[BigMackCam]

The more I look at the specs of new camera models (mirrorless), the less I am interest to buy anything...

I feel more-or-less the same, generally... however...

it's pure technological stagnation, same tech as five years ago at higher prices, and some mirrorless model are even worse than DSLR.

I'm not sure whether it's stagnation or the fact that we're reaching a point where not much that's truly new and/or significantly better is possible with the given component technology the camera manufacturers have available. And plowing big money into larger technological advances in the current market is a very risky strategy. Of course, the lack of innovation and technology advances may, in itself, be a factor in the declining market, but I'd say it plays a considerably smaller part than general saturation. Given that, you have to ask which, if any, of the brands has a sufficiently high appetite for risk to make the necessary investment for a real leap in technology...

[Rondec]

Sony feels obligated to release a new camera for every new sensor that they come up with. Sort of the opposite of Canon who used the same 18 megapixel APS-C sensor in about 10 different cameras.

As prices go up, brands sell less and they have to increase prices even more just to stay in one place. I can foresee that by the time we get to the A9 IX it will be priced at 699,999 retail price and Sony will sell exactly two copies.

[TedH42]

Definitely the law of diminishing returns applies here. As you pay more you get less extra as the value goes up.
But that said the manufacturing cost of these higher end cameras are actually quite a bit higher than lower end due to:
* Larger sensor - a lot more expensive (harder to get the quality yield for a larger area of silicone than smaller parts)
* Larger sensor = larger pretty much everything (shutter, mirror, prism, and processor to handle the large amounts of data)
* Lower sales volume = high cost of components and labour

The reasons given here are part of the reason for those diminishing returns.

I think another part is that as the customer expectation of quality goes up, their increased need/desire for that quality allows the manufacturers to charge the extra premium for that quality.

[biz-engineer]

I am concerned that digital imaging technology has bottlenecks not addressed by the size of the sensors. A larger digital sensor has the same limitations than a smaller sensor. And more image data requires huge processing power, seriously limiting the cost/quality of the digital imaging technology. Seen with such perspective, there is no true justification for big price differences imposed on customers.

[biz-engineer]

In other words, it seems to me that digital imaging technology is well suited for small electronic devices, but completely fails to rival film imaging in large format territory due to non linear cost increase of sensors and digital image processing that choke on big high resolution digital sensors.

[Alex645]

In other words, it seems to me that digital imaging technology is well suited for small electronic devices, but completely fails to rival film imaging in large format territory due to non linear cost increase of sensors and digital image processing that choke on big high resolution digital sensors.

Yes, here I would agree with you. It is the reason I still shoot medium format film and then have both dark and light room options.

However, if I were still a pro, digital MF is a better value over the long haul and for the time saved not having to develop and scan film. But as I now shoot in one year the equivalent of what I shot as a pro in one week, film is the economic solution for me.

But I can't justify 35mm film over a FF digital sensor, other than if I prefer the analog process and post-process.

Would I love Ricoh to price the 645Z for the same increase from a KP to a K1II? Sure. But if that was unsustainable for them (and it probably is), leading to the extinction of a digital MF model and lens system, I'd rather have it in production at the current price $500 less than the flagship Canon and Nikon FF.

[steephill]

I am concerned that digital imaging technology has bottlenecks not addressed by the size of the sensors. A larger digital sensor has the same limitations than a smaller sensor. And more image data requires huge processing power, seriously limiting the cost/quality of the digital imaging technology. Seen with such perspective, there is no true justification for big price differences imposed on customers.

Time will solve your problem. Ever since computers were invented processing power improves and storage capacities get larger and cost less every year. The first memory card I bought was a 1Gb CF microdrive which in today's money was over £600. You would be hard pressed to even find a card that small now and a card with 500 times the storage capacity is a tenth of that price. The cost of my original 6mp *istD + DA16-45 bought in 2004 adjusted for inflation would now buy you a 36mp K1-II + DFA28-105.


So I don't see any bottlenecks associated with digital imaging technology. It sounds to me that the rant is based on wanting a Rolls Royce but only being prepared to pay for a Mini.

[bdery]

The cost of making larger lenses and larger sensors does not increase linearly with the size. Hence, a FF of MF sensor is much much more expensive than an APS-C one. Of course there is a premium, like for any "luxury" item.

The only sustainable business model in the long term is to deliver value to customers in relation to the price they pay for the product or service.

Value is a relative term. For someone using MF, there is probably no value whatsoever in APS-C, and for someone used to paying 35000 USD for a camera, a 8000 USD model is extraordinarily valuable.

[biz-engineer]

So I don't see any bottlenecks associated with digital imaging technology.

The digital image is coded over 14bits at ISO100, remove 2bits for middle grey, remove 1bit for high light protection, so 11bits for coding all tones from black to white, compressed by a gamma curve (so that's even croarser tone definition), versus unlimited bit depth of analog film since it's analog there is no limit to how smooth film can record tones.


The other bottleneck is that the sensor is 24x36 mm, not 4 x 5 inches (inches! imagine that).


