Pros and cons of the new A7r II

<p>Is there any review out there that is both comprehensive and also a critical analysis of the new Sony and its practical performance and IQ? I realize it is just appearing in the stores. It seems on paper like a very high quality device, but there are probably some downsides to squeezing so many pixels into a full frame space. Limited dynamic range? Other? There probably will be limitations on what non Sony wide angles one can use on adapters with it and still retain high quality. Which are the best Sony and E mount Zeiss optics for it, prime or zoom? Thanks for any updates on tests.</p>

<p>This won't ship for a couple of months. It hasn't had a thorough review yet, and you really can't get much from the guys who had it for a few minutes and weren't allowed to keep the images.</p>

<p>I have owned an A7ii since December, and it is everything promised and more. There is every reason to believe the A7Rii will deliver as promised. The killer features are increased resolution, better high ISO perfomance and the electronic shutter, including a silent operation option. The somewhat metallic noise of the A7ii, even using the electronic first shutter, is distracting.</p>

<p>It is true that wide angle lenses for rangefinders, Leica and Zeiss, are blurry in the corners, to the point of looking smeared. This is due to the thick (2 mm) cover glass on the sensor with lenses with a short backfocus (e.g., Summicron, Zeiss Biogon). SLR lenses work much better in the corners, since they are designed to have a long back focus, but they are not nearly as sharp overall and much larger. Zeiss has a much better solution in the Loxia 35/2 Planar (MF) and Sony-Zeiss 35/2.8 AF. The Sony 28/2 is also a very sharp lens. Promised for July delivery, Zeiss has two new AF lenses, the Batis 85/1.8 and Batis 25/2 which resemble the lofty Otus lenses in design and performance.</p>

<p>Longer lenses, like a Summicron 50 mm work very well, but are somewhat softer in the corners than the Loxia 50/2 Planar, which is mated very well with the A7 in other respects. The Sony-Zeiss 55/1.8 is extremely sharp throughout, as is the Sony-Zeiss 35/1.4 Distagon ($$$). The Sony 70-200/4 is half the weight of the Nikon f/2.8 lens, and IMO twice as sharp. Sony and Zeiss are reaching for the moon with their A7 lenses, especially good primes.</p>

<p>I have Novoflex adapters for Leica M and Nikon SLR lenses. The Nikon adapter has aperture control for G lenses too. There are no real drawbacks, except you don't have AF nor EXIF data to share with the camera.</p>

<p>Thanks Andy, and especially Edward for your comments and A7 experience. It would probably not solve all issues with wide angle RF type lens compatibility, but the option of an A7R II with a much thinner sensor cover glass might be interesting. I find it hard to justify an A7RII for 50mm and longer existing RF optics, especially for those of us who already have a small stable of lenses for film Leica, V-C or Cosina-Zeiss Ikon cameras, and other older cameras. I imagine that Sony would not be interested in providing such an option as it would undercut sales of its own A7 compatible short focal length optics. I don't care about having AF for short focal lengths. </p>

<p>So the question of the A7RII for some of us is whether we want to ditch or not many of our current RF optics and invest in new, albeit very fine, new lenses for the A7 series. I wonder if any aftermarket modifications may be available for the sensor and its cover glass (I'm not a specialist and perhaps the two are practically inseparable), but if that is ever an option the question of firmware modification would no doubt also be necessary.</p>

They seem to have struck a fine

balance between pixel count and

sensitivity. That tells me that Sony is

playing the quality game, not the

numbers game. Having said that, I do

not feel that I need it.

 

FWIW I love Sony cameras and I still

have (and use) one NEX-C3 as well as

an A7.

<p>The cover glass will probably be the same in the A7Rii as in the earlier A7 cameras - 2 mm. This is a separate issue from the back lit sensor structure, which serves to increase the pixel size by putting most of the circuitry behind the cell window. The cover glass is protective, and serves as the IR cut filter. This is a significant element in the optical path, and affects the focus point at the edges and corners of the image. Lenses designed for the A7 compensate for this thickness. Changing it would render existing lenses for the A7 obsolete.</p>

