I purchased a new Samyang 8mm F3.5 UMC Fish-eye CS II today. After I made my first shot with it I immediately noticed a flaw. At f/3.5 the flare of bright objects is shifted towards the bottom (and just a little to the left).

I have quite few lenses and have never seen such a thing, Only my old Revuenon 1.4/50mm does it but in very subtle way. The flare is shifted to the left and it is barely noticeable, in fact I noticed it after purchasing the Samyang, and the effect is negligible (just few pixels difference).

The lens is usable after f/5 and flare seems completely disappear at f/5.6 which effectively makes the lens at best f/5 8mm.

I don't mind flares, as long as they are even around the object, but such flares make photos just ugly.

Is this effect considered normal?

Here are photos made with normal position of camera and rotated. The camera is Pentax K5.

enter image description here enter image description here

  • 1
    \$\begingroup\$ Is it the same over the entire image, or if you use that last image (the spot), but put it in each of the four corners, does it actually spread outward towards the corners? Also, if you put that spot dead center, is it the same? Many if not most lenses have a sweet spot closed down a stop or two.... \$\endgroup\$
    – twalberg
    Oct 16, 2020 at 16:56
  • 1
    \$\begingroup\$ @twalberg, yes, it is the same in every point on image. For example the images number 2 and 3 are from the same shot: the marshall logo being on far left of the image, while the wheat is in the center \$\endgroup\$
    – Qeeet
    Oct 16, 2020 at 17:10
  • 1
    \$\begingroup\$ Could be a decentered element i guess. \$\endgroup\$
    – Arji
    Oct 16, 2020 at 17:29
  • \$\begingroup\$ @twalberg, just tested in the evening on bright spots. Flares actually are not perfectly shifted to one direction on every point of the image. They slightly rotate: top right and bottom left corners - counterclockwise, bottom right and top left corners - clockwise. But the difference between rotations in extreme corners of an image is about 90 degrees \$\endgroup\$
    – Qeeet
    Oct 16, 2020 at 19:45
  • \$\begingroup\$ If they seem more radially-aligned, then that may just be normal unavoidable aberrations in the lens design, but if they're only partially radial, then, as @Arjihad suggested, it could indicate one or more lens elements are out of adjustment. If that's the case, then perhaps warranty terms apply, if you bought it new, or alternatively you could send it in for repair at your own cost. Or, since the aberrations seem to be minimized when stopped down, just avoid shooting wide open when you can.... \$\endgroup\$
    – twalberg
    Oct 16, 2020 at 20:20

1 Answer 1


After returning that lens and buying the new one, I can conclude with confidence that the first lens was definitely defective.

This new lens is not perfect either, but it performs much better. You can see those ghosting flares are going up and a bit to left, but the effect is less noticeable. The focus on both images is the best I could get - at infinity.

Finally, the lens is usable at f/3.5.

The images are cropped. enter image description here

  • \$\begingroup\$ I don't see any difference in the images that can't be attributed to camera settings and focus distance. \$\endgroup\$
    – xiota
    Nov 4, 2020 at 23:00
  • \$\begingroup\$ @xiota, maybe you are right. If so the focus on the first lens had to be set beyond what is marked infinity. But that is impossible without taking apart the lens \$\endgroup\$
    – Qeeet
    Nov 11, 2020 at 15:54

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.