1. Reminder: Please use our affiliate links for holiday shopping!

Hot/stuck pixels on Olympus sensors

Discussion in 'Olympus Cameras' started by Kirill, Jun 20, 2013.

  1. Kirill

    Kirill Mu-43 Regular

    78
    Feb 13, 2013
    Tallinn, Estonia
    Almost every Olympus camera I have tried/used since almost 10 years ago had at least one cluster of defective pixels on the sensor, those that refuse to stay black when you shoot a black frame with the lens cap on (at the lowest ISO at about 1/30) - and you see them. I do prefer Oly m4/3 cams to Panny ones due to many reasons, but finding one that is OK in this respect is really a challenge.
    A few days ago I dropped my E-PM1 from about 1 metre and the chair even cushioned the fall so I was shocked to discover the sensor dislodged on two axes and the shutter no longer operational. Repairing this would have probably cost as much as a new E-PM1, it is also not a straightforward arrangement in my country and I could have ended with a poor replacement sensor, so I just wrote the camera off and headed to the shops to buy a new one. A simple enough task, right? In fact I had to trawl through several outlets all over town before finally, in a state of utter frustration, I found a good one. Two days and about a dozen E-PM1s later I finally have as good a unit as the one that perished. I reiterate: I had to go through a dozen units (different retail chains, hence different batches) before finding one that showed no discernible hot/stuck pixels (most had more than one cluster). Surely this has got to be some kind of a record??
    For testing I used the old little programme called DeadPixelTest (change the threshold in the left-hand box from 60 to 30 as this is empirically the border value for defective pixels that will be visible to you in real-life dark images), I am sure there are other similar software options out there instead of relying on your eyes.
    What I am really curious about is the 16-megapixel sensor in the newer Oly cams, I have not ventured into that territory yet but plan on doing so as the E-PM2 prices in Europe are finally beginning to drop towards the 400-euro mark. If you actually bother testing the cams you are buying for hot/stuck pixels, what has been your experience with newer 16 MP Olympus models?
     
  2. entropicremnants

    entropicremnants Mu-43 All-Pro

    Jul 16, 2012
    John Griggs
    It's funny, but every Nikon, Panasonic and Canon I've ever had did it too.

    Oly uses Panasonic sensors except on the E-M5, E-PM2, and E-PL5. Then they use Sony, just like Nikon.

    EVERY sensor can do this, and it can be permanent, or come and go.

    Your unit with no stuck pixels? It will eventually develop them or so say the odds.
     
  3. keko

    keko Mu-43 Rookie

    23
    May 19, 2013
    Spain
    All and I repeat All the cameras I've owned so far have shown hot pixels, and I've owned quite a few... Nikon D700,D300,D200,D50,D40 Canon 400D Fuji S5 Pro among others...

    Some cameras such as the Nikons correct them in their jpg engines, but in RAW they tend to show up as you raise ISO.

    I don't think it is an Olympus or E-PM1 specific problem.
     
  4. entropicremnants

    entropicremnants Mu-43 All-Pro

    Jul 16, 2012
    John Griggs
    Further reading from the master of curmudgeonly technicalness, Ken Rockwell:

    Hot Pixels

    Notable quote: "Every camera does this..."
     
  5. uci2ci

    uci2ci Mu-43 All-Pro

    Jun 22, 2012
    Los Angeles, CA
    Sam
    Every sensor comes with dead/stuck pixels when manufactured. Oly pixels maps it before sellig it. But chances are that high energy particles hit the sensor before it arrives in your hands. Finding a pixel perfect sensor is an illusion: it was already pixel mapped and it will develope more stuck pixels over time. The Epm1 should have a pixel mapping function that you can use
     
  6. Djarum

    Djarum Super Moderator

    Dec 15, 2009
    Huntsville, AL, USA
    Jason
    I remap mine about every 6 months.
     
  7. Ian.

    Ian. Mu-43 All-Pro

    Mar 13, 2013
    Munich
    Ian
    They are not dead pixels. They are only sleeping.
     
