How Much Resolution Do You Want or Need? Read OP before posting.

Killerbob

Quite interesting article.

http://www.themarketbusiness.com/2015-09-08-canon-showcases-its-250-megapixel-sensor-in-the-extremely-high-resolution-camera

That's actually quite an interesting development, and they talk about this new fabled sensor here as well; http://www.bbc.com/news/technology-34173423

Have a read and wish...

spraynpray

For me, the most interesting part of the statement:


However, he said, improvements in other elements of an imaging system made large sensors unnecessary.
"You do not need such big sensors, because you could use smaller ones and just have better optics," he said.
"Simply increasing the number of pixels is not the only way forward."

WestEndFoto

For me, the most interesting part of the statement:

However, he said, improvements in other elements of an imaging system made large sensors unnecessary.
"You do not need such big sensors, because you could use smaller ones and just have better optics," he said.
"Simply increasing the number of pixels is not the only way forward."



In my view, the author of this article is clueless about camera systems. No FX system lens is ever going to benefit from 250 megapixels and probably not much more than a hundred. What is the difference between the 85 1.4 D and G at f/5.6? Practically nothing. At 1.4? OK, there is an improvement, but I doubt my D800 would be better than a D750. If lens manufacturers, and they are all about the same, have not been able to improve more than that in the last 20 years, I doubt you will see anything other than similar improvements in the next 20.

Spraynpray quoted the author saying, "You do not need such big sensors, because you could use smaller ones and just have better optics." Somehow he imagines optics being like computers, increasing exponentially every year. Lens designers are lucky to achieve 1% a year.

This author is knowledgeable about computers, not optics. Optics is everything. Everything else, including the sensor, is an accessory.

If he was honest and knowledgable, he would have pointed out that just like Nokia's 40 megapixel marketing gimmick sensor (anyone remember that?), there is no path forward with this.

I am afraid that if you want more resolution in photography, lots more not mere increments, then a larger format is the only answer. If you want it at a reasonable price, you will need to wait until medium format sensors come down in price.

There is no free lunch.

donaldejose

As to increased lens resolution, Sigma seem to be demonstrating that it can be done at reasonable prices with their Art series. We all know how sharp the 35 and 50 mm Art lenses are. Even the new Art zoom is very sharp.
http://slrgear.com/reviews/showproduct.php?product=1804

WestEndFoto

As to increased lens resolution, Sigma seem to be demonstrating that it can be done at reasonable prices with their Art series. We all know how sharp the 35 and 50 mm Art lenses are. Even the new Art zoom is very sharp.
http://slrgear.com/reviews/showproduct.php?product=1804

Sure, and I acknowledge that in the other thread about 1.4 primes. But the Sigma upgrade is like upgrading from 24 to 30 or maybe 36 megapixels. And those lenses are massive and represent 20 years of grinding incremental improvements in lens design. Further improvements of the same small magnitude will not come so easily. With this Canon sensor, we are talking about upgrading to 250 megapixels. Anybody that claims, "This simply requires further upgrades in lens resolution (assume FX)." is clueless about lens design. They even implied that it is possible with a smaller format, which will compound the challenge.

rmp

Need: 36 mp (OK, 90% of the time about 4KP)
Want: More, many more.

WestEndFoto

When it gets down to it, all we need is food, shelter and clothing. Everything else is a want and the notion that there are "needs" in this category worries me that society has become entitled, decadent, or both. Given my views, I find a debate about wants vs needs in camera gear rather silly, unless it is in the context of a professional earning money.

haroldp

Optics (and optical designs) that are much sharper are possible and available, they are used for industrial and military applications. They are just hideously expensive.

Leica introduced the 'APO Summicron' two years ago, which is corrected to industrial standards. It is amazing, and hideously expensive at USD $8,000 for a manual focus 50mm F2. It is also back ordered 3 years.

It's biggest weakness is that a 24mp leica sensor looks almost as good with the non APO 50mm F2 Summicron (this is my go to leica lens for 25 years, now sitting on a Sony A7II) at 1/4 the price.

Larger scale production using CNC techniques could drop these prices considerably.

Such lenses on a high res (perhaps 72mp) FX body could be an economical alternative to MF.

While larger sensors can benefit proportionately from technology improvements (remember 16mp Hassies), for most photog's, even pro's, once IQ passes the threshhold of 'good enough for the client' , other considerations take over and as smaller formats get better, the MF market shrinks.

.... H

WestEndFoto

Optics (and optical designs) that are much sharper are possible and available, they are used for industrial and military applications. They are just hideously expensive.

