Latest or all posts or last 15, 30, 90 or 180 days.
Welcome to
In-depth review coverage is by subscription.
Also by Lloyd: and
First-time visitor
Capacities up to 48TB and speeds up to 1527MB/s

Sony A7R II: General Commentary

Get Sony A7R II at B&H Photo.

This is a general commentary speaking to the body of my expanding review of the Sony A7R II. But the comments here reference in particular Sony A7R II: Posterization with a Near Optimal Exposure in Daylight (Portrait at Dana Lake).

The A7 system has some very fine attributes (sure is nice to carry and I love having an EVF and IBIS). Operational controls and haptics need improvement (I've commented upon the issues in my review), but those things might not concern many users.

When light or color or dynamic range is unusual, the file quality can become an issue, and this shrinks the operational envelope. The blue lake is not some test scene; it is a real scene of rare and special beauty (I’ve never seen its equal in Yosemite). Experiencing that incredible range of blue (continuously variable too!) in person is stunning. And so I wish to record the special things I see without worrying that the camera will fail to record it properly. So I am adamant that the 11+7 bit file format needs a Nikon-grade 14-bit sibling and that the RGB histogram needs work.

Were I to reshoot, I’d bracket and I would not polarize, but in this case I had wanted sheen off the face, which is why I polarized it. I knew the blue was intense and had I stopped to consider, I’d have known that this could be problematic for the red channel. But I was focused on a series of pictures (many more besides that one).

But here’s one rub: the Sony A7R II offers a coarse histogram unsuitable for discerning issues in small areas of the frame, and it offers no RGB histogram just for the zoomed-in area, so even after taking the shot it’s not possible to be certain of the exposure for critical areas, like the face or water. This is a flaw as one can see just by using a Nikon D810 which does it right. I have also observed misleading errors in the Sony RGB histogram when comparing to the actual raw data in RawDigger. Such things are a handicap for field use, but Sony could fix it. The only quick and practical solution at present is bracketing such scenes, which is not all that viable for facial expressions. Oh, and continuous shooting drops to 12-bit mode so how does one bracket quickly, except frame by frame? (though maybe bracketing is not continuous mode and so is not affected, not sure). The A7R II has various file quality land mines of this sort, which in my view undermines its merit as a professional tool. And yet is is very, very capable in ways that make it a better tool under some circumstances.

Sony could issue a firmware update for a lossless-compressed 14-bit file format as well as fixing some problematic behaviors, like dropping to 12-bit mode with Bulb mode, and long exposure noise reduction mode making things worse in some cases. And Sony could improve the RGB histogram behavior and quality. But the issues may run deeper, e.g., excessive processing of the raw data aka “cooking”. And the electronics might not be capable of quality sufficient for a true 14-bit format. Still, a high-grade file format would be a good start that would make the Sony system more credible as a professional tool (albeit of no interest to many casual or routine shooters, many of whom are probably viewing the blue lake on a display capable of little more than sRGB, if that and/or (hopeless) iOS which doesn’t even support color profiles).

As I've written before, Sony cameras feel like camera imitations, not the real thing (the Leica Q feels like the real thing, as does the Nikon D810 or Canon 5DS R). The Sony camera gestalt seemed to be deeply cultural and it may take a long time, if ever, to overcome some design patterns*. Which gives NiCanon some breathing room, in theory. But Sony has made some serious progress with the A7R II vs the A7R, and I applaud those advances—keep ’em coming. In particular, the Sony A7R shutter vibration that did not exist (my findings were repudiated by Sony) is now fixed in the A7R II (how does one fix a non-problem?)—to my lasting relief I can shoot without having a great deal of work ruined, as was the case with the A7R.

I like the A7R II and I will be buying one. Now I just wish for Sony to improve it via firmware updates. If Sony wants to listen, they can ask me for my wish list*, and that goes for any vendor out there. But no vendor ever asks. Maybe Sony can be the first? My goal for all brands is to see the offerings improved, which is why I spell out everything I find, as I have for years.

* At least eliminate the Applications menu which just annoys by occupying space.

John W writes:

These points may serve as evidence that this next wish isn’t possible in the A7R II:
'Sony could issue a firmware update for a lossless-compressed 14-bit file format'

It’s starting to smell like Sony is/was compensating for bandwidth deficiencies in the A7R II’s implementation. Having done signal-processing type firmware work myself, tricks like dropping to 12-bit mode and the lossy RAW format, all sound like “cutting corners by putting in a bunch more work” that resource constrained firmware systems sometimes have to do.

Certainly the 11+7 bit format is a legacy inheritance now, which might excuse its continued presence if you squint just right. But dropping to 12-bit mode is a curious limitation, which I doubt is capricious.

DIGLLOYD: Agreed. The drop to 12-bit mode could indicate a quality issue within the sensor/electronics pipeline. Sony claims 14-bit files and 14-bit pipeline, which offers hope. But the lossy format can hide some issues. However, 14-bit lossless-compressed file format might also provide a useful noise dithering in dark tones.

In general, the next quality front is surely dynamic range. A camera that delivers a robust 15 bit dynamic range would be highly desirable (though few if any lenses can deliver that). And technology for bit counters on a pixel could make arbitrary dynamic range is possible (4 bit or 8 bit or whatever bits for overflow on each pixel).

diglloyd Inc. | FTC Disclosure | PRIVACY POLICY | Trademarks | Terms of Use
Contact | About Lloyd Chambers | Consulting | Photo Tours
RSS Feeds | Twitter
Copyright © 2008-2017 diglloyd Inc, all rights reserved.