  8. madmaxmedia

    madmaxmedia Mu-43 Veteran

    335
    Feb 20, 2010
    Don't you just remap them on Oly bodies, and then forget about it?

    I think Adobe apps now even automatically remove them when you open raw files.
     
  9. entropicremnants

    entropicremnants Mu-43 All-Pro

    Jul 16, 2012
    John Griggs
    Won't solve intermittent hot pixels. Since they are circumstantial you can't quite sure when they'll be there. Others may be there all the time.

    It's just a fact of life. I have yet to have a picture seriously affected and I used dark frame subtraction always for long exposures. But then I may have more patience than most.

    I did 4 or 5 minute exposures in a graveyard at night in Todmorden, West Yorkshire -- that's a long wait with dark frame subtraction but I liked the shot so it was worth it to me anyway lol.

    <a href="http://www.flickr.com/photos/teamgriggs/6238325657/" title="Cemetary at Night, Todmorden by Entropic Remnants, on Flickr"> 6238325657_ff6ca9ac53_b. "1024" height="700" alt="Cemetary at Night, Todmorden"></a>
     
  10. madmaxmedia

    madmaxmedia Mu-43 Veteran

    335
    Feb 20, 2010
    Ahh, yes- intermittent hot pixels from longer exposures are indeed different. For those the in-camera dark frame subtraction seems to do pretty well? (I don't do a lot of long exposures, although I really like your photograph!)

    But I think OP was talking about the alway stuck/dead/hot pixels if he's testing at 1/30 second exposure. I'm surprised he saw so many bodies with actual clusters of pixels.
     
  11. RobWatson

    RobWatson Mu-43 Hall of Famer

    Maybe I misread but he is not "seeing" them some piece of software he uses "analyzed" the image and "found" them.
     
  12. entropicremnants

    entropicremnants Mu-43 All-Pro

    Jul 16, 2012
    John Griggs
    I may have missed what he was talking about too, perhaps. But I just think the extremity of his situation is hard to explain. The sensors just aren't THAT bad.

    If he is using analytical software, it was probably designed by somebody over on DPReview's forums and therefore completely unrelated to real photography... :rolleyes:
     
  13. Mikefellh

    Mikefellh Mu-43 Top Veteran

    939
    Jun 7, 2012
    Toronto, Canada
    First dead pixels would be permanently black, stuck pixels are permanently on, and hot pixels are what we have here. Here's the ext as described in my E-300 manual:

    "During shooting at slow shutter speeds of 30 or more seconds, noise may appear on- screen or the image may be overly bright in the top left part of the screen. These phenomena are caused when current is generated in those sections of the CCD that are not normally exposed to light, resulting in a rise in temperature in the CCD or CCD drive circuit. This can also occur when shooting with a high ISO setting in an environment exposed to heat. The NOISE REDUCTION function helps reduce this noise."

    As others already stated, this a common phenominom with ALL brands and sensors, and not limited to Olympus. Proof is the quote above refers to CCD sensors, since the E-330 Olympus has been using Panasonic NMOS sensors, and the newest cameras use Sony's CMOS.
     
