Sony to unveil a 50 MP new A7 body at Photokina?

[RLPhoto]

That's all fine and dandy but in the Real world, hardly of any use. Only a handful of lenses actually can resolve the details of a 36mp sensor. Forget about a 50mp sensor on that Sony glass.

That's not how resolution works. The final value is a result of the combination of all components (lens and sensor or film) and is always lower than the weakest component. But strengthening any component leads to an increase for the entire system, even if that specific component is "past" the resolution of the weakest component, i.e. the weakest component is not a hard limit in the way you're imagining it to be.

At any rate, you can discern plenty of lenses on 24 MP APS-C sensors which means they will show an advantage on 50 MP FF.

That said, it would be difficult to see any difference between 24/36/50 MP except on the very largest prints. We're hitting the point of diminishing returns with respect to even 36" prints.

Eh. Not so in practice, you just getting more crappy detail of the crappy part of the lens by adding more MP. However moving to a bigger format with more MP, now that adds real detail gains.

 

[jrista]

That's all fine and dandy but in the Real world, hardly of any use. Only a handful of lenses actually can resolve the details of a 36mp sensor. Forget about a 50mp sensor on that Sony glass.

That's not how resolution works. The final value is a result of the combination of all components (lens and sensor or film) and is always lower than the weakest component. But strengthening any component leads to an increase for the entire system, even if that specific component is "past" the resolution of the weakest component, i.e. the weakest component is not a hard limit in the way you're imagining it to be.

At any rate, you can discern plenty of lenses on 24 MP APS-C sensors which means they will show an advantage on 50 MP FF.

That said, it would be difficult to see any difference between 24/36/50 MP except on the very largest prints. We're hitting the point of diminishing returns with respect to even 36" prints.

Eh. Not so in practice, you just getting more crappy detail of the crappy part of the lens by adding more MP. However moving to a bigger format with more MP, now that adds real detail gains.

I wouldn't call it "crappy" detail. The sensor simply begins to oversample the lens when the lens is diffraction limited. When the lens is not diffraction limited, and aberrations are smaller than a diffraction spot, then even at 50mp, we are still a LONG way from oversampling a lens at f/4, even f/5.

I would rather have the sensor oversample the lens, and be able to LEGITIMATELY do away with an AA filter, than always be undersampling. Once we finally achieve that, then the majority of sensor-frequency noise reduction will affect SUB-detail frequencies. We could pretty much eliminate noise, downsample by a factor of two, and still have ultra high resolution images that are sharper and clearer than anything we have today.

 

[surapon]

Surapon, that actually looks like the perfect analogy of a 50MP Sony camera ;D Bravo!

Ha, Ha, Ha, Dear Mr. fragilesi
That is a great Idea from our friend Mr. c.d.embrey, I just find old photos to show some of our friend ( Who live in the big city around the world) and do not know What Mr. c.d. embrey talk about. Yes, I live in the Boone-dock and see this kind of truck all the times, But not Pretty like this one.
Thanks again
Surapon

 

[unfocused]

I doubt if this rumor is true. But, it is enlightening that so many people believe it could be.

It just adds to the impression that Sony is throwing every possible piece of crap on the wall trying to make something stick.

Some people mistake that for innovation. But really, it's just a company that is in third-place and way way back and is desperately trying to capture a little more market share before their creditors and stockholders tell them to hang it up.

 

[moreorless]

For one, it could pretty much eliminate the reach gap between APS-C parts and FF, assuming you could maintain a high frame rate (and we know that's possible...Canon achieved 9.5fps at 120mp.) You could crop any part of a 50mp frame, and have the same kind of reach as a 20-24mp APS-C camera.

True and I spose you can argue an advantage for action shooting both in cropping from a larger area and that the AF points can cover a larger part of the crop. I don't see this being a big benefit from the FE system though given its AF performance, handling and lack of tele options though.

Don't get me wrong I certainly wouldn't say no to 50+ MP on an replacement for my D800 when I need one but even as mostly a landscape shooter who often prints large(A2 and up) its not something that would have be rushing out to buy a new camera.

 

[preppyak]

Proof, please.

http://www.lensrentals.com/blog/2012/03/d800-lens-selection
http://www.lensrentals.com/blog/2012/03/d-resolution-tests

D800 is capable, with the very best lenses, of bringing more resolution to an image. But, it is only with the best (1050+ lines average) lenses that it makes the difference between a 36MP and 22MP.

And that is of course working off a tripod in lab perfect conditions. In more practical conditions (especially handheld), you'd see even less gain with anything but the best lenses. Basically, if you want the most of the D810/D800E, you want the Zeiss lenses.

Of course, this isnt a knock against a 36MP sensor. Its capable of more in some circumstances, and in every other one, its on par with a smaller sensor. But, if you're considering a 50MP sensor (and the hassles that come with that), its good to know what you are actually gaining

 

[jrista]

I doubt if this rumor is true. But, it is enlightening that so many people believe it could be.