Leica introduced the 'APO Summicron' two years ago, which is corrected to industrial standards. It is amazing, and hideously expensive at USD $8,000 for a manual focus 50mm F2. It is also back ordered 3 years.

It's biggest weakness is that a 24mp leica sensor looks almost as good with the non APO 50mm F2 Summicron (this is my go to leica lens for 25 years, now sitting on a Sony A7II) at 1/4 the price.

Larger scale production using CNC techniques could drop these prices considerably.

Such lenses on a high res (perhaps 72mp) FX body could be an economical alternative to MF.

While larger sensors can benefit proportionately from technology improvements (remember 16mp Hassies), for most photog's, even pro's, once IQ passes the threshhold of 'good enough for the client' , other considerations take over and as smaller formats get better, the MF market shrinks.

.... H

I agree with all of that. I think that there is both a “theoretical cost no object” limit and a practical “economics engineering limit”.

The “theoretical cost no object” limit is determined by diffraction. At f/20.0, most cameras sensors have exceeded this limit. At f/1.4, hundreds of megapixels might be required before a sensor exceeds the ability of the lens. There are online resources that can be found to calculate this limit that I have played with previously. I don’t have the time at the moment to dig them up.

The “economics engineering limit” is very well illustrated by your Leica example HaroldP. However, I will point out that 72 megapixels is a long way from the 250 megapixel Canon example, and we are already up to $8,000. Also, because of diffraction, your choice in f-stops will likely be limited to wide open, so it won’t be good for landscapes. A budget priced Medium or Large Format lens can probably be made for less than this that would exceed the resolution as the optics only need to be decent, not cutting edge. The sensor, however, is another matter. But sensor prices are coming down, so maybe it is only a matter of time.

My point is that in time, the ability to image the demodex peeking out from behind a model’s eyelashes may be cheaper to achieve with medium format than FX. While I also agree that “good enough” is a huge factor, there will always be people that want to image those demodex. Like me!

haroldp

@WestEndFoto

We are in 'violent agreement', your points are well taken.

I think fashion will be the last bastion of MF since very high IQ is needed, and the cost of the equipment is the cheapest part of the shoot, and inconsequential.

For landscapes and anything else that does not move too quickly (except in California), and DOF is needed, the tiled panorama is the moral equivalent of large format (it actually is a large sensor), and is already the dominant format for serious landscape shooters. I can get hundreds of mp, all of which are shot with the center of a very sharp lens, no MF can match this.

As to the 'good enough' phenomnon, in the last year, the Nikon I have seen most frequently used by pros at weddings is the D300. Flash predominates, so ISO does not matter, Nikons DX lenses are sharp, and 12mp is enough to keep the customers happy.

A limit I do not hear too much about because we are not getting close to it yet (maybe military apps are) is that at about .7 micron pixel pitch, we approach the half wave of the longest visible light, ant the sensor may become a 'filter' into the visible spectrum.

Meanwhile my equipment is already better than I am.

This will all be fun.

WestEndFoto

@WestEndFoto

We are in 'violent agreement', your points are well taken.

I think fashion will be the last bastion of MF since very high IQ is needed, and the cost of the equipment is the cheapest part of the shoot, and inconsequential.

For landscapes and anything else that does not move too quickly (except in California), and DOF is needed, the tiled panorama is the moral equivalent of large format (it actually is a large sensor), and is already the dominant format for serious landscape shooters. I can get hundreds of mp, all of which are shot with the center of a very sharp lens, no MF can match this.

As to the 'good enough' phenomnon, in the last year, the Nikon I have seen most frequently used by pros at weddings is the D300. Flash predominates, so ISO does not matter, Nikons DX lenses are sharp, and 12mp is enough to keep the customers happy.

A limit I do not hear too much about because we are not getting close to it yet (maybe military apps are) is that at about .7 micron pixel pitch, we approach the half wave of the longest visible light, ant the sensor may become a 'filter' into the visible spectrum.

Meanwhile my equipment is already better than I am.

This will all be fun.

Hear hear!

Ironheart

A limit I do not hear too much about because we are not getting close to it yet (maybe military apps are) is that at about .7 micron pixel pitch, we approach the half wave of the longest visible light, ant the sensor may become a 'filter' into the visible spectrum.