<p>The extent effectiveness of this compensation can be seen in the Loxia 35/2. The MTF curves for this lens indicate a fall off in resolution and increase in coma at the extreme edges. However the MTF curves are measured in air, not on the camera itself. Looking at actual images, the 35/2 is sharp from corner to corner. There is some fall off in sharpness, but it is barely noticeable. The same is true for the Loxia 50/2, which is sharper than my Summicron 50/2 v2 (1964) on the A7ii, and on a par with the $7K Summicron 50/2 ASPH in a test by Steve Huff (tested on an A7ii and Leica M240 respectively). In comparison, my Summaron 35/2.8, Zeiss Biogon 35/2.8 and 28/2.8 produce ugly smearing in the corners on an A7ii, but not when used on a Leica M9.</p>

<p>Odd, isn't it, that we quibble about extreme corner sharpness on these cameras, but have accepted smearing for years in Nikon DSLRs. There's a new benchmark, and it's hard looking back. In this new world, we can (and often do) use lenses wide open without sacrificing quality, and consider f/8 the threshhold of diffraction.</p>

<p>Hi Edward,</p>

<p>Are you sure about this statement?</p>

 

<blockquote>

<p>Lenses designed for the A7 compensate for this thickness. Changing it would render existing lenses for the A7 obsolete.</p>

 

</blockquote>

<p>If that is true, what about the Nikon 800e versus 800? I assume Nikon lenses that work fine on the 800 would do equally well on the 800e with its thinner protective glass. </p>

<p>I am not very knowledgeable about the physics involved (I do understand the difficulty of ray angles on pixels for short back focus RF lenses, but little beyond that) but cannot the camera firmware be set for different optics, a bit like the M9 compensations for differing lenses (which I guess is mainly to compensate wide aperture vignetting)?</p>

<p>Probably something worth querying Sony about, as there are many of us who would love to buy a Sony and use our RF optics with it.</p>

<p>Firmware can correct for many things, like vignetting and distortion, but not for coma and curvature of field. Light striking a plate of glass is refracted least when perpendicular, and more as the angle of incidence increases (Cosine Law). Light approaches the sensor from various parts of the lens in a cone, converging on the focal plane. Passing through the cover glass, this convergence is disrupted. When refraction occurs, you also get some dispersion depending on wavelength.</p>

<p>Zeiss takes this into account in designing their premium lenses for the Sony A7. The cone of convergence is calculated to compensate for this refraction at the edges of the field. The cover glass is treated like another, predictable element in the lens when calculating the optical design. In the simplest sense, the field curvature is inverted from its usual shape, along with lateral CA.</p>

<p>In order to minimize cost and weight, Zeiss has elected to let distortion slip in the Batis lenses, while optimizing resolution and CA, and the cover glass "problem." Except for that compromise, Batis lenses are nearly equal to Otus lenses in performance, but at 1/4th the price and 1/2 the weight. The distortion approaches 3%, but in a predictable fashion, and is corrected in firmware to less than 0.5%.</p>

<p>I wish I could cite sources, but I do a lot of research before spending Zeiss prices for lenses. The design issues have come out in several essays, which are lost in the fog. I have some knowledge of physics and optics, and have added some things above to make the principles clearer to a casual user.</p>

<p>I understand that the angle of incidence, if far off the perpendicular, is a problem for pixel arrays which unlike film emulsions are sensitive to angle. The M8 was corrected somewhat for this by having small lenses or prisms in the cover glass that changed the angle of incidence otherwise seen by the sensor faced with RF optics. Does the M9 have a similar disposition? I assume so, as the width of the capture field is much greater.</p>

<p>Other than the better ISO performance and tiltable monitor of the Sony, the lesser number of pixels of the M9 (18MP) versus the Sony A7R II (42 MP) may not be that significant, but that would be a reason for upgrading if the Leica wide angle lenses could be better accommodated on the Sony. I am sure that I am not alone in having invested in Leica glass of that type and now faced with few opportunities to obtain higher IQ (other than that of the expensive Monochrom 240P for B&W).</p>

<p>The M9 has a thinner cover glass, 1 mm if I recollect correctly. Nor is there an anti-aliasing filter. Furthermore the microlens array of the M9 sensor is designed around existing Leica lenses, in which the rear element approaches the sensor quite closely, even in the 50mm Summicron. Wide angle Leica lenses (and Zeiss ZM lenses) behave as well in the M9 as on film (actually better, because the resolution is so much better and cleaner than film).</p>

<p>Digital sensors are sensitive to the angle of incidence, but this causes vignetting, not a loss of resolution. The thick Sony cover glass displaces the plane of focus by refraction and increases dispersion. Even flat glass increases the optical path by about 1/3 of the thickness in the center. The effective thickness increases with the angle of incidence.</p>

