Photography’s future is software & computation

10 thoughts on “Photography’s future is software & computation”

  1. Lots to unpack.

    Photography as more than just a photo, but as a medium for communication. When I worked at IG this was something that was implicitly understood and explicitly incorporated into design decisions. Image based communication, while subject to misinterpretation (as is text or audio) transcends language and culture in ways that other mediums can’t. Memes are one example, TikTok dances another, and more subtly replicable (and replicated) types of photos, often tagged, i.e. #followmeto #lookinguparchitecture

    The crossroads that camera makers find themselves at is a road that they’ve been on for some time. The point made by Tamaki-san about how to incorporate computational photography into cameras, and the reluctance by manufacturers to do so, is in equal measures understandable and baffling. Given the results yielded by combining computational photography with the limited optical capabilities of phones (OK, one of the main reasons is to account for those limitations) it’s tempting to think about what the combination of high-end optics, large sensors and computational photography would deliver.

    Also understandable, but something of a shame, is the dearth of exploration by camera manufacturers of combining phones with cameras. Samsung did this with their Galaxy Camera and Galaxy NX, before they exited the dedicated camera business. Nikon made a half-hearted attempt with the Coolpix S800c. Panasonic had the CM1. Sony was perhaps the most experimental, with their QX line, which were “lens cameras” that combined large sensors with high-end lenses that communicated wirelessly with phones.

    Sony still participates in the space with phones like the Xperia PRO-I. Lastly, and perhaps most intriguingly, was the Leica T (followed by the TL and TL-2), which is now also abandoned as Leica exited the APS-C segment. The Leica T and its successors were physically very slim, had a large sensor, and were built around a touchscreen interface, with limited physical controls (which could be customized) and were designed to work with both L-mount and M-mount lenses (they lacked good tools for manual focus, however). There were other notable attempts, including the Light L16 and the ill-fated Zeiss ZX1.

    On the one hand, it feels like there’s the opportunity for camera makers to resume this type of experimentation, on the other they’re mostly facing harsh business realities that make it hard to do so and most importantly, there’s the (lack of) a consumer use-case (vs. phones) that hampered these experiments in the past.

    Phones do many things well to very well and some things uniquely, while being more than good enough cameras for almost everyone, and skilled photographers can push them to create great photos.

    Cameras, increasingly, are designed to cater to the needs of a shrinking group of people who buy standalone cameras, which means investment by camera makers into features that differentiate cameras from phones, catering to photography use-cases that phones can’t (yet).

    Paradoxically, this makes cameras less general purpose and thus further shrinks their addressable market, compounding the already tough situation.

    1. In my life as an investor, I met two companies that had each individually come-up with a proper solution – chipset, embedded OS with all sorts of connectivity and ability to interoperate with iOS, and Android. They got no traction with the camera makers even though it was start of the decline.

      Interestingly, Leica tried to go for it with Model T and they didn’t really quite understand what a magical device they had on their hands. It was literally the perfect camera. With UWB and Bluetooth 5.0 it would have been a perfect connected camera. They could have used this as a springboard for a new camera experience. I argued with Leica folks quite a bit about it.

      PS: I still have a QX in my hardware drawer.

      1. That’s a shame re the companies you refer to, perhaps that little bit too early…

        Agree re Leica T series. I owned a TL-2 on three different occasions – it was the first Leica I ever bought actually – for exactly this reason; so close…yet so far away. Also a shame that they’ve given up on it.

        Also have a QX-100 floating around somewhere!

        1. I still use TL2 as my travel camera. It is way easier to carry it as a backup camera and not have to deal with weight and big lenses. They did a really good job with it.

          1. I am not quite sure that we all quite understand that photography is less about gear and technology. Instead it is all about the very action and experience itself.

  2. Hi Om,

    Couldn’t agree more. Manufacturers are barely beginning to scratch the surface of computational photography. And yes, innovator’s dilemna has hit them hard. They have been fighting a performance war for so long that they forgot that 95% of the market does not care about camera performance but is in photography for the enjoyment of a creative hobby. Payback has been very harsh, but deserved.

    I see computational photography taking three directions :

    (1) More performance, from the mainstream brands. We are seeing this in the transition from AF, to face-detect, to eye-detect, to animal-eye detect, to bird-eye-detect and probably a lot more to come in the same vein. Improving multishot stabilization comes to mind. This has been hugely beneficial to a type of photographer, and will continue to be, though only to a minority.

    (2) Done for you photography. Already, we are seeing Samsung phones replacing a burnt out moon with a simulated moon, and post-processing apps that replace skies in a couple of clicks. This will no doubt come to cameras and will be very popular on Instagram.

    (3) I would bet the third improvement will be more artistic. Here I’m thinking of creative profiling, such as what Pixii has done with the Monochrome mode, taken even further to capture the resurgence of analogue love without leaving the digital domain. I see a large market opening up here, for those who cannot be persuaded to go back to film, scanning, darkrooms … but want a much more involving experience than what phones and mainstream cameras offer.

    There is a fourth, but I think most manufacturers have caught on to it : phone apps that allow to interface the camera with the smartphone ecosystem. Interestingly, this seems to be the domain of small, high quality cameras. Leica and Hasselbald both have such apps, and even small Pixii, again, are on the way. But I’m not sure this is worth talking about in the future tense, as I’d be surprised if all manufacturers didn’t have a great app in the near future.

    Thanks for the very interesting post! Looking forward to more 🙂

    1. There is a lot of opportunity for companies like Pixii here as they need to go more on computational photography and less on the traditional stuff that camera people do. As I mention above, UWB/Bluetooth LE 5 and better/lower power WiFi means, the symbiosis of camera and phone can go quite far. Sadly, camera people don’t know how to think in “software” terms.

      1. So true. We are witnessing a new shift. Mirrorless caused a major transformation in the industry. I believe AI and connectivity will cause another, too. Good for us, the consumers 🙂

  3. Actually, I believe option n°3 corresponds to what Kazoo Tamaki refers to when he says “We don’t need to satisfy customers who just want to play with images or want a cartoon look. We have to follow the tradition of photography.”


    1. I think even people who play with images for a cartoon look start with a normal image 🙂

Comments are closed.