Rolling Shutter, Panning, BIF - Feather Test

The vertical lines of a fence suddenly lean to the right or to the left - what's going on? Everything is fine - it's just part of my little panning experiment. The rolling shutter effect is real, even for the Nikon1 with its fast shutter read-out time of 1/80 sec.

The effect may be worse for other models - illustrated in another thread by nzmacro's funny sparrow photo. Nevertheless I was eager to find out which effects I might expect when I pan small birds in flight at fast shutter speeds.

A bird wasn't available, thus a feather 6.50m away became my target. The orientation of the feather was important, as the rolling shutter effect mainly distorts vertical lines. As arranged, the feather's fine structures displayed vertical and horizontal elements.

Before we study the panning effects, two shots showing the feather without movement of the camera (Nikon V3 + CX 70-300). Selected was manual mode (M), no VR.

V3 - mechanical shutter, 6fps, no VR, 1/4000 sec

V3 - electronic shutter, 10fps, no VR, 1/5000 sec

Anybody willing to bet whether the mechanical or the electronic shutter does better?

Over the next days I'll post photos illustrating what the panning movement does to the fine structures. Often quite horrible results...

Don't tell anybody, but yesterday I wanted to do the same comparison test with the J5. Until the penny dropped - the J5 doesn't have a mechanical shutter. Ouch!

From my experience, there is no much difference between mechanical shutter and electronic one with VR

Faith Yeung wrote:

From my experience, there is no much difference between mechanical shutter and electronic one with VR

Thanks for the feedback. I used to think the same. And the differences are not so huge. Still, I find it interesting - see my next post.

All images shot in RAW, processed in DxO 11 using Prime, microcontrast 25. The PGNs were imported into Photoshop Elements, clicking on "automatic". No further editing.

#1. V3, mechanical shutter, no VR, 1/2500 sec [fast panning]

#2. V3, mechanical shutter, no VR, 1/4000 sec [slow panning]

With the mechanical shutter, fast panning wipes out most finer vertical structures. Even a slow panning movement and 1/4000 sec doesn't retain much vertical fine detail.

#3. V3, electronic shutter, no VR, 1/5000 sec [fast panning]

Now let's try the electronic shutter. Again, fast panning wipes out vertical fine elements. No surprise so far. But then...

#4. V3, electronic shutter, no VR, 1/16000 sec [fast panning]

A surprise - the vertical fine structure is visible again. Is it the fast shutter speed?

The high ISO 5600 isn't appealing, but under perfect conditions high shutter speeds of 1/8000 sec or more are possible. Further tests are necessary.

******

Next I tried to set the VR on "active", which had always been my preferred BIF setting. We start with the mechanical shutter, just as in the series above.

#5. V3, mechanical shutter, VR active, 1/2500 sec [fast panning]

#6. V3, mechanical shutter, VR active, 1/4000 sec [slow panning]

#7. V3, mechanical shutter, VR active, 1/4000 sec [fast panning]

Apparently the mechanical shutter doesn't profit significantly from the active VR setting. The vertical fine structure is mostly gone. It seems to me, however, that the horizontal elements are clearer than in the shots taken without VR support.

But a selection bias may be in play - from several shots I tended to select the clearest.

#8. V3, electronic shutter, VR active, 1/10000 sec [slow panning]

#9. V3, electronic shutter, VR active, 1/10000 sec [fast panning]

#10. V3, electronic shutter, VR active, 1/16000 sec [fast panning]

I'd say that the electronic shutter cooperates very nicely with the "active" VR.

CX Bob wrote:

It looks to me like the active VR is only 'cooperating' with the very high shutter speed. Other than that, I'm not sure if the 'active' setting is of any benefit in itself. I can't remember what the highest speed of the mechanical shutter is, or if it's not the same with the V1 and V3. I'm also about to drop into bed right now, so I'm not sure about anything. Still, I can't help thinking how different reality is from 'theoretical' things. I mean, if it were possible to do a 'perfect pan', then a 1/8 of a second shutter speed would work fine, wouldn't it? I'm beginning to think that the shutter speed is the most significant thing in most bif applications, and the advantage of the electronic shutter is simply its ability to provide those higher speeds. Keep up the good work! ( :

Thank you, Bob. [And I'd also like to thank Jonikon for his rolling shutter observations in the thread "Grey Wagtail in flight (60fps)". It was the starting point for my test.]

I need to look closer into this matter. As a consequence, I might ignore the mechanical shutter entirely - except for flash photographs.

