Nikon D5 Extended Test

The last week in June saw me in Norway doing something a bit different from the norm – photographing eagles all week with a 500mm prime – real hard work!

But I thought the task would be made slightly easier with the Nikon D5 running with generation 2 firmware; that is v1.10

And after a solid week of shooting my verdict is – WOEFUL!

The auto focus is still as predictable and user-friendly as a sack full of weasels, but what I found truly appalling is the image quality at lower ISO values – and by lower I mean sub-3200 ISO!

Shooting in Manual Exposure/Auto ISO is the most efficient way of shooting any action, especially with long glass,  but allowing the Nikon D5 to choose its ISO speed just highlights its single massive drawback – poor low ISO performance.

The shot above is at ISO 250, 1/2000th sec and f8.  The crop shows the simple adjustments done to the shot inside the Lightroom Basic panel, and as you can see there is nothing untoward there.

But just look at that appalling level of noise in the underside of the wings – as I said before – WOEFUL.

In this next image, we see the same eagle shot at the same time with the D5 (left) and a D4S (fitted with a 400 f2.8, right). Both cameras are in d25 AF mode, 1/2000th sec, f8 and ISO 1100:

Both images have had a ‘zeroed’ process applied to them in Lightroom followed by a process version swap to kill the excess contrast added by Lightroom in the background.

Again there is excess noise under the wing together with detail degradation in the D5 shot on the left.

As a final comparison, here is the same moment in time caught on the D5, D4S and 1DXMk2. All 3 images have a flat, neutralised process in Lightroom with no added output sharpening:

You do really have to view these images at their full size.

On the whole, I have to say that the 1DXMk2 is the worst image in terms of IQ – both Nikons have it beaten to death – which I must admit surprises me considering the ideal lighting conditions.

But as for the two Nikon shots the D4S still produced the slightly better IQ, lens differences aside, there is still slightly more noise in the D5 shot.

The other error in the D5 shot is due to the sack of angry weasels – the auto focus – the shot is not sharp.  But funnily enough the previous frame was:

As ever, the Nikon AF tends to bounce around a little bit. Even though the D5 has the new Multicam 20K system there is still the same problem of subtle focus bounce that I personally try and negate by shooting at f8 – hoping that the extended DoF will mitigate its visual effect.

But it doesn’t always work, and the D5 will still sometimes drop focus completely on the ‘pick shot’ when the eagle hits the water.

Over the course of the week, I tried pretty much every conceivable permutation of Blocked Shot Response/Subject Motion/AF Mode settings that made any sense – and a few that didn’t – and to be honest they were all as bad as each other.  In the end, I settled on BSR/SM settings at default ‘out the box’, and Group AF mode – but that was way less than perfect.

Ole Martins eagles do represent possibly the most testing scenario for any camera auto focus system, but overall I have to say that for this particular job the D5 AF is an epic fail and a retrograde step – the D4/D4S faired much better.

And both Nikon systems get kicked into touch by the Canon system,  but the IQ of the 1DXMk2 lets it down somewhat, especially in comparison to the Nikon D4/D4S.

I’m certain that better IQ can be had from the Canon system if only Canon would give its users a firmware update to record uncompressed raw; something I’ve been banging on about for years.

But this post is about the Nikon D5 image quality at lower ISOs – and in all honesty, it’s CRAP.

Further Nikon D5 AF thoughts:

Typically of Nikon, they bring out another firmware update just after I use the damn thing for a week.  I’ve not tried the new firmware yet but the ‘added AF modes’ of a single row and single column closest point have a certain smack of desperation in my eyes.

I know that OEM AF calibration is done in both the horizontal and vertical planes.  So to give a camera the ability to use its horizontal calculations and to ignore its vertical ones, and vice versa says to me that there is an imbalance between its x and y axis ‘workings out’.  Couple this with forcing the AF to pick the closest point on the subject under that row or column is basically a case of ‘hedging your bets’ even further.

Nikon should have done exactly what Canon did, and simply refine their existing AF system instead of adding a shed-load of these tracking sensors – there are just TOO MANY points resulting in too much information, and any errors between vertical and horizontal are just being amplified.

I cannot find any visual representation of the two new modes, called group-area AF (VL) with 5 points, and group-area AF (HL) with 11 points.  But if they are as described then they will be ignoring the fixed tracking points. If that is indeed the case, and these modes actually give a marked improvement, then the whole system is a waste of time and effort because it is the plethora of fixed tracking points that form the main distinction between the Multicam 3500 FX and 20K systems.

But hey, that’s just my opinion and I’m not really in possession of all the facts yet.

Become a patron from as little as $1 per month, and help me produce more free content. Patrons gain access to a variety of FREE rewards, discounts and bonuses.