Pages: [1]   Go Down

Author Topic: Nikon VR issue  (Read 4381 times)

Peter_DL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Nikon VR issue
« on: January 03, 2015, 12:41:22 pm »


Based on a recent thread on the Minimum shutter speed
I did some related tests, with some surprising results (at least for me) as far as the VR feature is concerned.

Nikon D5300 (APS-C, 24 Mpx), AF-S 3.5-5.6/16-85 G ED travel zoom,
shot at 16 mm, handheld, AF-S (non Life View), Exposure Delay Mode off, VR off or on

With VR off:  the results are essentially in line with the below quote. Proceeding from slow to high shutter speeds, from 1/60 on there is a good / increasing chance to get sharp image details (ca. 50:50 at 1/60). There is some dependency on the (my) human body’s motion per shot. Based on a large number of shots, the results suggest that shutter speeds of 1/250 or shorter are needed for reliable good results - within the limitations of the lens of course.

Situation changes with VR on:  there is a relatively higher chance to get sharp results at slower shutter speeds of 1/25 to 1/60. The 1/60 are now a quite safe choice. But – big but – further proceeding to higher shutter speeds there is a "bad shutter speed" of 1/200 where the image details almost always go south (see attached image, from left to right: 1/60, 1/200, 1/320). At 1/200 it more looks like a double image than just blur. Shutter speeds shorter than 1/250 are needed to overcome this effect again.

The VR-on behavior @ 1/200 seems to be a technical flaw here,
not sure though if and what it means "in general" for other devices / units.

Any similar experiences around here ?

Peter

--

The 1/FL rules was never good for critical work, IMHO. Ansel Adams, in The Camera (from the second series), on page 116, talks about handholding a normal lens camera -- he doesn't specify which one -- and says that he needs 1/250 second for critical sharpness. On a 35mm camera, that's 1/(5*FL).
I've done testing to find out how well I can handhold various cameras and lenses, and I advise you to do the same…

Here are some test results that show that 1/500 is pretty good with a 58mm lens on a D800E, and 1/250 isn't quite as good..

--

Attached image, VR on,
from left to right: 1/60, 1/200, 1/320
Logged

Some Guy

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 729
Re: Nikon VR issue
« Reply #1 on: January 03, 2015, 02:33:12 pm »

One other possibility is to turn off the AF and just leave the VR on.  Just focus "manually in Live View" and see if that matters for the VR portion.  Could be the AF moving a bit (hysteresis? or slop where it hunts a bit.) and throwing the focus off.  With the combination of AF and VR both on, it's harder to nail down which is the worse influence.

My D800E doesn't always get the AF nailed consistently shot-to-shot.  Even the FoCal software shows a lot of its hunting diamonds around some areas in the focus tuning curve where is constantly refocuses and tries again.  I can fire a barrage of 6 shots in a row off a tripod on a fixed object, and one will always be sharper than the rest.  Maddening at times.

SG
Logged

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: Nikon VR issue
« Reply #2 on: January 03, 2015, 04:40:39 pm »

Historical concensus has been that VR has potential negative effects with shutter speeds above 1/500th.  What you say you are experiencing is a bad spot around 1/200th.  My eyes aren't good enough to say vibration blur or focus blur.  I don't handhold enough to really say, I've esperienced anything like this.  Mostly if I'm handholding, it is indoors with a flash bracket attached and off camera flash.

It could be a camera thing...every camera has specific shutter speeds where they generate increased self induced vibration, but that is usually at much slower shutter speeds.  It could also be a phenomena specific to this lens model.  I would search out others with this lens for anecdotal experience.

I will say that 16mm, it would not even dawn on me to use VR at shutter speeds approaching 1/200th.

Based on a recent thread on the Minimum shutter speed
I did some related tests, with some surprising results (at least for me) as far as the VR feature is concerned.

Nikon D5300 (APS-C, 24 Mpx), AF-S 3.5-5.6/16-85 G ED travel zoom,
shot at 16 mm, handheld, AF-S (non Life View), Exposure Delay Mode off, VR off or on

With VR off:  the results are essentially in line with the below quote. Proceeding from slow to high shutter speeds, from 1/60 on there is a good / increasing chance to get sharp image details (ca. 50:50 at 1/60). There is some dependency on the (my) human body’s motion per shot. Based on a large number of shots, the results suggest that shutter speeds of 1/250 or shorter are needed for reliable good results - within the limitations of the lens of course.

Situation changes with VR on:  there is a relatively higher chance to get sharp results at slower shutter speeds of 1/25 to 1/60. The 1/60 are now a quite safe choice. But – big but – further proceeding to higher shutter speeds there is a "bad shutter speed" of 1/200 where the image details almost always go south (see attached image, from left to right: 1/60, 1/200, 1/320). At 1/200 it more looks like a double image than just blur. Shutter speeds shorter than 1/250 are needed to overcome this effect again.

