Friday, 27 December 2013

Focus Shifters

The 50mm f/1.8G was Nikon's update to their cost effective 'D' and initial reviews where favourable. However, after using this lens for a number of years now I was still having issues with misfocus at wider apertures on my Nikon bodies. This behaviour was consistent across 2 other 50mm f/1.8Gs too but recently someone pointed out an optical phenomenon called focus shift that may be a potential cause...

The 50mm 1.8D had been in my bag for a little while and it was a great cheap lens - it was sharp and lightweight although it annoyed me that the focus ring would spin when the body's motor focused the lens. The 1.8D also beat it's newer brother in a number of sharpness performance areas as seen from the Photozone reviews of both lens on a D3x: this tied up with my own experiences with only the wide open performance of the 1.8G being sharper along with better edge performance and colours being more pleasing.'s 50mm f/1.8D vs f/1.8G MTF charts

The Photozone review of the 1.8G noted more barrel distortion but more importantly "some focus shift when stopping down (residual spherical aberration)" which wasn't reported on the 1.8D. At first glance, this statement doesn't really jump at you but this is important as it essentially tells us there are optical attributes (uncorrected spherical aberrations) that may give the camera's focus system misinformation when the lens is shot at an aperture that isn't f/1.8.

The old rules that we learned early was that lenses were sharper when stopped down a little, even if it was only a partial stop (this is very true on my 85mm f/1.4D - even a 1/3 stop shows a visible improvement although this 85mm is a different class of lens) so I would be shooting my 50mm f/1.8 lenses between f/2 and f/2.8 in an attempt to gain some additional sharpness. The focus shift of the lens has meant that the focus has been compromised - as I tend to shoot f/2 .. f/2.8 (the exact apertures reported to show the focus shift problem the most!) I was suffering the potential results of focus shift. When the lens is stopped down further, the DOF covers/hides the focus shift. Indeed, my initial results with the lens wide open were very good but I wanted to bleed just a little more sharpness. For the 50mm f/1.8D, f/2 was clearly better than wide open but this is being aware of your lenses characteristics.

click for 100% crop/screengrab, D800 f/1.8G at f/1.8

To examine and prove to yourself if your lens has a focus shift issue is simple: disable autofocus on the body and lens, on a tripod manually focus the lens (via liveview) at full aperture, shoot and confirm focus at full aperture is good - without changing anything else, stop the lens down to f/2 (or f/2.8) and reshoot. If you notice that the f/2 image has it's point of focus moved, we've got focus shift.

Therefore, the advice for shooting a lens that has focus shift? Shoot wide open or shoot at an aperture that would mask the focus shift (for my typical shooting distances, that seems to be f/4 and down). Given that the 50mm f/1.8G appears to have been redesigned to give better wide open performance than it's older brother, I have my fix - simply ignore the old rule (sharper when stopped down a little) and I'll just shoot wide open and not worry since the 1.8G is already sharper than the 1.8D in this regard.


www.Siamese-Dream.Com said...

Thanks for posting this. This is really the first time I have heard of focus shift.

Since I am using a D7000 (a DX sensor) and since the DOF of a DX sensor is larger than the DOF field of your full-frame D800, I wonder if DX cameras are LESS prone to focus shift? (Or at least, when there IS focus shift, the effects are less noticeable due to the increased DOF of a DX size sensor.)

Also, for the test that you suggest, is there a reason to use live view instead of viewfinder view? (aside from the fact that it is often easier to manually focus in liveview than it is in viewfinder view?)

Thanks in advance.

Ray said...

It's difficult to say FX vs DX dof covering up for the focus shift but it was very evident between my D800 vs D300. But I used a photoclub's D7000 with their 50 f/1.8G and that exhibit the same issues I saw with my D800 and 50mm.

In terms of using liveview to verify manual focus - the ability to zoom in on LV to such levels makes LV perfect for verifying focus. The optical viewfinder just can't give you that detail. Of course the subject needs to be static and your tripod support needs to be solid as well as remote shutter trip to ensure no shake/bumps to your focus

Post a Comment