After reading this question, it made me wonder: while the auto sensor cleaning is happening, would the human eye be able to detect any movement?

Point 4 of the accepted answer to this question suggests that the vibration is at a frequency of around 35-50khz. (For some cameras.)

Instinct would make me think that this is just too fast for the human eye to detect, however if the amount of movement is enough, I guess this could feasibly be seen as a blur.

I've never tried to trigger it while the mirror is locked up, but my guess would be that it's not possible to manually trigger sensor cleaning while the mirror is locked up, therefore with normal behaviours of a camera, you wouldn't be able to test the theory.


2 Answers 2


It is imperceptible, either as direct movement or blur.

Source: I removed the lens from my Fujifilm X-series mirrorless camera and activated sensor cleaning. The rear LCD says "sensor cleaning", so I know it's not disabled in this case, but even straining my eyes and concentrating, I can't see anything.

  • \$\begingroup\$ I'm not familiar with how Fuji does it. But with Canon the auto cycle done when the camera is powered on/off is done with the shutter curtains remaining closed so obviously nothing can be seen. On the other hand, if one goes into the menu and tells the camera to "Clean now" the shutter opens (but the mirror stays down) and closes twice during the cycle which is also longer. Looking around the mirror and using a bright flashlight I can't see any movement of the IR filter. The sensor itself does not move during auto sensor cleaning. \$\endgroup\$
    – Michael C
    Commented Jun 3, 2016 at 5:01
  • \$\begingroup\$ The shutter is open. (This is the normal state, as the main sensor feeds the EVF.) \$\endgroup\$
    – mattdm
    Commented Jun 3, 2016 at 11:09

That's much too fast (1-2 orders of magnitude) to see as movement, though you're right it could appear as a blur.

I did an experiment as a kid using mirrors attached to old speakers and bouncing a laser pointer off them to produce Lissajous figures. The speakers were run at audio frequencies (probably middle C and an octave or 2 higher) and the figures appeared static rather than as a dot tracing them out (i.e. they appeared as a blur). I don't suggest doing this with a camera sensor - you'd be more likely to damage the sensor than see anything interesting. The direction of movement is wrong and the amplitude too small.

  • \$\begingroup\$ though the second para of your answer can be considered not related to this question, it is pretty interesting. Could you show me it with a diagram? \$\endgroup\$
    – user152435
    Commented Jun 2, 2016 at 18:33
  • 2
    \$\begingroup\$ Worth pointing out: bouncing a laser off most camera sensors will damage them. \$\endgroup\$
    – Blrfl
    Commented Jun 2, 2016 at 18:45
  • 1
    \$\begingroup\$ @user152435 my 2nd para was meant to back up my assertion that even at large amplitude and a few hundred Hz all you see is a blur. Rather than my recollection of something I did 20+ years ago, I suggest you Google "lissajous mirror speaker" there are several good links with instructions and video. \$\endgroup\$
    – Chris H
    Commented Jun 2, 2016 at 19:11
  • 1
    \$\begingroup\$ @Birfl there wouldn't be any point bouncing a laser off the sensor as the movement is presumably planar. However I have access to enough kit to do this with a weak enough laser, so I may be able to check. But note added \$\endgroup\$
    – Chris H
    Commented Jun 2, 2016 at 19:12

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.