Jump to content
stenrasmussen

Front focus in incandescent light

Recommended Posts

After reading about the issue on various sites I decided to investigate the alledged issue, and sure...it happens!

I tried at least hundred shots with all sorts of AF-S, AF-C, focus/release priority, ADL on/off, different WB, etc. There is a consistent front focus (camera focuses proximal to the target). I haven't got to the stage of quantifying the error but will of course do so.

The issue is very much real and I wonder if anyone else can add to a discussion on this topic?

Edited by stenrasmussen

Share this post


Link to post
Share on other sites

Øivind, you set me straight...it isn't the low light it is the light temp that fools the camera. (Hence the title change).

Here's a more controlled evidence where two different light sources were used. Pretty clear (to me at least) that a little trick needs to be played with AF fine tuning when getting under the incandescent situation!

post-519-0-86175600-1361733524_thumb.jpg

Edited by stenrasmussen
  • Like 1

Share this post


Link to post
Share on other sites

Sten, did you use the true incandescent light bulb? I've heard about this problem and vaguely thought that the AF sensors see things more into the IR region than the image sensors do. The absense of blue portion and the richness of near IR portion in the incandescent light might cause the shift of the entire wavelength range detected by the AF sensors to the longer side, which might cause the focus shift?

The LED light source with the simulated incandescent light color doesn't seem to cause focus shift, which might support my assumption: LED doesn't emit IR, unlike the light bulb.

I also wonder if a Baader UV/IR block filter would mitigate the problem.

Edited by Akira

"The eye is blind if the mind is absent." - Confucius

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

Share this post


Link to post
Share on other sites

I just used a regular, old fashioned light bulb. Your theory about a lack of the shorter wavelengths is certainly interesting. I don't have a Baader UV/IR block filter so cannot answer that.

I will probably mitigate the issue by learning what AF tuning compensation I need to dial in under such conditions.

This is not a lens specific problem as it can be repeated with other AF lenses.

Edited by stenrasmussen

Share this post


Link to post
Share on other sites

I just did a test with my D5100 on the glowing element of the kitchen stove (old-fashioned US non-flat-top with exposed element) while making myself a cup of tea. :) I compared it to one that was not turned on (light source of the kitchen warm fluorescent bulb replacements, with little IR radiation). While not an ideal focus target and with fairly low light levels, I found no difference in focusing accuracy. My D40x IR-720 would typically completely blow out on that target and things closest to it. As I recall from some older threads on DPreview, D5100 was one of the bodies that had some complaints about (front? back) focus in incandescent light. Could alignment of of each individual body's AF sensors have some effect here, a sensor not ideally aligned having more CA?

Edited by otoien

Øivind

Share this post


Link to post
Share on other sites

Øivind, what type of light are you referring to as "kitchen warm fluorescent bulb replacements"? So long as they are light bulbs and not the LEDs, they should emit healthy amount of IR that is influencial enough.

Also, I would think you would need to eliminate the influence of the "glowing element" when you do the test with the "replacements" for more accurate results.

Edited by Akira

"The eye is blind if the mind is absent." - Confucius

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

Share this post


Link to post
Share on other sites

Fluorescent lights do not give off much IR, the reason they are saving energy ("power saver bulbs"). It does record some in the red channel, but much less IR than a typical outside scene. Just to make you happy I tested with an LED light source, without seeing any difference. :) Of course filtering is needed if one completely wants to eliminate IR. The kitchen stove element is a massive IR source and completely overwhelms the background IR and visible light.

Edited by otoien

Øivind

Share this post


Link to post
Share on other sites

Øivind, thanks for further description of the setup.

However, the problem in this particular test is the influence of IR light on the AF sensor, not the image sensor. The image quality in terms of the color balance is not important.

So long as the stove is on, and there is lots of (near) IR as you say, then the difference of the lighting (LED or power saving one) may make no difference in terms of the focus. You would have to extinguish the fire completely in order to test the LED or power saving fluorescent light, which I would not require you living in Alaska around this time, though, :)

Edited by Akira

"The eye is blind if the mind is absent." - Confucius

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

Share this post


Link to post
Share on other sites

We do have electricity in Alaska, and the LED light was tested with the electric kitchen cooking stove off. :D

That is the control experiment.

Then, when the heating element is on and emitting IR overwhelming all other sources, if the AF sensor was sensitive to IR light, and had a shift (CA) because of the longer wavelength it should have been seen as a shift in AF when focusing on the element, as the sensor would focus partly according to IR and not the visible light which forms the exposure.

Another possibility would be that it is mostly sensitive to the blue end of the spectrum, and in incandescent light it would receive very little blue and be inaccurate. In a third experiment I eliminated that possibility by focusing on red fabric filing the view which also resulted in accurate AF.

Edited by otoien

Øivind

Share this post


Link to post
Share on other sites

Øivind, thanks for the detailed description of your experiment.

Another reason for my assumption of the deep red/near IR sensitivity of AF sensor is that the AF assist light (or patterns) emitted from the flash is mostly red. On the other hand, the AF assist light on the mirrorless cameras are mostly NOT red: my Nikon 1 J1 had a green LED and my Sony NEX-5R has an orange one. Panasonic G1 and GH2 didn't have AF assist lamps, and, interestingly, the red light on the flashes (dedicated to both 4/3 DSLR and mirrorless m4/3) didn't light when used with G series bodies. I suspect that the image sensors are not as sensitive to red as AF sensors due to the efficient UV/IR cut filter.