You will have to get above 2 gigapixels (2000 mp) on a FX before this is an issue. I think we have a little ways to go, even with military applications.

donaldejose

Perhaps we should think about "reasonable cost high resolution." For example, take reasonable cost lenses like the 85mm, and 105 mm Nikkors and the Sigma 35, 50mm Art lenses. They are very sharp. Now put those lenses on a new Nikon reasonable cost body containing a 50+ megapixel sensor. Do you achieve a desirable and noticeable increase in image sharpness? If so, you have a "reasonable cost high resolution" camera system. We should soon have the answer to this because I expect Nikon to put out a 50+ megapixel sensor (hopefully in a DXXX style body) before a year is up.

haroldp

At each generation of fabricators, the cost of a semiconductor ( like a sensor) is mostly a function of it's size, not it's pixel density. How many can you get out of a wafer is the key economic question.

As they develop, a 50 (or 100mp or 200mp) sensor may cost no more than a 24mp sensor of the same size.

Optics being mechanical will always be a major cost and limiting factor.

... H

haroldp

You will have to get above 2 gigapixels (2000 mp) on a FX before this is an issue. I think we have a little ways to go, even with military applications.

Small sensors like cellphone cameras will likely push these limits first. I believe some are close to a micron pp now.

spraynpray

http://www.dpreview.com/articles/2389085425/canon-plans-high-res-future-with-120mp-dslr-and-8k-cinema-eos-in-development

donaldejose

Apparently Canon thinks the lenses will be able to handle 120 mp sensors.

WestEndFoto

Apparently Canon thinks the lenses will be able to handle 120 mp sensors.

I am pretty skeptical. Nikon thinks that you can shoot at ISO 400,000 too, but nobody does that.

However, I would love to be proven wrong, because if Canon can do it, Nikon can do it.

One thing to keep in mind, a lens that can resolve 120 megapixels is only going to be able to do it on a wide aperture. As an example, the 28 in my signature is diffraction limited at f/5.6 (stopped down, resolution decreases because of diffraction). Therefore, if they increase the resolution past "x", then you will only see the increased resolution when it is opened up past f/5.6.

Now the questions is what is "x". If you go to the following link, you can probably figure it out:

http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm#calculator

I am going to spend some time fooling around with this on the weekend.

Paperman

An old article which summarizes diffraction limits on a table ...You are safe up to 1885 Mp ( at f1.0 ! ) if you own a $10k f0.95 Leica ... :-* . If you are a landscape photographer, better not ask for more than 60 Mp ...

https://luminous-landscape.com/do-sensors-out-resolve-lenses/

donaldejose

Looks like there is still room for higher mega pixel sensors and better lenses. The weakest link seems to be in red colors.

haroldp

The weakest link is my skill, but I still want better iron.

... H

haroldp

If the announced new Nikon 1 is 22mp in a 1" sensor, that is the pixel density of about 160mp in an FX sensor.

Approx a 2 micron PP.

... H

donaldejose

I have been expecting Nikon's next high pixel sensor to be about 56 mp. However, they now have the technology to produce a 100+ mp sensor if they have the processing power to handle it.

WestEndFoto

An old article which summarizes diffraction limits on a table ...You are safe up to 1885 Mp ( at f1.0 ! ) if you own a $10k f0.95 Leica ... :-* . If you are a landscape photographer, better not ask for more than 60 Mp ...

https://luminous-landscape.com/do-sensors-out-resolve-lenses/

This is interesting. It basically says that my 36 megapixel D800 out resolves ANY lens at greater than f/5.6ish due to diffraction. Probably my only lenses in my signature that take full advantage of the resolution on my D800 (and only if I am shooting at f/5.6) are my 28 2.8 AIS, 85mm 1.4G and maybe my 135 DC 2.0 and 200 f/4.0.

My gut based on experience of shooting says this might be a little on the low side, but it sounds like it is in the ballpark.

It also means that if Nikon does make a higher resolution sensor (>36 megapixel) and you intend on taking advantage of it, you better be shooting at wider than f/5.6.

And if you are shooting DX, you can basically divide the 36 megapixel limit (I think it is a little higher, but for illustration purposes, lets go with this) by the ratio between the area of DX and FX. Basically, then, with a DX sensor, the lens will need to be diffraction limited at less than f/5.6 to take advantage of anything more than 18 megapixels. If you are advocating DX because of its pixel density, you better be using a top end lens.

And if you are shooting micro four thirds, anything more than about 10 megapixels is useless.

Sigh........I am looking forward to medium format coming down in price so I can get my full body shots with crisp eyelashes. Or perhaps Nikon needs to produce some fabulous lenses and I will need to shell out for them.

spraynpray

Erm, Dx - smaller sensor, but higher pixel density - effect of all that on diffraction - needs some thinking about. I regularly use my D7100 down to f13 and occasionally f16 for my macro shots and never had a problem.