The VR-on behavior @ 1/200 seems to be a technical flaw here,
not sure though if and what it means "in general" for other devices / units.

Any similar experiences around here ?

Peter

--

--

Attached image, VR on,
from left to right: 1/60, 1/200, 1/320
Logged

Peter_DL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: Nikon VR issue
« Reply #3 on: January 04, 2015, 09:32:59 pm »

 
Many thanks for the comments,
yes, it seems as if the VR here would amplify a specific camera vibration at 1/200th rather than defeating it,
however, I also see that the issue as described is not particularly widespread, and maybe an upgrade of the camera will resolve it
(waiting for the rumored D7200). In the meantime I'll check some other lenses.

Peter

--
Logged

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: Nikon VR issue
« Reply #4 on: January 05, 2015, 02:26:26 pm »

 
Many thanks for the comments,
yes, it seems as if the VR here would amplify a specific camera vibration at 1/200th rather than defeating it,
however, I also see that the issue as described is not particularly widespread, and maybe an upgrade of the camera will resolve it
(waiting for the rumored D7200). In the meantime I'll check some other lenses.

Peter

As an aside, Peter, you might check on the 18-200mm DX G VRII lens.  It is the only DX lens I ever bought.  For what you can get it for it is a real gem.  Yes, it has barrel distortion at the wide end and pincushion at the long end, but as a general travel and walking around lens, it is quite remarkable.  About the only real issue is zoom creep when pointed downward.

D7100, f/22 for stars on the lights, 18-200mm at 35mm.
Logged

Peter_DL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: Nikon VR issue
« Reply #5 on: January 06, 2015, 06:26:44 am »

 
Thanks for the pointer, - actually I have a couple of further DX lenses to resort to,
however, the 16-85 was my favorite for walkaround photography. Testing my further lenses:

10-24 / 3.5-4.5 G ED:  no VR, no related issue
18-105 / 3.5-5-6 G ED:  same issue as described above
85 Mirco / 3.5 G ED:  no issue, VR works fine
55-300 / 4.5-5.6:  no issue, VR works fine

The 16-85 and 18-105 are from the same generation, introduced 2008,
and maybe that’s the root course here vs. the much newer D5300 24 Mpx camera.

Anyway, it seems that Nikon has started their expected spring announcements of new cameras and lenses today (D5500, 55-200 II).
I’ll wait what else comes out.

Peter

--
Logged

MarkL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 475
Re: Nikon VR issue
« Reply #6 on: January 06, 2015, 08:22:14 am »

My D800E doesn't always get the AF nailed consistently shot-to-shot.  Even the FoCal software shows a lot of its hunting diamonds around some areas in the focus tuning curve where is constantly refocuses and tries again.  I can fire a barrage of 6 shots in a row off a tripod on a fixed object, and one will always be sharper than the rest.  Maddening at times.

This drives me nuts too. I notice it most when I use my X100 as cdaf seems to absolutely nail it or totally miss and the latter is obvious at shooting time. I've even considered buying an LCD loupe and usig the camera in live view when I don't need af speed.
Logged

Peter_DL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 544
Re: Nikon VR issue
« Reply #7 on: January 11, 2015, 06:00:20 am »

  
Just for the purpose of documentation:

a very similar finding was reported here,

as well as in particular here:

quote >>I came across one unpleasant thing regarding D5300, that would be good to highlight. In certain shutter speeds the photos are blurry. … When shutter speed is set to 1/200 s, we can observe a significant "ghosting effect" (doubling of an image), …<<

Indeed there seems to be a certain "bad shutter speed" of 1/200 in conjunction with VR.

Peter


--
Late Supplementary:

In the meantime I reproduced the above shown blurring / ghosting with the new Nikon D5500 + 16-85 lens (VR-on)
at the critical (bad) shutter speed of 1/200.

The adverse effect can be reduced by using the Exposure Delay mode,
thus reducing the influence of the mirror slap, but even with Exposure Delay mode it is not a perfectly reliable solution on a statistical basis of images.

Further, the adverse effect is less pronounced with the newer 18-140 lens,
which supports the assumption that the old VR module of the 16-85 or 18-105 lens plays a role here.

The most significant improvement is however when changing to the D7200,
even with 16-85 lens, be it due to the higher weight of the camera damping the vibration, or the different mirror/shutter unit compared to the D5300/D5500.


« Last Edit: July 16, 2015, 06:00:38 pm by Peter_DL »
Logged
Pages: [1]   Go Up