I'm not sure if there is any IR cut filter in front of AF sensors, though.

Edited by Akira

"The eye is blind if the mind is absent." - Confucius

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

Share this post


Link to post
Share on other sites

I just used a regular, old fashioned light bulb. Your theory about a lack of the shorter wavelengths is certainly interesting. I don't have a Baader UV/IR block filter so cannot answer that.

YOu could try the opposite approach and use a more easily available IR filter and see where that focuses.

Øivind, what type of light are you referring to as "kitchen warm fluorescent bulb replacements"? So long as they are light bulbs and not the LEDs, they should emit healthy amount of IR that is influencial enough.

Compact flourescesnt bulbs ie the ones that fit into a conventional light fitting are notorious for missing certain bands of the spectrum. I was playing around with slide copying a while back and found that the ones I have were leaving my image completly lacking in red.

Share this post


Link to post
Share on other sites

...and here are the results from testing out the D600. It backfocuses! On an average not as badly as the shift seen with the D800 but enough to "irritate".
I also include an updated compilation for my D800.

Edited by stenrasmussen

Share this post


Link to post
Share on other sites

Does this mean (if true!) that cameras would mis-focus in any kind of warm light - for example, sunrise or sunset light? I haven't seen any evidence of this, and my D800 doesn't focus any worse at sunset than it does in normal daylight. :(


Dave Paterson

Share this post


Link to post
Share on other sites

Sten, thanks for sharing the additional interesting result. One would re-realize how unforgiving a denser image sensor is of tiny bit of focus errors!

Edited by Akira

"The eye is blind if the mind is absent." - Confucius

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

Share this post


Link to post
Share on other sites

What happens if you don't change the camera white balance setting? Is the camera using that to compensate focus due to different wavelengths?

Share this post


Link to post
Share on other sites

Sten, I have the same with my D4. It front focus clearly at warmer color temperatures. Quite annoying, especially with my fast primes. Changing the AF fine tune value when color temperature changes is not very workable.

Share this post


Link to post
Share on other sites

I noticed this behaviour on my D7000: Incandescent lighting causes it to back focus every time. When using the DX 35mm 1.8G at a distance of 3 metres, focus is some 20 cm behind the object. In daylight everything is perfect. I would consider a new camera that focuses better in warm light but it seems that even the very expensive cameras are not perfect in this behaviour.

Share this post


Link to post
Share on other sites

This is claimed to be fixed in the new firmware releases for both the D800 and D4. I didn't have any time to test it but can anyone confirm?

 

Hi, Nick, the problem discussed here is not even mentioned either in the new firmware release or any past announcements from Nikon.

 

Sorry to say, but I don't think this problem can be solved on the user-updatable firmware basis.  The AF inaccuracy problem of my first sample of D7000 could be addressed on the software basis BUT only at the service center.

 

I believe that this affect from the incandescent light bulb is the inherent problem of the AF sensor and not the image sensor.


"The eye is blind if the mind is absent." - Confucius

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

Share this post


Link to post
Share on other sites

As it turns out if I dial in the opposite equivalent AF fine tuning value it works well in incandescent light.

My 85/1.8G requires a -6 under daylight and if I make it a +6 it addresses focus correctly under incand. light.

The same applies to my 50/1.4G and 28/1.8G...more or less. Haven't tested my 80-200/2.8 (dual ring) yet.

Share this post


Link to post
Share on other sites

Akira,

I know it's not mentioned in the changes list from Nikon, but I've seen several posts (at dpreview) claiming this issue has been resolved. Both for the D800 and D4. Even with shots before and after. That's why I asked.

Thread at dpreview:

http://www.dpreview.com/forums/post/51231410

D4 shots:

http://www.dpreview.com/forums/post/51243938

We've seen issued resolved in firmware more often without Nikon mentioning it so it could have been solved.

It is a bummer it does not appear to be fixed acvording Sten. I will test it tonight.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


  • Similar Content

    • By Kyle
      So over a month ago I vowed to "post more" from my mountain trip....one fixed keyboard later and a new CPU, here I am. Enjoy. 
       
       




    • By armando_m
      Young pregnant mother , struggling under water to get a nice shot , all with the intention to keep a nice memory of this period in her life
       
      I was afraid of not getting good images, but things did work out
       
      Here is a SOOC image , after seeing this she repeatedly asked if I could fix it, and I replied "sure", I probably didn't sound to convincing as she kept asking

      Edited / cropped


      A few blurry patches maybe due to the plastic container I used, plus she used lotion to block the sun and it soon created a film which attached to the surface of my container , I had to remember to wipe it underwater every time I used it
    • By armando_m
      click up please

      Cropped to 8x10
       
      Thanks for looking
    • By armando_m
      Crop to 8x10 and cloned out a couple of people on the left side of the frame
      please clickup

      Thanks for looking
×
×
  • Create New...

Important Information

By visiting this website you are agreeing to our Terms of Use, Privacy Policy & Guidelines.