I have a photograph, taken woth the K1 on tripod, ISO200, a bit of shadow raised, but I can't print it larger than about 24 x 36" otherwise it looks fuzzy even at normal viewing distance. For the detail that's in that landscape photo, printed 32 x 48" , I'd need around 140Mp and a 2 x largers sensor , something like 50mm x 70mm, and THAT in digital would cost more than two kidneys. My only option would be to bring my gimbal, calibrated "nodal" point, and stitch 3 rows of 3 exposures, hopping that elements in the landscape didn't move too much. I now have no doubt why Ansel Adam used a large format camera, although it was heavy piece of equipment to carry on his hikes.

[Serkevan]

Pareto principle applies more or less everywhere. You know the drill, 20% of the effort for 80% of the performance. Want 25% more? Pay five times more.

-R&D costs of cameras with FF and MF sensors go bananas partly because of the extra processing power needed, but not only that: shutters and mirrors have to be larger and heavier, increasing the mechanical complexity. While the overpricing applies mostly to digital, a good Pentax 67 still sells for a bunch of money now, used.
-Lenses have to be substantially larger. Well-performing glass is expensive, more so if the image circle they cover is twice as large as APS-C. My Tamron 17-50 barely covers anything beyond the corners of the APS-C crop.
-Inflation is a thing because our beloved economic gurus insist that rampant consumerism is the only way forward (towards a cliff but yay). A 2012 camera that cost $2000 is equivalent to $2236 today, which isn't *that* much but it's still over 10%.

That said, APS-C is definitely the best value for money: economy of scale helps, sizes are large enough that costs related to miniaturization are kept in check and quality is still well ahead of typical displays (namely screens under 8K resolution).

---------- Post added 10-17-19 at 08:56 AM ----------

I can't print it larger than about 24 x 36" otherwise it looks fuzzy even at normal viewing distance.

I have a wallpaper, downsampled to 1920x1200p, on my... 32 x 52 cm (13 x 20") monitor and it looks swell. It was taken with the K-7. The K-1, particularly with Pixel-Shift, should be more than capable of printing even larger than twice this...

[kaseki]

The cost of making larger lenses and larger sensors does increase linearly with the size. ...

To be a bit more exact, production cost of scaling up a lens assembly goes roughly as the cube of the relative change in dimensions for materials, and as the square for glass polishing and coating, assuming that the quantity is large enough to justify the larger sized equipment needed. Grinding might lie between the square and the cube. If the back focal length has to scale differently than the dimension scaling, then that may change the properties of lenses and modify this dimensional rule. Moving part inertia scales as the cube and friction as the square so auto focus needs proportionately more power for the same speed. This affects the camera body design.

[bdery]

To be a bit more exact, production cost of scaling up a lens assembly goes roughly as the cube of the relative change in dimensions for materials, and as the square for glass polishing and coating, assuming that the quantity is large enough to justify the larger sized equipment needed. Grinding might lie between the square and the cube. If the back focal length has to scale differently than the dimension scaling, then that may change the properties of lenses and modify this dimensional rule. Moving part inertia scales as the cube and friction as the square so auto focus needs proportionately more power for the same speed. This affects the camera body design.

My original post was missing a "not" between "does" and "increase"...

when you increase the surface area of a lens, you also increase its thickness, which scales up the cost of the blank. Also, higher-resolution sensors call for higher-quality glass, so the design elements become more complex.

[biz-engineer]

Yet, no one pointed out the price of a large format camera in today's money compared the same system made of digital tech. If I remember correctly, even for a niche market, the price of a large format film camera system is ten times less than the price of the same size system with a CMOS sensor in it. IMO, it's just that a large format film cap may equate 250Mpixels worth of detail recorded on film plate during the exact time of the exposure time, while reading out and processing data from a 250Mp sensor into a CPU will take a dozen seconds on top of the exposure time. How much is the cost of a 4x5 film sheet covered with a silver emulsion compared to 4x5 CMOS sensor with 250Mpixels all working good. I can only conclude that digital tech for cameras is well suited for small formats, and far from being cost effective for recording light over large area.

[Serkevan]

A large format film camera system does not require readout speeds in the Gb/s to work, so of course with tech at the current level film will be an order of magnitude cheaper (for the body). But you have to remember that digital FF took five or six years to be viable compared to crop sensor cameras. Give it another decade and we'll have MF available at reasonable prices... for the bodies at least

[BigMackCam]

I have a photograph, taken woth the K1 on tripod, ISO200, a bit of shadow raised, but I can't print it larger than about 24 x 36" otherwise it looks fuzzy even at normal viewing distance. For the detail that's in that landscape photo, printed 32 x 48" , I'd need around 140Mp and a 2 x largers sensor , something like 50mm x 70mm, and THAT in digital would cost more than two kidneys. My only option would be to bring my gimbal, calibrated "nodal" point, and stitch 3 rows of 3 exposures, hopping that elements in the landscape didn't move too much. I now have no doubt why Ansel Adam used a large format camera, although it was heavy piece of equipment to carry on his hikes.

Whilst I realise your thread isn't about this in isolation, I wonder what you mean by "fuzzy" and "normal viewing distance"? I ask because these can be pretty subjective things, and folks have been printing larger than this with far fewer megapixels for years, with acceptable (even good) results. I also wonder if and how you enlarge your photos, and sharpen the final output, to suit the reproduction size and medium? I'll admit, I'm not asking from a position of experience and authority here, as I very rarely have my photos printed, and I've never had anything printed at the dimensions you're mentioning. But it seems odd that your 36MP landscape shot can't be processed to look good at those dimensions, assuming a realistic viewing distance for such a print