This is continuation of last year's summary of my thoughts on a subject of a digital camera and its future.
First of all, I'm not writing about a camera I recently parted with, because I didn't. Even more, with holiday season coming I was on a lookout for something (GAS, you know...), but struggled to see an object that could be of a valuable addition and/or replacement to one of currently owned devices. So FZ1000 is definitely safe, with its versatility, more than satisfying image quality, very good handling and option for 4K. Then there is Olympus 12-40mm f/2.8 lens: with its sharpness uniformity having no contender amongst other systems and lovely contrast, it makes the perfect day-to-day piece of glass. Atached to E-M1 body it balances well, and this camera was the first I managed to complain that it offers too many direct access buttons (to be fair, only after using the lever) and needs refreshment course in operation once not used for longer. Then there is E-PM2 - as good image quality as any other MFT body in as little as not many package, with equally diminutive Panasonic 45-150/4-5.6 complementing range of angles of view. Last, but not least, the beast: Nikkor VR 70-300 mm f/4.5-5.6 on Nikon V1 body.
So what happened? Quality of images produced by these cameras (except for V1), in typical conditions I photograph (outside, daylight, no rain) and typical output being WQHD (2560x1440) display or 1000px wide export for web is sufficient for me. Full stop. I occasionally find challenging conditions, but not often enough to justify foray into large sensor teritory. Operation of the cameras I own, even though not free from quirks, is also satisfying (Nikon excluded). In the current market climate there are not a lot of objects desirable enough to reach to the pocket.
What is currrent market in my opinion then? Most of all, progress in sensor sensitivity improvement slowed down, so new generations of cameras aren't much different than what they replace. Changes seem to be focused mainly on polishing operation ergonomics (which is great!), but that hardly justifies price tag coming with 'just released' being actually 'mild facelift'. But it definitely is great market for buyer, with cameras perfected in their n-th generation, and old-but-as-good-as-new models available for a fraction of price asked 2 years ago: E-M5, which still is a great piece of equipment (see reasons above), can be had new for £349. With new models inevitably coming, that pushes even greater pressure on the market, and soon its players will need to adjust. One already seemed to have found a solution: Sony uses its lead in sensor manufacturing as well as innovation in camera forms and capabilities and focuses on the upper end of price brackets. Canon and Nikon do their usual (not a lot). All the rest try to innovate, but the scale of sales is against them.
What might happen?
This is a speculation based on my observation of digital imaging devices market and my experience with some of those devices. I started editing this post come time ago, and so the opinions were evolving, rather than being a snapshot of the current state.
Again I'm going to divide into specific areas, and refer to the previous 'forecast'
4K - quite surprisingly, Ultra High Definition TV receivers seem to be selling well, so the gap created by the lack of content will help leverage adoption of 4K recording in digital cameras even more.
8K - wooah, not too fast? Maybe not - since 4K painfully paved way for the underlying technology, 8K actually isn't that far away. Its advantage is obvious : 8K=4*4K . Simples ;) But most of all Sony new 42Mpix sensor is exactly that: 7952 x 5304 easily cropped to 7680×4320. Easily downsampled to 4K. Or heavy cropped to 4K. Take your pick.
I hoped to see the end of megapixel war, and even though Canon had to breach 50M barrier, the main player seems to be more pragmatic: use as many sensels as it is really needed, and only so many so reading them fast is not more challenging that required. The slight bump in pixel count from current best 36Mpix (7380 x 4928 - just shy of 8K) over the same surface should be offset by using BSI structure.
C'mon, it was supposed to be about photography, wasn't it? Well, I feel this is where the biggest shift is going to happen: with 4K, and more so with 8K, having 24/30fps with so many details will cause traditional 'still' approach to be greatly reduced. With serious consequences.
Stabilisation - back in the stills domain for a moment, spatial resolution of the sensors became so high, that any minute movement causes image to be blurred at pixel level. To the extent that electronic first curtain shutter of Nikon D810 is the most notable upgrade of that model versus D800, and good enough reson for many to switch. Sony A7RII headline feature is 5-axis IBIS, greatly expanding its shooting envelope. With video good stabilisation is even more crucial, as with affordable readout methods pretty much limited to rolling shutter only (I'd love to see some clever mechanical video shutter one day, unless stacking methods finally enable global shutter for masses), so jello and all other forms of vibration are big enemies. With advent of drones and their gimbals, it seems obvious this technology will filter into recording devices. Or upscale from current mage stabilisation systems. And so the video will be still :)
Form - naturally digital still image camera grew on legacy of film, and so video recorders came from studio boxes. Yet nowadays they share the same key components, and their functionality overlaps more and more. Surely, many will remain as their specific specialized form suits well in many cases. But as still/video will blend, the form will ivevitably morph as well.
So let's talk about film SLR: it needs space for a roll of the film, space to expose it, space to collect it, a lever actuated with the right thumb to transport it and a hump housing a prism right on axis with the lens.
Now let's talk about digital mirrorless camera: it doesn't need space for the roll of the film to the left of the lens, it doesn't have to use vertical form of grip because thumb is relieved from duty and neither EVF needs to sit right above the lens (luckily in this case designers can peek at rangefinders and place it in the top left corner of the body). The point is, as much as retro is visually attractive (by being reminiscent?), it actually acts against the ergonomics of a camera: if holding it is as unnatural as a smartphone when taking a photo, isn't that another point against having dedicated device? Cameras keep downsizing, there is less and less room for fingers of the right hand to hold on to, yet the lens does not shift to the left to regain some space. No, there must be 'traditional' chunk of chassis which whole purpose is to cut into left palm supporting the lens. And of course there must be no good grip: cameras 20 years ago didn't have one, so why would you want such comfort today? As much as I can understand that approach for pocketable models with fully collapsible lenses, in case of bodies with interchangeable lenses it is obvious that the glass this is going to be the most protruding part of the combination, so adding decent grip will not make the eventual setup any thicker! And that grip can house bigger battery, more controls or whatever sits to the left from the lens.
Traditional form is terrible when it comes to space: its T shape means that most of the volume of a box around is wasted. Also weight distribution (biased to the left and front - where the lens is) causes handling issues and contributes to blur at exposure (shutter press movement, lens front drop).
Alternatives? Handycam-like: weight supported by lying on the thumb, rather than squeezed, better weight distribution (mostly front-back balance), possible to hold with both hands and use EVF, more useful surface for controls, more efficient use of volume (lighter and smaller device).
Some experiments in that field can already be seen (Canon XC10)
Modularity - in some respects extension of the form. Once controls and preview are detached, by using external modules (smartphone, field monitor, VR headset maybe) then only a box housing sensor, storage, battery and electronics is needed. Potenitally can be wrapped into adaptor with viefinder/controls to become more traditional device if required. Or attached to some handy gimbal. Still to gain traction, but some devices already popping up (Blackmagic, DJI Zenmuse X5, DxO One)
External controls - There was no breakthrough in that field over the last year. There is however one thing I would be happy to change my opinion about: focus point selection method. I suggested a solution that would improve manual focus point selection, as use of 4-way controller is not an optimal method. But seeing excitement about capabities of autofocus of new Sony cameras, I'm happy to vote for this road of improvement. After all, purpose of automation is to speed things up and make more reliable at the same time. If AF can select the right subject as fast (or faster) than fidling with D-pad, and on top of that track it - I'm in. Moreover, it does not require any new fancy controls I suggested (4-way around shutter button, next to lens, on the lens?) and still allows for manual selection, should a photographer or conditions require that. By extension, this can be, and is, applied to many other settings. So my viewpoint now is: why some of those controls are still required, and the unfortunate answer is: because automatic setings are often rubbish. So I'm looking forward to reduce number of buttons by improving functions of intestines.
Internal controls - IoT seems to be the buzzword of today, chips in mobile phones gain more procesing power that supercomputers in XX century, their software reacts on touch, voice, shake and location of the device. And then there is ringfanced world of cameras, where tethering is big halo, redefining function of a button impossible, menus crippled by ergonomic idiots (they can be ridiculous)/sadists (they infuriate users)/psychos (one must be out of mind to complicate simple things so much). It is not all doom and gloom though: Panasonic figured out yet another use of 4K video, employing it into post-focus function. Similarly, Olympus managed to add in-camera focus stacking to their products. The list of missing features is still long: as much as I love the idea of OLED depth of focus display on Zeiss Batis lenses, I cannot quite figure out why none of the cameras offer such function on their displays.
So what is going to happen to digital cameras? In short term major driving forces seems to be:
- rolling of backside illumination process across all formats. That will reduce influence of sensor size to total light collecting area (low light sensitivity) versus required/achievable depth of field
- increase in sensel count to 4K/8K, increase in readout speed
- increase in processing capabilities
- progress in lens manufacturing: complex aspherical surfaces and use of Fresnel lens being already used. Shape tuning ('liquid lenses'), metasurfaces, computational optics wait round the corner
On a subject of the sensor format, 1" type (13.2x8.8 mm) and full frame (36x24mm) seem to be defining types. The former is significantly more capable than imaging devices in phones and compact cameras, still allowing for miniature (pocketable) devices. The latter, being of size of traditional film frame, making it natural reference. Micro Four Thirds (17.3x13mm) advantages are proportions (4:3) and fact of using common lens mount by more than one manufacturer. APS-C: With pressure from 1" from the bottom and full frame from above, it is endangered. On the other hand this is 'native' video format, so 4K/8K should boost its importance. It seems to be the most popular, but also the most fragmented format, and there will be some victims. I would love to see some steps to avoid that, like Samsung switching to Fuji X mount - imagine 4K video (no luck with Fuji body) with those lovely primes.
Back to Earth to see what changes could happen in my camera bag:
- Nikon V1 - first on the relegation list. Recently introduced Nikon J5 is what V1 should have been and is on the list as a potential target. Unless V4 with newly announced Sony sensor will smash in.
- Nikkor 70-300 VR - there is a patent for Fresnel version, which should make it even lighter. I doubt it shows any time soon, and this being a great lens, little reason to think about switching, except... announcement of PanaLeica 100-400.
- M.Zuiko 12-40 - fantastic lens, nothing comparable seen so far, definite keeper.
- Olympus E-M1 - new sensor and EVF will surely make way to MkII, but will it be enough to upgrade? Not any time soon. New firmware version brings more reasons to keep it.
- Olympus E-PM2 - I love its size, especially considering the potential. Permanently mounted Panasonic 45-150 lens is so-so, but I like the way it renders (warm colours, significant falloff) and do not use this range so often to be a problem. M.Zuiko 40-150/2.8 too large and expensive for my limited use.
- Panasonic FZ1000 - 1. use new Sony sensor, 2. add touch screen, 3. improve EVF (especially optics), 4. fit bigger battery, 5. reduce lens to 300mm reach: that will make it smaller and lighter overall, of smaller diameter and less unbalanced (better handling), 6. add more video pre-set capabilities. Sony RX10II will be in troble as its 1st version.