Stephan’s Blog

Stock Photography

EOS 5D Mark II: The ‘Black Dot’ Phenomenon

with 15 comments

My noise behavior photos have stirred quite some interest. I’ve uploaded these photos to my new blog on the day the camera was released to the public.

After pixel peeping, someone has discovered black dots right next to blown-outs. If they show, then they always appear a few pixels to the right of the blown-outs in landscape-oriented photos. The higher the ISO the more pronounced the effect appears to be.

Below is a crop of the ISO 3200 image that I’ve shown in my noise behavior post. You can clearly see these ‘black holes’. Some notes:

  • To the best of my knowledge, I turned off all in-camera processing such as picture style and noise reduction. I wanted to show results as close to the RAW file as possible.
  • I turned off all processing in ACR5.2 (Photoshop CS4 trial version), e.g. no ‘Recovery’ or ‘Blacks’ adjustments, no sharpening or noise reduction, no adjustments of curves or vignetting.
  • The same black dots appear in the results I get from the version of DPP that comes along with the EOS 5D Mark II.
  • To post it on my blog, I converted the results to JPG with the least amount of compression (setting 12 in Photoshop).
  • The crop below has been blown up by a factor of about 2.5 to clearly show the phenomenon.

Black dots in a night shot taken at ISO 3200

So what’s the problem? I don’t know but I hope it can be fixed in firmware. The sensor is supposedly identical to the one used in the Canon EOS 1Ds Mark III which doesn’t appear to show this behavior. Although the optics in front of the sensor has been changed to reduce the micro lens gap, I can’t imagine this is causing this phenomenon.

Another indicator that it is not a sensor issue is a shot taken at ISO 50. The same areas that show black dots in the ISO 3200 shot appear more like some sort of flare. I wonder if these flares are caused by the lens optics or by the sensor optics.


Same area at ISO 50

I’d like to contact a Canon representative who’s close to optics or firmware development, but don’t know who to get in touch with. So if you can give me a name, please post a comment. I will not publish it to protect the representative’s email address.

I should note that it is not only my camera that shows this problem. There is already a variety of examples shown on various other sites.


Written by glubsch

December 7, 2008 at 10:06 am

15 Responses

Subscribe to comments with RSS.

  1. Steve.

    Thanks for taking the time to post this. Just wanted to add my voice to yours on this issue.

    It was your original cityscape shot that inspired me to drag the 400mm f/5.6 out to the -10 degree celcius toronto waterfront to run the same tests on our skyline.

    Shooting both JPG and RAW, running from ISO50 to ISO6400 I noted that the higher the ISO, the more pronounced it was… ISO50 it appeared to be a purplish type chromatic issue while at 100 it was faintly noticable, becoming quite pronounced by 400. Certainly not as severe as yours but it’s there.

    Full set posted here:

    Interestingly, its evident on one ISO3200 shot but a following ISO3200 with high ISO noise redux set to ON it was not evident.
    Need to run some more tests with a notebook in hand so I know exactly what I’m shooting when.

    JPG’s have been posted full res, untouched direct from the DCIM folder with links to download the full res shot. With the link above hover your mouse over the skyline shots which will identify what ISO they are.


    December 7, 2008 at 10:22 am

    • Hey Chris, thanks for your comment and the posting of your photos. And, yes, I can clearly see the dots in your photos as well. It appears they are only prevalent if the lightsource is small. The larger blown out areas don’t show this.


      December 7, 2008 at 3:30 pm

  2. Hi Stephan. Thanks for taking the trouble to post this blog entry. My camera has the same issue.

    Is it my imagination, or are there very faint grey dots in the ISO 50 image, in the position of the obvious black dots seen in the high ISO image?


    Mike Louw

    December 7, 2008 at 11:17 am

    • Helle Mike, I’m not quite sure I see it the same way. To me they just appear to be some tiny flares.


      December 7, 2008 at 3:32 pm

  3. Have you considered sending your findings to Rob Galbraith. He was the one who found and analysed the focus problems with the 1DsMkIII. I’m not saying he will want to follow this one up but getting a notice posted on his blog would increase the exposure of this problem in a way that Canon would be more inclined to take notice of.


    December 7, 2008 at 7:56 pm

    • I’ll give it a try. Good suggestion. I already sent him an email.


      December 7, 2008 at 8:06 pm

  4. Hmmm, the bits of flare that these black dots seems to cover at higher ISO settings look a lot like the dreaded “Microlens artifacts” that show up in astrophotos with some high sensitivity Kodak microlensed CCD-based astrocameras. Many astrophotographer will thus not use these microlensed cameras. See for example:
    I am not referring to the blooming, the spike from the bottom, but rather the spike at the top, caused by certain types of microlenses. Recall that Canon is employing new microlenses in this 5DM2 CMOS sensor.


    December 7, 2008 at 10:29 pm

  5. […] EOS 5D Mark II: The ‘Black Dot’ Phenomenon My noise behavior photos have stirred quite some interest. I’ve uploaded these photos to my new blog on the day […] […]

    Top Posts «

    December 11, 2008 at 4:31 pm

  6. […] [photo credit : Stephen Hoerold] […]

  7. […] [photo credit : Stephan Hoerold] […]

  8. Hi. The fact that these black dots get worse with higher ISO indicates to me that it’s a problem with saturation of the analog amplifier that sees the charge from the photodiodes first. Higher ISO, is, I think, cranking up the gain of this amplifier. Analog amplifiers, when they saturate, like from dealing with the previous blown out pixels, take a while to recover and will give very wrong outputs until they do. So, I think, the analog amplifier reading out the blown-out pixels saturates at maximum output level, but does not recover before seeing the charge from the adjacent pixels as the the pixels’ charges are fed at a fixed rate to the analog amplifier.

    John Mann

    December 15, 2008 at 8:04 pm

  9. It looks like disabling highlight tone priority, lighting optimizer, and noise reduction fixes the black dot problem for now.

    Must be a software thing give that it doesn’t apply in all situations. With that, it sounds like Canon is about to release a firmware update to cover this issue.

    Ariel Bravy

    December 18, 2008 at 1:06 pm

  10. […] [Via Digital Photography Review, image courtesy of Stephan Hörold] […]

  11. Terry

    January 8, 2009 at 1:42 am

  12. Yes, the 1.0.7 firmware appears to have done the trick, at least as far as I can see with a few quick real-world test shots. A relief indeed for cityscape and astrophotographers at least!

    Mike Louw

    January 9, 2009 at 5:08 pm

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: