...re-reading some columns by Nikonista Thom Hogan...and seeing that Fuji is announcing another new, free firmware upgrade for the X-Pro-1, got me thinking about the future of DSLR upgrades.
I recall Canon's commitment when it released the original F1 that they would not release a new model for at least a decade. Their purpose was to demonstrate to professional photographers that they were committed to their pro-level SLR and that photographers could purchase the camera knowing it would be fully supported.
For most of the past decade, gear enthusiasts have been spoiled by the remarkable and near continuous improvements in DSLRs. But, those improvements are increasingly coming at the margins. Higher megapixels, increased dynamic range, high and low ISO improvements, while nice, are generally needed only under very specific conditions for very specific purposes.
Any honest assessment would acknowledge that for 90-95% of subjects and conditions, the cheapest entry-level Canon and Nikon will produce results that under real world conditions will be indistinguishable from the flagship models.
Fuji has followed a path with their X-Pro-1 of releasing firmware updates to keep the camera current and boost customer satisfaction and loyalty. (To be fair, Canon did much the same thing when it released it's major firmware upgrade of the 7D – extending the practical life of the camera and effectively giving customers a free "7D.20" version of the original.)
So, having said all that, I am going to go out on a limb and predict that the upgrade cycle for the 5DIII to 5DIV will equal or exceed the cycle between the 7D and 7DII and that we are entering an era in which upgrades will be fewer and further between.
At the same time, I am going to suggest/hope that Canon and Nikon will offer more significant firmware upgrades during the interim.
This will certainly require some adjustments to their business model, but in a sense they are simply going back to the model that both companies followed successfully for decades. That's one reason why I believe Nikon and Canon are better positioned for long-term success than companies like Sony, which got into the digital camera market during the boom era and do not have the institutional memory or experience to easily adopt to longer development cycles and more modest sales growth.
I see Canon as particularly well-positioned for this change. They have aggressively developed products for new markets, especially the booming cinema market where growth is fed by the seemingly unquenchable thirst of the internet for new video content. Their recent emphasis on security cameras also shows they are prepared to move into another fast growing emerging market. I am less convinced that Nikon is equally well-positioned, but then I don't follow them as closely as I do Canon.
So what's the point?
In part to feed off the idea of "10 years from now" and in part to get people out of the rut of trading insults over dynamic range and other esoteric subjects that do not sell cameras and do not matter to the vast majority of photographers.
What is your prediction? Will we see fewer upgrades in the future? Will we see more substantial firmware upgrades? Given that Canon and Nikon need to continue to sell products, do you think they will become more aggressive at selling lenses, strobes and other peripherals? Will you spend less money on photography in the future, or will you just spend it differently and how?
Very interesting thoughts!
Personally, I'd like to see a slower cycle. Longer period of time (4 years plus) between major body releases. For example, the 1D X is a phenomenal camera, and I don't see it "needing" to be replaced for years to come. Even if Canon put a sensor with more DR into it, that would only affect lower ISO settings, which are so rarely used when photographing action...I really wouldn't see the point.
I do indeed hope to see more firmware updates. So much can be done with firmware, now that we have eminently more capable hardware. I see no reason new features, say focus peaking or even more information displayed in Canon's Transmissive LCD in their OVFs, couldn't be added through firmware. For features that require fairly significant development time, I don't see why us, as customers, shouldn't even pay a small fee for feature-enhancing firmware. I am not saying we should pay thousands of dollars just for a significant firmware update, but a few hundred for a firmware update that moves the 1D X into the future a few years from now, keeping it a competitive camera...I think there could be a business model augmentation to be found in such a paradigm somewhere...
I cannot say whether I'll spend less money on photography in the future. I think I'll spend what I feel I need to spend, which is what I've always done. Sometimes my photography expenditures are high (or even very high...such as this year, over $15,000 spent on photography so far, where as last year, I only spent a few hundred), and sometimes they may even be non-existent. When it comes to camera upgrades, I do need a compelling
reason to spend my money.
Both the 1D X and 5D III are quite compelling to me. The 5D III simply because it is very good at everything, and even though it may not have all the best technology everywhere, it is still a superior product overall. I do mostly bird and wildlife photography these days, and landscape/macro photography more seldom. That is in part because I really love birds and wildlife, but also in part because I like to maximize the detail...in highlights, midtones, and shadows...of my landscapes, and simple fact of the matter is, it is probably the most demanding form of photography for dynamic range...and Canon cameras don't measure up. Sure, I can get around the problem of Canon's read noise, and I do...although it DID cost me more money, as I had to buy an expensive set of Lee filters and the Lee filter system in order to balance contrast on-site, and Topaz tools and Nik tools in order to extract the maximum amount of detail from my shadows in post. That was an investment of well over a thousand dollars, above and beyond the money I spent on my camera bodies.
