think you will see those streaks dissapear, and the image you get back will only be dust specs. You could also point it at a plain wall, but the key is that the lighting MUST be good.</p>

[leo_papandreou1]

<p>This happened to me. The pattern wasn't as big or as geometric as yours but like yours it would migrate when I used the dust removal function. An air blower made it go away completely. I can also confirm what S Arena wrote about good lighting.</p>

[matthew lies]

<p>Ha! I just took this out of my kitchen window, just to see what would happen. Pretty cool if you ask me. :) I did point at blank wall and nothing came up, so I'm guessing there's no dust...good! Thanks for the help!</p>

<p><a href="http://s47.photobucket.com/albums/f172/liesmattjon/?action=view&current=DUST3.jpg" target="_blank"><img src="http://i47.photobucket.com/albums/f172/liesmattjon/DUST3.jpg" border="0" alt="Photobucket" /></a></p>

[s_arena]

<p>That is pretty cool. I think you should submit that photo to PPG with a catchy title like: "Magic Dust"</p>