<p>Magnification of the M9 viewfinder is 0.58x in order to accommodate 28 mm lenses without an auxiliary finder. Consequently it is very difficult to focus a 90 mm lens without a magnifier, and focusing a 135 mm lens is hit and miss. M240 and newer models have live view, and the latest models accept an auxiliary finder with eye level live view. The A7 has 100% live view with peaking and magnification as focusing aids. The focal length doesn't matter, because the viewfinder is always the same as coverage as the lens. Focusing is easy and precise, even for longer lenses.</p>

<p>For most purposes, your Leica lenses will work just fine. There are limitations if you look closely, as the following examples may help illustrate.</p>

<p>I uploaded an image with the same scene (in January) used in my 50mm lens test, using the Sony A7ii with a Summaron 35/2.8 lens. This is a full resolution (6000x4000) camera JPEG, so it will display only as a link in this thread. Notice that the center is very sharp but the bare branches at the edges are smeared if viewed on a pixel basis. <br /> http://d6d2h4gfvy8t8.cloudfront.net/18047155-lg.jpg</p>

<p>The following is the same scene (September) taken with a Leica M9-P and the same Summaron 35/2.8 lens. There is some loss of sharpness at the edges, but no smearing.<br /> http://d6d2h4gfvy8t8.cloudfront.net/18047156-lg.jpg</p>

<p>Good information, Edward, thank you. I heard that the otherwise excellent 35mm aspherical Summicron does not do well in regard to edge resolution on the A7r (probably similar on the new A7rII). I use it and the 21mm f2.8 aspherical for much of my work, so it would be unfortunate if both smeared edges with an A7 camera body. I also use the 35mm VC f2.5 classic lens as well as their 12mm ultrawide and presume both these wide angle rangefinder optics would behave badly with the A7.</p>

<p>The 240 is too expensive for me as I already invested with the M9, the trade-in value of which is probably only one half the new price. The new A7rII is tempting, but not a perfect match. I guess I could sell much of my Leica gear and acquire the Sony together with some good Zeiss optics designed for digital. The difference between18MP and 24MP is not great (despite other advantages of the newer M), but a jump to 42 MP might be more significant, especially with apparent Sony advantages like the higher ISO performance, tiltable monitor and other. I wouldn't need to change if I could happily to stay at or within 12 x 18 inch enlargements, but that is an issue for me.</p>

<p>42 MP is only about a 40% jump in resolution over 24 MP. Absence of an anti-aliasing filter in the A7Rii adds another 25% advantage over the A7ii. That's significant, but not revolutionary. According to Sony, the odd number was chosen for optimum compatibility with 4K video with a full frame sensor.</p>

<p>Judging from the distance the Elmar 21/2.8 extends from the rear, similar to the geometry of the Zeiss Biogon ZM lenses, I expect that there will be noticeable smearing in the corners. Whether it spoils the shot depends on the subject and degree of enlargement. Neither Zeiss nor Sony make a prime lens of this length at this time. However the range is covered by a Sony-Zeiss 16-35/4 Vario-Tessar, which is very sharp, flat field, but with distortion which is well corrected in firmware (net < 0.5%). I suspect that Sony or Zeiss will announce a 20mm or 21mm prime in the near future - it is a "hole" that needs filling.</p>

<p>I have a Nikon 20/2.8 AFD which works surprisingly well on the A7, but from my limited experience with it, the corner sharpness is not that good, and probably not as good as your Elmar ASPH, and not even close in the center.</p>

<p>Your 35mm ASPH will probably work well. If not, I recommend the Loxia 35/2. Some examples at f/2 are below...<br>

<img src="http://d6d2h4gfvy8t8.cloudfront.net/18047220-md.jpg" alt="" width="680" height="453" /><br>

<img src="http://d6d2h4gfvy8t8.cloudfront.net/18047219-lg.jpg" alt="" width="600" height="400" /><br>

<img src="http://d6d2h4gfvy8t8.cloudfront.net/18047218-lg.jpg" alt="" /><br>

<img src="http://d6d2h4gfvy8t8.cloudfront.net/18047217-lg.jpg" alt="" width="600" height="400" /></p>

<p>http://www.stevehuffphoto.com/2015/03/16/battle-of-the-champions-leica-m-50-apo-vs-sony-a7ii-50-zeiss-loxia-by-brad-husick/</p>