4
\$\begingroup\$

Nikon says that banding may occur at high ISO sensitivities with autofocus on the D4 or D800:

D800

Noise in the form of horizontal lines may appear in pictures taken with AF-S Zoom Nikkor 24-85 mm f/3.5-4.5G (IF) lenses at ISO sensitivities over 6400; use manual focus or focus lock.

D4

Noise in the form of lines may appear during autofocus at high ISO sensitivities. Use manual focus or focus lock. Lines may also appear at high ISO sensitivities when aperture is adjusted during movie recording or live view photography.

What would explain this behavior?

Improve this question
\$\endgroup\$
4
  • \$\begingroup\$ At a guess only, the frequencies used in USM motors may get into the sensor electronics. High ISO = higher amplifier gains. Basically they are saying that their design is deficient. \$\endgroup\$
    – Russell McMahon
    Commented May 18, 2012 at 17:41
  • \$\begingroup\$ @RussellMcMahon: You're saying that the sensor is being subject to EM interference? Why would the SWM cause interference? \$\endgroup\$
    – bwDraco
    Commented May 18, 2012 at 18:09
  • \$\begingroup\$ it appears from your answer that my surmise was correct. This is poor design of the camera on Nikon's part. The increasing sensitivity to noise with increasing ISO is a fundamental. matter. As the sensor will usually not be acquiring signal while the lens is focusing, except in tracking modes, it sounds like the USM oscillator is just too noisy. Very badly done by Nikon. \$\endgroup\$
    – Russell McMahon
    Commented May 18, 2012 at 18:46
  • 1
    \$\begingroup\$ Since it's a problem with one specific lens, perhaps something about the SWM in that lens is different/problematic? This is one of the first smaller lenses to receive AF-S, perhaps compromises were made that cause this? \$\endgroup\$
    – Dan Wolfgang
    Commented May 18, 2012 at 19:14

2 Answers 2

Reset to default
4
\$\begingroup\$

I have a D4 and can see very very slight banding from 10000/12800, and if under exposed can be seen at 25600 especially at 51200. If over exposed banding is visible slightly on 25600. If post process +1 exposure then banding is visible above 8000/10000 iso. Tested with AF on and OFF and makes no difference, is very random and is mostly seen in the darker areas only especially towards the right side of images. DO NOT POST Process with gain on exposure, far better to over expose and then darken image, you can then take photos upto 51200 without seeing or very little banding. Shame because the d3s at 102400 had very little banding ( well my copy did anyway - just general very light consistent banding that was acceptable) - see imaging resource images of the 2 cameras and you will see what I mean especially at 25600 and 51200 in the darker parts of the manikin woman image with the green leaves towards the right. It may be true that there is some contribution when using AF and USM lens but it is not the main reason, its just there and very random.

Improve this answer
\$\endgroup\$
2
  • \$\begingroup\$ +1 for presenting an actual test. If this is the case, the camera's electronics (and their interaction with the sensor), not necessarily the SWM, could be responsible for this banding. \$\endgroup\$
    – bwDraco
    Commented Sep 26, 2012 at 2:15
  • \$\begingroup\$ Yes that is the case as non-usm and usm lenses, af on or manual still shows. \$\endgroup\$
    – Nicholas
    Commented Sep 28, 2012 at 8:25
5
\$\begingroup\$

This problem is caused by electromagnetic interference generated by the SWM in many Nikon lenses, and possibly by the camera's own electronics as well.

There are cases where Canon lenses with USMs have caused banding. See this dpreview forum thread for an instance of sensor banding caused by a USM lens mounted on a Canon DSLR.

Improve this answer
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.