[Winder]

The future of photography is code – TechCrunch

[stens]

Yep. Concur with this article. Phone cameras have become better with each generation. I am afraid that it won’t be Canonikon that finishes off Pentax, but Apple, Google, Samsung, etc. of course, they’ll most likely also displace Canon and Nikon.

I’m not ready to give up my K-1 yet, not really anywhere close. Phones don’t have a quality telephoto or true macro, among other features they lack. But the images keep improving: Portrait mode actually works, and the phones are leveraging the image streams, powerful processors, and equally powerful software to make great images, as noted in the article.

The camera companies so far don’t have anywhere near the on-board processor power, nor do they have the requisite software engineers to make use of those processors. Man, I’d like an A12 chip and Apple software in my K-1...but I don’t see that happening (and Apple would try to include a nonremovable battery...). Maybe Sony...but even they can’t keep up with the newer tech giants. I don’t see anyone of the mainstream camera companies who can make any great leaps in processors or software, unless they would ally with one of the tech giants. But what would be in it for Apple, Google, or Samsung? Phone cameras and interchangeable lens cameras don’t have much in common, except for creating art with light.

I’m returning to Kenya again this winter. But, as it’s a working trip, and I’m not going on safari, I’ll just take my phone and travel light. I’ll get some great pictures.

[tvdtvdtvd]

Modern smart devices are truly a wonder. My 'old' IPhone 6 takes really great photographs which look good without any additional processing.
Color rendition, white balance, exposure, etc, are all very good to excellent, and I don't have to tweak a single setting to make that happen.
Newer smart devices have only improved in capability. Yes, I can take a better photograph with my K3, but I have to work for it. I do think it's
time the grandaddies of photography, (Pentax, Canon, Nikon, etc), put a little more R&D into the intelligent processing capabilities of their cameras.
Not as a replacement for RAW photography but as an adjunct. I do think there will be a place for 'real light' photography for many years to come
but that doesn't mean a modern DSLR can't do much better when set to full auto mode, (or even partial auto).

[biz-engineer]

The downside of smartphones is that they don't feature a K or M42 mounts, so until Google or Apple or Samsung come up with a K mount simulator and M42 to K simulator adapter, it won't be possible to get Takumar style bokeh images out of the phone. The recourse to computational techniques in mobile phones means they have reached the limits of improvement on the hardware side. That indicates that the limits of phone capability are going to saturate very soon. Those computational techniques can easily be implemented in standalone cameras, if so, standalone cameras will still be more fun than phones.

---

Last edited by biz-engineer; 10-24-2018 at 11:56 PM.

[microlight]

Very interesting indeed. I suspect that we’ll know if and when the crossover point is reached, when pros start using their pocket multimedia centres (they’re not just ‘phones’ any more) to take pictures that they earn a living with. Until then (and maybe even not then!) I prefer to use the Pentax, not only for the quality of the photos but for the tactile pleasure of handling it.

I already mostly use my iPhone 6+ for video in preference to the K-3II as it’s a whole lot of less work: the stabilisation alone knocks even the older K-5 type of mechanical stabilisation into a cocked hat, and since I have the jellifying software stabilisation turned off in video on the K-3II, it just adds another substantial processing layer of having to de-shake in post. I’m no video buff (it’s family events mostly) so in good light (and there’s the caveat) I’ll use the iPhone without a second thought. More creative stuff in low light, wide aperture for bokeh or differential focus, then it’s typically back to the Pentax, usually with the lovely little M40/2.8, or A50/1.4, and VirtualDub.

[Andrea K]

In the future the photography wil be a voice command to Google Assistant or Siri.

[BigMackCam]

Interesting article - thanks for posting.

Where it misses, for me personally, is the lens aspect...

It's not enough for a lens to be optically near-perfect in a technical sense. The character imbued by each lens is terribly important (at least, it is to me, and judging by these forums, I don't think I'm alone). If we leave our DSLR / mirrorless cameras at home and start shooting with iPhones, we'll all have that same lens and a bunch of clever processing to produce effects and emulations such as shallow depth-of-field bokeh, and everything will start to look the same except for in-camera- or post-processing adjustments.

