Nikon D5 - first impressions

[Ilkka Nissilä]

I just came to think about why my auto AF fine tune results are so consistent. I used daylight for the target and the background is dark gray seamless paper which is very homogeneous and reflects little light. I've always wondered how much of an influence light reflected from the background can have on AF testing and believe that it can have some influence if it is not controlled. Some users report great variation in their auto AF fine tune results and I believe the test conditions should be carefully controlled after which it works very well.

I have reprogram my shooting with the D5 ... I shot a concert this afternoon and I did my characteristic overshooting to compensate for AF variations. But now they're all in focus so overshooting was a waste of time, effort and card space. I love it.

However I also did some street shots with the subjects in the shade and background lit by the sun and the matrix meter underexposed the subjects quite badly. I'm not saying the D810 would not have a similar thing but with that camera the recovery zone is deep (at low ISO). The D810's successor (presumably with the Multi-CAM 20k) will be some camera. However for me the D5 saves a lot of time in post because of its consistent auto white balance across a broader range of lighting spectra and the consistent autofocus (I don't have to shoot excess frames) and the smaller files. It is a pity Nikon couldn't have the D810's base ISO signal quality in a fast camera; it slightly reduces the value of the camera to me. I guess I will be carrying the D810 for shoots in sunlight for now. It really has phenomenal image quality at ISO 64-1000. I was shooting catwalk shows a month ago in bright sunlight and the D810 with 200/2II and 300/4 PF and although I was shooting into the light there was just no issue with the backlit subjects' faces. However with the 200/2 some of the shots of the approaching models were not in focus (linear points have a hard time on finding detail in the skin of young female models. I am looking forward to seeing how the D5 handles this situation. Certainly its ability to focus in low light has been a revelation.

By the way I find the 300/4 PF to be one of the best autofocusers of any lens I have used. It's super consistent.

[Frank Fremerey]

Thank you Ilkka. As far as I have comparable experience to yours, I can absolutely copy your statemets and underline every word.

We urgently need a fifth generation D8XX!

[MFloyd]

I had excellent experiences with highly backlit subjects on the move: exposure; dynamics; AF:

I'm still waiting for an example where the (proven) lower dynamics of the D5 sensor translates in lesser pictures; may be, the sensor gives you less exposure latitude; but that's probably all, given that the total EV latitude of print rarely exceeds 7; and screens 9, the latter provided that you set your black point at very low levels.

So I couldn't entirely agree with Ilkka's third statement, although I had some cases where subjects were underexposed and had to turn on spot metering.

Further having high dynamic range in both high and low ISO situations seems to be a matter of choices to be made; Nikon appears to have taken the high ISO route.

[tommiejeep]

MF, very nice.  I'm not looking at a D5 but I still use the D3S for fast action sports in very changeable light.  Needing the speeds and associated rise in ISO(rarely get close to Base ISO)  I'd rather have the camera skewed towards high ISO.   What I look for is very accurate Auto WB and Auto ISO.  Never know which way a play is going to go on the field and what the light will be.

Tom

[Frank Fremerey]

Auto WB is perfect or next to perfect in both D5 and D500 ... The D500 is better than the D3 in every single aspect.

The D5 adds super Hi ISO, slightly better AF and slightly higher speed to that.

Using D3 and D5 in parallel, I must confess the D5 ergonomics is better.

Alas. I like the gripped D500 used in parallel with the D5 better. See above. Landscape mode D5.
Portrait mode D500. At least for my hands.

[Ilkka Nissilä]

There are a number of approaches to mapping the information from a low ISO high dynamic range shot into one that can be printed or displayed on the screen. This can be done by dodging and burning, or using masks to make local adjustments to the image. In my case I typically leave the subject slighly underexposed in the original capture in order to preserve background highlights. I then lift the midtones towards higher values using curves and compress the highlight tones to lower contrast and increase the contrast on the shadow side. This brings out noise if there is any in the shadows, and is cleaner to do with a D810 than D5 (at low ISO). The effect that I'm looking for (reduced highlight contrast and increased shadow contrast) is intentional and spot metering and increasing the exposure for the main subject would not produce the same result.

When photographing e.g. people in a sailing boat, I find extreme contrasts quite common and yet I must deal with the situation. A fill flash can be helpful obviously, when the sails are white and suitable for bouncing, but sometimes the geometry of the situation is against bouncing and one can solve the situation by taking advantage of high dynamic range (e.g. by dodging a face). The D810's sensor is immensely helpful in this kind of portrait in a sailing boat in daylight shots. Success often requires a combination of precise exposure and careful post-processing.

With the D5 I'll be able to take advantage of radio-controlled flash and position it using a clamp to a suitable location for bouncing. I look forward to this actually, though  I'd rather use something less expensive. (Yes, I'm aware of Yongnuo etc. but I'd rather not have so many batteries and components to manage.)

[MFloyd]

Dear Ilkka, thank you for sharing your thoughts & experience. Indeed, you have a very valid point.  I will have to make it out for myself, as, for the time being, I have no pictures-of-sailing-boats-with-people; but this will be the case within a fortnight, as I was asked to photograph a sailing competition on nearly close contact.  In the meantime, a picture taken with my former D4s.  We'll see how the D5 will behave in similar circumstances i.e low ISO (100) 

[Frank Fremerey]

Ilkka. I love your insightful posts full of practical knowledge and consideration. I am always fascinated how many approaches there are to get to a certain result. Thank you!