  14. Kirill

    Kirill Mu-43 Regular

    78
    Feb 13, 2013
    Tallinn, Estonia
    OK, maybe I need to clarify a little.
    1) Of course I understand that cameras/sensors of all manufacturers are prone to this issue, I am not talking in absolutes here - but in degrees. I have been using various digicams for over a decade now and have indeed witnessed the issue not just in Olympus cameras, but statistically and in my personal experience, Olympus cameras seem to be the worst in this respect.
    2) That helpful software I came to use (and I will gladly post a link later today) does not concoct results, it merely and reliably substitutes my own eyes in the task of finding problem pixels. When it shows no such pixels at the aforementioned threshold then I can be certain that the camera has no discernible hot/stuck pixels at normal (handheld) exposures at base ISO, i.e. in normal shooting conditions, simple as that. At longer exposures NR kicks in, no issues there.
    3) As for all cameras developing such pixel clusters over time, let me ask you this: which of the following two cameras is more likely to have several discernible clusters one year after purchase - the one that had none at first or the one that already came with several? :)
    4) I don't get this "remove in PP" argument. Why waste time on PP for something that can and should be avoided at the moment of photography if you chose your camera well? :)
    5) The pixel remapping function built into the Olympic cameras is utter cack. I used it on many occasions over many years in various Oly cameras, faithfully following the instructions and always achieving zero results.
    I think that covers about every point raised by those who were kind enough to try and help me out here :)
    At the end of the day I just want to know how widespread, based on your experience, this issue is in the latest crop comprising E-PM2 and E-PL5 because to get the best price I will probably have to order online from another country and I hate the hassle of returning stuff.
     
  15. Brian Beezley

    Brian Beezley Mu-43 All-Pro


    I bought an E-PL2 three weeks ago. After using it for a few days I spotted a bright white pixel in one image at 100% crop. I looked a little more carefully and found a few more errant pixels, some colored. So I ran the pixel mapping function. The offending pixels disappeared. After reading your post, I decided to check more methodically. I shot a fully dark frame at 1/30 second, amplified the residual noise with contrast/brightness, and examined the entire image by panning it across my monitor at 100% crop (RawTherapee has Navigator, a graphical crop tool that makes it easy to ensure you don't miss anything). I found no stuck pixels. Pixel mapping had worked perfectly.

    Brian
     
  16. Kirill

    Kirill Mu-43 Regular

    78
    Feb 13, 2013
    Tallinn, Estonia
    I am glad for you, but in fact the last time I tried it was last year for two E-PL2s and it failed in both cases. I don't think the problem pixels were white, though. To my eyes they are always some shade of colour. Maybe it can detect white ones best. In any case, gives me more peace of mind when the camera has none outstanding when I buy it. I am convinced that if a sensor shows pixels like this from the moment I pick up the camera, it is more likely to deteriorate faster. My logic/knowledge is that all sensors have them BUT such pixels are automatically mapped out at the factory. So if I see them when I buy the cam it means they appeared in the rather brief and inactive period between production and purchase, which is not a good sign. Maybe I am just a perfectionist, of course about ten years ago I started off without checking for this stuff but quickly came to realise how I hate seeing such spots in dark images. And I do notice them sooner rather than later, I can't help it.
     
  17. keko

    keko Mu-43 Rookie

    23
    May 19, 2013
    Spain
    Could you ellaborate on the exact procedure you used in store to determine the number of hot pixels, so I can reproduce it with my cameras and let you know the outcome?

    Without any rigorous testing, my subjective impression is that both my E-P1 and E-M5 have never evidenced a hot pixel in actual pictures, whereas my nikon D700 and Fuji S5 have shown some in equivalent use/situations.
     
  18. Ian.

    Ian. Mu-43 All-Pro

    Mar 13, 2013
    Munich
    Ian
    post withdrawn
     
  19. Kirill

    Kirill Mu-43 Regular

    78
    Feb 13, 2013
    Tallinn, Estonia
    The free softaware I use can be downloaded from here :
    StarZen Technologies, Inc; DataFlex tools and books

    I set my camera as follows prior to taking a shot with the lens cap on:
    M mode, lowest ISO, aperture closed all the way down just to be on the safe side, shutter speed at 1/30, focus switched to manual because otherwise some cameras will hunt and hunt.

    To determine if there are any hot pixels that will be visible to my eyes, I set the left-hand parameter in the software to 30 instead of the default 60. The right-hand parameter is fine the way it is. Then I load the JPEG file (you can test another format as well) and click Test and hold my breath :) If no pixels are listed at the threshold of 30, the sensor is good enough for me. If you re-run the test with that value set to 20 and still no pixels are listed, then your sensor is truly excellent :)