Better technology gives me more control, more capabilities, and less need to spend extra money on the side to extract the kind of detail I really want from my photography. To me, the most compelling camera that I would be most interested in buying next is one with as many megapixels as Canon can reasonably pack into a sensor, and at least 13 stops of dynamic range. I would probably settle for more megapixels, and "the same old DR", if that is all Canon can come up with...but I have to be honest...I'd part with my money with more difficulty than otherwise.
I believe Canon is a highly innovative company...but I do think there needs to be some compelling developments in their future product releases to keep me buying. If they don't innovate compelling new technology, and especially innovate and progress enough to keep up with the competition, it will be ever more difficult for them to get me to spend my money on their products. I probably still will...I have too much invested in Canon glass not to...but it will be late, and lazy, never at release when they are getting the most revenue from their products...instead I'll wait for the best sales time can offer, and snatch up their products at a significant discount, resulting in the minimal amount of revenue for Canon. I'll be late to the game with new gear, possibly by a couple years, but that will be the only way I feel as though I'm getting my moneys worth.
In the most extreme case, I'll break down and buy the most compelling product from a competitor. Sony's A7r is intriguing, because it could make an excellent landscape camera that would work with EF mount lenses. I don't really need AF or other high performance factors for landscape photography, so the reports of the A7r's poor AF performance with adapted EF lenses wouldn't be a huge deal. The remaining problem is...it costs quite a bit, $2700? For "just" a landscape camera, even though the sensor is amazing, eh, I'd still have to wait until a particularly deep sale occurred before I could justify the purchase.
So, to me, compelling innovations included in compelling upgrades. That's the most important thing Canon can do. They don't necessarily have to compete feature for feature
with their competitors, but I do think they need to keep abreast of their competitors in terms of the level of their products.
Reading some of the excellent commentary here, especial Jon Rista's take on the amount of headroom that remains in ISO improvements (I don't for a second pretend to understand terms like quantum efficiency, so I have to take his word for it)
You don't necessarily have to take my word for it. Here is a little visual diagram demonstrating Q.E.:
The two diagrams are identical in every way, except the number of electrons stored in each photodiode. The amount of photons raining down on each sensor is identical. The microlenses and color filter array behave identically. The only difference is that the one on the left has 40% Q.E. and the one on the right has 60% Q.E. A photodiode works by converting incoming photons into electrons, an amount of electrical charge, which is stored and can then later be read out via a special process. Photodiodes are not 100% efficient, especially at room temperature, so instead of converting every single photon it detects into an electron, each photon that strikes the photodiode increases the probability of an electron being released.
For photodiodes that have a 40% quantum efficiency, it takes approximately 2.5 photons for that potential to become realized, and release an electron. That means, every two to three photons detected, in a roughly even distribution, one electron is released. For photodiodes that have 60% quantum efficiency, it takes approximately 1.667 photons for that potential to become realized, and release an electron. That means every one to two photons detected, in roughly even distribution, one electron is released.
In other words, it takes less time for a sensor with higher quantum efficiency to achieve the same exposure. In the same amount of time, the exposure for a sensor with higher quantum efficiency will be greater. At high ISO, a sensor with higher Q.E. will have a greater exposure...more electrons, greater charge per pixel, requiring less gain, to achieve the same result. Less gain, less noise.
It should be relatively self explanatory now to understand that once Q.E. approaches 100%, where you basically get 1 electron for every 1 photon detected, you've literally reached the limits of what physics will allow. There is no Q.E. above 100%...not literally. Someone may come up with a means of improving sensor sensitivity even more...historically, changes to photodiode area is the only thing that affects the maximum number of electrons that a photodiode can hold before leaking charge or increasing in temperature (any additional photon strikes once you reach the photodiodes capacity either result in releasing an electron, which means another electron has to be "lost" somehow (into the neighboring photodiodes, into the wiring current, etc.)...or the photon is converted into heat.) It may be possible to put multiple layers of photodiodes into a pixel...allowing electrons to be collected "three dimensionally"...in area and in depth, which might increase maximum well capacity and the potential of converting a photon, therefor increasing charge, and therefor signal strength, at higher ISO settings. So far, layered photodiodes have only accompanied sensor designs like Foveon...but there is no reason, necessarily, that we would have to employ all three colors at each pixel, we could stick to a bayer design and gain the benefits of gathering photons in area as well as depth.
Anyway...hope the diagram helps.