In one sense, that could be seen as a great leveller, in that our abilities and techniques become the sole differentiator... in the same way that certain classes of motor racing allow just one manufacturer, model and preparation of race car, and the results are primarily based on driver performance. But this is photography, and each lens plays a unique role in giving our images a certain look and feel. If I'm shooting with an 85mm lens, some days I might choose my Jupiter-9, while on others I'll go with the Helios-40-2. Same field of view, but wildly different rendering, and that has a big impact on the photographs I take. It also requires a little knowledge, experience and judgement in picking the right tool for the job - and that's part of the fun.

Of course, technologists (and I say this respectfully, having spent much of my career in technology-related roles) will say those lens characteristics can be emulated in software, and to some extent I'm sure they can. But those emulations won't capture every nuance of a lens. At best, they might offer a vague approximation, in the same way that film emulations in post-processing vaguely approximate the real thing.

So, whilst I look forward to seeing what can be done with code - and I've no doubt it will be both impressive and, in the right applications, very useful - I'll stick with my interchangeable lens cameras and a selection of - thankfully - far from perfect lenses...

---

Last edited by BigMackCam; 10-25-2018 at 03:29 AM.

[Not a Number]

It's just image processing. Do it offboard on a laptop or desktop. Who da thunk it?

[brofkand]

The computations that Apple and others apply to their photos are mostly to make up for deficits in their designs. The phones are thin and light so they can't put in large sensors or zoom lenses or a Xenon flash.

DSLRs do not have any of those deficits. I do not see a need to apply "computational photography" to DSLRs.

The phone has killed point and shoots and probably even low-end DSLRs like the Pentax K-S1, but I don't see DSLRs or MILCs as a whole going away because of phones. Some brands will be casualties but that's business. The photographer's control, image quality, and flexibility are what DSLRs and MILCs do so well, and I don't see that overtaken by phones.

For most of the past year or so, I only took my phone on trips. It was okay, I was able to work within the limitations, but in a few weeks I'm going to Bermuda and will be bringing my Pentax gear along.

[photoptimist]

The reason phones will never fully replace "cameras" for photography is that these purpose-build cameras have:
1) a wide-range of interchangeable lenses (not just a couple of modest-focal length fixed lenses);
2) a device body shaped for photography (not a hard-to-hold phablet);
3) dedicated tactile buttons, dials, and wheels for responsive controls (not a laggy touchscreen);
4) an eye-cup viewfinder for immersive inspection of the scene (not an open-air display).

[photoptimist]

It is true that the role of code will surely increase. That's especially true for casual photography where non-photographers simply want a pleasing SOOC image for sharing without bothering to learn any of the complexities of managing exposure, lighting, color, etc.

Yet I can't help but fear that it will lead to a homogenization of photography -- a vast mass of decent shots heavily auto-post-processed according to Apple/Google/Samsung's standards of what constitutes a good image. The mass production of decent photos certainly serves the basic needs of capturing and sharing memories of events but it does not serve the more advanced needs for creativity and art.

These kinds of tools also create a challenging chasm for users of non-smartphone cameras. The advanced code and high-performance processors enable a smartphone to do in a fraction of a second what it took a Photoshop user minutes to do and months to learn. A novice photographer might be quite frustrated that they are having a hard time getting images out of their brand new "pro" camera that look as good as their smartphone images.

Makers of dedicated cameras will need to find ways to replicate or borrow the auto-post-processing power of smartphones. I wonder if one solution is to link the camera to a smartphone with both looking at the same scene. Both cameras fire at the same time and then the big camera loads its image into the smartphone for automagical post-processing. Even if the big camera has an ultra-wide angle or ultra-telephoto lens, the smartphone's view of the scene would help cue the lighting and color management software.

[derekkite]

Sure. And every new phone iteration which has improvements in the camera has a better sensor and better lens.

If your photography is wide angle good light stuff, great. I use my phone camera all the time to document my work. But when I shoot what I want, I have a large sensor and nice glass.

The phone satisfies most people who want to take shots. Fine. They are remarkably good.

In the marketplace something cheap and low image quality can't compete against what most people already have in their pocket. People who want more than an $800 phone know they will have to spend a bit of money to get it. Hence the sweet spot on the market is about twice that and higher.

[MossyRocks]

They are starting to run into the hard physical limits of what can be done on with small sensors and tiny lenses. Because of this large sensors and larger better lenses will always win out when discussing image quality (the quantifiable objective measures). There are sensors with a pixel size that is rapidly approaching the wavelength of red light (800nm pixel size vs 750nm for the bottom of the red) so there are very real physical limitations with ability to gather light and noise that are being hit. Then there is the question of producing lenses that are capable of providing the necessary sharpness and light gathering ability. At a 800nm pixel size a perfect lens would need to be something like an f/.6 lens or there about to not be diffraction limited again running into some hard limitations with lens design and how it functions. Notice that there the phrase ideal lens was used, cheap mass produced lenses used in cellphones will not be these optically perfect ideal lenses. A tiny 3mm focal length lens at f/.6 is going to have a lot of issues to overcome.

Far too many people think that having better gear that is capable of giving them better image quality will result in better pictures (subjective measures). Picture quality has far more to do with the person behind the camera than the camera in front of the person. For example I take reasonably good landscapes but I suck at portraits, having only managed to get 3 that are pretty good. The first was just dumb luck with the first roll of film I ever sent through my Spotmatic F with the S-M-C 55mm f/1.8 lens, the second was one was when I shot a roll of B&W film of all the family members one Christmas with the spotmatic and 135mm f/2.3 Vivitar lens, and the 3rd was a couple weekends ago when one of the cub scout girls at fall family camp said "take my picture" and I quick snapped it with my K-3 and cheap 28-300mm f/3.5-6.3 tamron super zoom. In all 3 cases it wasn't the gear that made the photos great, especially the film shots, it was my pure dumb luck in getting the image right

However when I do the photography activity each year with the cub scouts I always point out that it mostly doesn't matter what camera you use to take a picture since most people are not limited by their gear, even pros and serious amateurs. Granted there are very specific cases where gear is limiting but if you are shooting in a normal range of light and at what would be wide to smaller telephoto ranges (basically 28mm to 200mm on 35mm) at normal speeds, most people will never be limited by their gear even if it is a cellphone.

Also for most people the high automation with cellphones is a good things as it eliminates an awful lot of garbage pictures. The cellphone will usually get the focus, exposure, and white balance good enough most of the time. Add in that most pictures are viewed on tiny screen or are small prints and gear really becomes less important.

Finally anything that can computationally be done with cellphone images can also be done with higher base quality DSLR or MILC images. Besides I like having a camera that will let me take a "crappy" photo, at least it is my photo instead of what the camera thought I wanted.

[pres589]

There's no way to properly mount my SMC DA 15 Ltd. to a smart phone. No sale, not interested.

[clickclick]

Yep, neat stuff, but I'm in the "it's all about the lens" camp. I truly suffer from LBA, and the reason is I've been bitten by the character aspect of the different lenses. I'm spoiled by a neighbor with nice gardens and tons of lovely blooms. I get calls from him when new stuff pops. I then go look at my pile of lenses and grab different combinations of body (K3II or K-1) and lens based on my mood. For some, I know this may sound nuts, but my Sigma Art 35 and CZJ Biotar just produce totally different results, and when one fits, it fits. I just don't see that happening from someone's programming that can't see and feel what I'm experiencing at the time. I also wonder when photos taken with such technology start "looking alike". Sure, you can argue that you could write the CZJ effect in software, but somehow, I just don't think it's the same as bending the light rays on the way to the film or sensor.