[MFloyd]

Senscore just published their D5 sensor results http://www.senscore.org

This are the conclusions: "Nikon's latest full frame low-light sensor cannot match the previous generation's excellent noise performance, but in every other aspect, it manages to improve on its predecessor. However, the moderate improvements are somewhat disappointing for a new generation flagship camera. Obviously, a camera is much more than just sensor performance, but the fact that this is the best sensor Nikon can offer 4 years after the launch of the D4 is worrying all the same. Still, the D5 is a great camera and the new #1 in the database."

Also here http://nikonrumors.com/2016/07/08/senscore-posted-their-nikon-d5-test-results.aspx/

It is also noteworthy that the D5 "beats" the Canon 1 DX Mk II in every domain, even with regard to noise, where the D5 is the weakest. From my perspective - compared to the D4s - I observed the improvement in color and tonal range.

[Ilkka Nissilä]

I'm not sure what the "noise" column means at senscore.org, could it be the read noise and exclude photon shot noise? That would make the rest of the data reasonable. All the color sensitivity, tonal range are functions of the SNR. 

I've noticed (as others have before) that the high ISO noise reduction (turned on in camera, activates the Noise Reduction feature in Capture NX-D when it is used as the raw converter), creates quite bad looking blotches of blur at very high ISO (e.g. 8000-12800 are the images that I'm looking at) and I by far prefer High ISO NR OFF. I'm pretty shocked at how inconsistently the algorithm handles detail. I haven't used such ridiculous ISO settings in previous cameras so I'm not sure if this is a new bug or something that always was there but I didn't notice since I was working with lower ISO settings. I think the problem is so severe that I feel the need to contact Nikon about it. It just can't be right. Anyway, without NR, processed in NX-D, the files look good, and at low to medium high ISOs (up to, say, 3200) I haven't noticed offensive artifacts but I haven't looked carefully where these start to occur. I think I'll just turn high ISO NR off for now, and consider trying other software for noise reduction (e.g. I have a license of DXO Optics Pro 9, but I need to update to get it to support the D5 I am sure). Other options include the Google Nik Collection which I have and I could try. Or just leave them without noise reduction, the D5 high ISO noise seems well behaved (as long as the algorithm is turned off!).

I am happy about the camera's very rapid response time. 12 fps is just totally ridiculous (if I turn it on, I seem to end up with a minimum of 3 frames most of the time even when I try to just capture one).    I single shot mode (which is what I'm more used to), the camera still feels faster and the AF is very sure footed. The build and ergonomics are a perfect fit for my hands and it doesn't seem that I want to pick up any of my other cameras for a while.

[MFloyd]

I don't use Nikon postproduction software anymore; everything is processed through Lr/Ps. Every "in-body" controllable image processing is turned off, except long-exposure NR. And I'm wondering whether Adobe will further fine tune the D5 processing algorithms, because there seems still to be glitches in some exposure situations.

[Ilkka Nissilä]

If you report the issues to Adobe perhaps they will fix them. I use ACR for D810 and D750 images and mostly Nx-D for D5.

[MFloyd]

If you report the issues to Adobe perhaps they will fix them. I use ACR for D810 and D750 images and mostly Nx-D for D5.

I haven't found a systematic pattern for the time being. Question: why are you using NX-D for the D5, and ACR for your other bodies ?

[Ilkka Nissilä]

I haven't found a systematic pattern for the time being. Question: why are you using NX-D for the D5, and ACR for your other bodies ?

The NX-D process is too slow for D810 files on my computer to use on a routine basis but for D5 files it is ok. I like it that the camera settings are followed as initial settings for the raw conversion, so I don't have to do those routine steps to get started. In the past I found I got better results using Nikon software for high ISO (6400) images from D3/D700 sensors than using ACR so I kind of figured the D5 would be similar since the sensor is of the same lineage. I haven't done a systematic comparison of high ISO images between ACR and NX-D yet. I should do that.

For the D810 I normally just use ACR or LR and only use NX-D for individual images in specific cases where the image benefits from it. I remember one instance where the LoCA correction in NX-D saved a crucial image of bride and groom exiting the church; activating this feature meant waiting for 1 minute for the processing to complete (!) but the outcome was perfect. In general I am hoping Nikon to get their software up to speed (NX2 in its last versions was several times faster than NX-D on my computer) so it becomes a more viable option for those of us who want to use it.

Adobe sometimes neglects to support some products, e.g., there is no profile for the VR 200/2 II (the last time I looked) so if I wanted to correct the vignetting wide open I have to do it manually or use other software. I don't normally mind the slight vignetting but if I have to do an asymmetrical crop of the image, then the vignetting can become asymmetric and that's annoying̣.

[MFloyd]

Thank you Ilkka for your answer. Your 200mm f/2 VRII lens seems to be taken into account since version 6.3 of ACR https://helpx.adobe.com/fr/x-productkb/multi/lens-profile-support.html#Nikon