The electronic shutter is king. How useful are my results? How relevant can tiny structures be? I'd argue they matter quite a bit. Look at the wood on the left side of the feather. In photos where vertical structures are wiped out, the wood looks unnatural.

The V1, V2 and V3 all offer the same maximum shutter speeds: 1/16000 for electronic, 1/4000 for the mechanical shutter. But when the limits are the same, it doesn't have to mean the "active VR" does exactly the same in the many N1 bodies. I'd assume over the years Nikon's engineers went through several different versions.

Whether only the high shutter speeds cooperate well with active VR remains to be seen.

By the way, the J5 has only one VR mode. Maybe it's the former "active VR"?

Other than that, I'm not sure if the 'active' setting is of any benefit in itself.

I believe the video folks prefer to use the "active" VR.

Thank you for your findings.

However, in most situation, I shoot 1/200 with VR and electronic shutter to keep the lowest ISO for "steady" birds in order to maintain fur details and sharpness

Stefan BW wrote:

All images shot in RAW, processed in DxO 11 using Prime, microcontrast 25. The PGNs were imported into Photoshop Elements, clicking on "automatic". No further editing.

#1. V3, mechanical shutter, no VR, 1/2500 sec [fast panning]

#2. V3, mechanical shutter, no VR, 1/4000 sec [slow panning]

With the mechanical shutter, fast panning wipes out most finer vertical structures. Even a slow panning movement and 1/4000 sec doesn't retain much vertical fine detail.

#3. V3, electronic shutter, no VR, 1/5000 sec [fast panning]

Now let's try the electronic shutter. Again, fast panning wipes out vertical fine elements. No surprise so far. But then...

#4. V3, electronic shutter, no VR, 1/16000 sec [fast panning]

A surprise - the vertical fine structure is visible again. Is it the fast shutter speed?

The high ISO 5600 isn't appealing, but under perfect conditions high shutter speeds of 1/8000 sec or more are possible. Further tests are necessary.

******

Next I tried to set the VR on "active", which had always been my preferred BIF setting. We start with the mechanical shutter, just as in the series above.

#5. V3, mechanical shutter, VR active, 1/2500 sec [fast panning]

#6. V3, mechanical shutter, VR active, 1/4000 sec [slow panning]

#7. V3, mechanical shutter, VR active, 1/4000 sec [fast panning]

Apparently the mechanical shutter doesn't profit significantly from the active VR setting. The vertical fine structure is mostly gone. It seems to me, however, that the horizontal elements are clearer than in the shots taken without VR support.

But a selection bias may be in play - from several shots I tended to select the clearest.

#8. V3, electronic shutter, VR active, 1/10000 sec [slow panning]

#9. V3, electronic shutter, VR active, 1/10000 sec [fast panning]

#10. V3, electronic shutter, VR active, 1/16000 sec [fast panning]

I'd say that the electronic shutter cooperates very nicely with the "active" VR.

Faith Yeung wrote:

Thank you for your findings.

However, in most situation, I shoot 1/200 with VR and electronic shutter to keep the lowest ISO for "steady" birds in order to maintain fur details and sharpness

You are clearly in a different situation than me, with my interest in small birds in flight. Fortunately we can both use Nikon's excellent electronic shutter. When I read posts in other mirrorless sections, users often still prefer the mechanical shutter.

Next week I may have more time for experiments. Checking the J5 and V2, probably, to see if they behave in the same way.

If anybody wants to do similar comparisons, feel free to post it in this thread.

CX Bob wrote:

I'm beginning to think that the shutter speed is the most significant thing in most bif applications, and the advantage of the electronic shutter is simply its ability to provide those higher speeds.

I've just checked out (via dpreview's camera feature finder) alternative BIF cameras, to achieve high shutter speeds. If I select a max continuous frame rate of at least 7 and a maximum shutter speed of 1/16000, only three ICL cameras come up: Fujifilm X-E2S, Nikon J3 and Canon EOS-1D. Not much choice.

Edit: Today I've repeated my feather test with the J5. Disappointing... Maybe I'll also try the V2, in spite of its AA filter. Now I regret I haven't bought a S1 when it was offered for little money.

Could it be that the Sony sensor in the J5 doesn't cooperate well with the software, or that it can't handle high shutter speeds or... whatever. I am no technician. The lack of feather detail in my panning test just has got me thinking whether Nikon's engineers simply couldn't solve very real problems arising from the change when they went from Aptina to a Sony sensor.

Maybe that's the trivial explanation why a V4 never appeared.