But What About Images?

Okay, okay, I hear you. But I’m not going to deluge you with lots of test images and results in this blog. I’ll leave the full discussion of what I’m seeing to my review and book. Still, you want something, so here’s a taste. 

Let’s not light my studio (according to my meter it’s running about 4 lux unlit). Let’s stick a great lens on my tripod and lock its tripod collar down (that way I can just attach each body to the lens and nothing else changes). I’ll shoot my usual Lego scene, but within that scene I have a small miniColorChecker card we’ll be looking at (and yes, I almost always shoot the ColorChecker upside down, as the two rows I’m most interested in are at the bottom, and there’s something about having to visually ignore two rows above what I want to look at...).

Next, I’m going to grab my D4 and D5 bodies, set the ISO to 6400 on each, then set in-camera noise reduction to Normal. And now for the kicker: I’m going to put white at the mid-tone (about two and a half stops underexposed). I’ll then take that JPEG and set its white and black points correctly. 

Here’s what we get:

D4: 

d4-6400.jpg

D5:

D5-6400.jpg

If you can’t see that difference, well, I can’t help you. 

Okay, first a bit more explaining to do: the same capture area is active in both images. The D5 has 809 pixels across in that area, the D4 has 706. So, yes, one impact is that we’re going to make the pixels slightly less visible for the D5 at the same output magnification. Not enough to significantly distort what you’re seeing, though.

Still, there’s a lot more going on here than first meets the eye. First, the underlying noise values are lower on the D5, all else equal. As noted earlier, it might be 2/3rds of a stop improvement over the D4 just from that. 

But Nikon has significantly improved the noise reduction and rendering in JPEGs, too. Some of the patches are startlingly better (yellow, white, lime, orange), while all patches are visibly cleaner (even at per pixel level). Edge handling is terrifically improved. As I look at low light D5 JPEG images, I’m struck by how far I can press the camera than the D4 and get something very usable. 

All is not perfect, though. Note the color blocking that’s happening. Technically, red and blue are drifting darker due to the noise and how it’s then handled. 

Still, this JPEG performance certainly would change the way I’d shoot from the D4 to the D5. I might be able to push another stop of ISO and turn down the noise reduction. I can’t quite get a handle on that yet with early testing; it’ll take a lot more testing in different types of light to get a full handle on what’s happening.

What I’m hearing from most of the pros I know already using the D5 seems to agree with what I just wrote: they’re pleased that they can push their camera harder in low light than they did the D4 and still get better results. At least with JPEGs.

So what about raw? Another test for another day, and I still have to wrap my head around all the hidden changes that ACR is doing (one third stop increase in exposure, for one). 

text and images © 2019 Thom Hogan
portions Copyright 1999-2018 Thom Hogan-- All Rights Reserved
Follow us on Twitter: @bythom, hashtags #bythom, #dslrbodies