lamiaceae Loc: San Luis Obispo County, CA

Hi guys (and gals):

Anyone seen this before or have any experience with this? I have been noticing a dark line with some rainbowing running vertically near the left edge of the landscape frames from one of my cameras, my secondary one. The line seems to be essentially one pixel wide but affecting the ones on either side of it as well. It does not seem to appear in all images but nearly all lately when it runs through a dark area in the image; affecting both light and dark pixels. Yet images that seem all light in tone in the area where the line might appear seem to be missing it. See the attached image file. It is only a small portion of a whole frame only featuring the problem area. And I added a little arrow with PS CS6.

My camera is a Pentax K-20D (2009 vintage). I tried running the Pixel Mapping application of my camera but that had no effect. Luckily it is about 5% from the left or bottom of my images so I can shoot to crop it out later or I may be able to use the Photoshop Clone stamp to take it out if it does affect a needed area. My other camera is a Pentax K-5. Perhaps it IS TIME to think about a Pentax K-5 II / K-5 IIS / K-3! A shame though as my K-20D was purchased new and has not actually been used all that much. Any thoughts as to what I am experiencing and what may be done about it? Thank you.

(Download)

watchcow Loc: Moore, Oklahoma

Does this show up in raw images as well? It really looks like you have a stuck pixel in your sensor. Since the sensors are read by row or column, a single failed transistor will result in the rest of the pixels in the row downstream from the failure unreadable or they will read as random noise. When this happens due to a failed transistor in the sensor, there is no cure other than replacing the sensor. It might be worth a call to Pentax. they may have had a bad batch of sensors and know of this issue and are willing to make repairs affordable or an upgrade/replacement affordable.

lamiaceae Loc: San Luis Obispo County, CA

Thank you for the information. Yes, it shows up in both Raw (the way I usually shoot) and in test JPG's. The problem seems rather strange though. The line is at 12.6mm from the left edge of a full sized unprocessed 262.8mm x 3995.5mm image. So in most cases I could probably crop it out or possibly retouch it out if it were in an important part of the image. As you can see in this image shot at ISO 200 I get a black vertical line with color fringes so it looks like a row of dead pixels or at least as you said one dead one affecting the entire row. And yes it does show in a Sensor Dust Alert test as an irregular broken line. Not a solid black line though. That seems odd. Also doing a high ISO (lens cap on 30s exposure shows nothing but black and random color noise, but no line. Also a test image taken at ISO 64,000 shows a noisy image with a lighter than most of the image line (see second image). That again seems odd. A dead line of pixels should be pure black not middle to lighter tones at high ISO and little light. Very strange it seems to me. Could it be possible that some tiny piece of debris is on the sensor that has some unusual active property affecting the sensor below the cover Hi/Low Pass & AH filter? Say a speck of magnetic Iron Oxide or Iron Metal, or some other mineral / chemical speck that is photo-electric Cadmium Sulfide it self causing say a discharge into a pixel (say "shorting out" a line of pixels)? I have seem postings to the Pentax Owners Forum about similar lines but the details seem different. A few people had "under exposed" bands along the left side. Time to physically clean my sensor carefully perhaps. Oh, by the way, both image fragments posted just happen to be of out of focus areas of the background of the images. Not indicative of my photography. LOL Thanks again.

(Download)

watchcow Loc: Moore, Oklahoma

If it were solid black i would worry more. if you look around at the technology that drives most CMOS and CCD single sensor cameras, they use a filter layer that looks like a colored checkerboard. your sensor only sees tonal values, not color. it is by sampling these tonal values around a single sensor site that the camera makes a decision based on the strongest light response from these filtered sites which color that pixel in the middle is supposed to be. your "raw" sensor data is really not raw. it is still subject to this oversampling that determines colors. So even a row of dead black pixels will get assigned a color because the system that does this sampling will assign a value to that location even if the information from that pixel site was blank.

http://en.wikipedia.org/wiki/Bayer_filter

lamiaceae Loc: San Luis Obispo County, CA

Thanks, the Bayer article was interesting. I've seen something similar long ago, say before I had gone digital with my photography. Not sure I get why Mr. Bayer would bother mimicking the physiology of the human eye since we "see" in our minds "balanced light." Our brains have auto-white balance so to speak, we don't notice more green sensitivity in vision if it is there. We notice more opposite color pairs, R-C, G-M, B-Y.

Wow, a RGBW or R-G-B-IR cell would be cool. I do Infrared photography with a converted digital camera. It would be great to have a "four" color image. As it is the Green Channel is used to record the IR. So you end up with a weird R-M-B-Y image once the B & R color channels are swapped in Photoshop. The biggest pain or issue with IR is you need to set a custom WB. This is with a 665nm filter that is, ones filtered closer to 880nm give you essentially a IR B&W image like IR film. Some people with the cash for more specialized lens and filter equipment have both the high & low pass filtration removed and shoot both IR and UV. UV photography involves more issues so I'll leave that here.

Like I said, luckily the problem is very close to the left edge (only, so far), so I can just crop the line out or try some other Photoshop tricks to get around it.

watchcow Loc: Moore, Oklahoma

It's still worth a call to Ricoh/Pentax to see if they have any ideas or will be generous enough to cut you some slack. Remember, if you don't have, it might because you didn't ask. Nikon repaired a D70 for me for the cost of shipping when that camera was 5 years old because the fault that developed was a known issue and they had made a decision to support those cameras because of a poor choice of components in the card reader circuitry.