It just adds to the impression that Sony is throwing every possible piece of crap on the wall trying to make something stick.

Some people mistake that for innovation. But really, it's just a company that is in third-place and way way back and is desperately trying to capture a little more market share before their creditors and stockholders tell them to hang it up.

LOL. I find it interesting that you automatically think it would be crap. I don't necessarily think the Sony CAMERA that included this sensor would be great...and I do think that Sony has thrown out too many products lately for all of them to stick. However, on the sensor front...I think a 50mp FF sensor with high DR would be awesome.

I think we also have to figure that at some point, Sony is going to get all the factors right. Their bodies are not great today...but if they respond to customer feedback and fix the issues with them, they could become competitive with Nikon and Canon bodies. They already are somewhat competitive, just because of the sensor. I am not a big Sony fan, I don't generally like how they do things...but I'm seriously considering an A7r. I don't need a fancy body packed with perfect ergonomics and features. I just need a camera with a big huge sensor and tons of pixels and DR that I can just plop onto a tripod. The rest is all manual...the features don't matter so long as the data file contains the information I want.

 

[preppyak]

LOL. I find it interesting that you automatically think it would be crap. I don't necessarily think the Sony CAMERA that included this sensor would be great...and I do think that Sony has thrown out too many products lately for all of them to stick. However, on the sensor front...I think a 50mp FF sensor with high DR would be awesome.

Not to mention, if Sony had the hackers Canon does with Magic Lantern, a 50MP sensor with their DR means you could have a video camera that is nearly capable of the resolution and DR of the Red Dragon sensor. Hell, even managing to hack 4K out of that camera would be a big step forward

As I said on the first page, Sony introducing a new camera wouldnt surprise me. They've released as many cameras as lenses in the last 4 years for E-Mount. I'd be more surprised if they said they were releasing 5 lenses instead.

 

[RLPhoto]

Proof, please.

http://www.lensrentals.com/blog/2012/03/d800-lens-selection
http://www.lensrentals.com/blog/2012/03/d-resolution-tests

D800 is capable, with the very best lenses, of bringing more resolution to an image. But, it is only with the best (1050+ lines average) lenses that it makes the difference between a 36MP and 22MP.

And that is of course working off a tripod in lab perfect conditions. In more practical conditions (especially handheld), you'd see even less gain with anything but the best lenses. Basically, if you want the most of the D810/D800E, you want the Zeiss lenses.

Of course, this isnt a knock against a 36MP sensor. Its capable of more in some circumstances, and in every other one, its on par with a smaller sensor. But, if you're considering a 50MP sensor (and the hassles that come with that), its good to know what you are actually gaining

Basically what I said. The benefit between 22 and 36mp bodies is negligible when using let's say the Tamron 24-70mm. Once you get the zeiss 135mm APO, the difference is stark.

However, doesn't change my point that even at 36mp it's hard to get the detail out of it and let alone a 50mp sensor in 35mm format. MF cams could get the majority of that detail out due to the lower lens magnification.

And when you put it in print, the difference is even more eh...

 

[unfocused]

I doubt if this rumor is true. But, it is enlightening that so many people believe it could be.

It just adds to the impression that Sony is throwing every possible piece of crap on the wall trying to make something stick.

Some people mistake that for innovation. But really, it's just a company that is in third-place and way way back and is desperately trying to capture a little more market share before their creditors and stockholders tell them to hang it up.

LOL. I find it interesting that you automatically think it would be crap...

It's a colloquialism. Not meant to be taken literally. The saying refers to throwing up a barrage of things in the hopes that one or more is successful or "sticks." Of course, "crap" being an internet-safe term for the real saying.

 

[Lawliet]

But, it is only with the best (1050+ lines average) lenses that it makes the difference between a 36MP and 22MP.

The joy of quoting numbers without understanding them...

Those tests simply measure when contrast falls under a certain threshold, but that doesn't mean the lens can't resolve structure in hair, skin or fabric, just that the contrast is slightly lower. Nice to compare relative performance, worthless for the question at hand. Well, actually less then worthless as it leads to wrong conclusions.
(And then came frequency seperation based techniques and made the reduced high frequency contrast an even more moot point.)

To find out whether a lens can benefit from higher sensor resolution one has to look at the different types of aberrations and their respective cumulative effects - a much more complex endeavor.

(Also I don't think that the ability of my strobes to freeze motion depends on the support of the camera. Although the shutter plays a role because of the shorter flash duration required to balance against the ambient.)

 

[awinphoto]

Oh Dear Lord, it's time to put all of my gear on ebay or better yet the canon rumors selling thingy and start stocking up on Sony gear... because we know all their lenses are top notch, and their quality control, ergonomics and functionality is all completely top notch! Also let me quickly sell my new imac and get that nuclear warhead (new mac pro) plus a couple thousand of dollars in back up hard drive space just to handle this sexy new sony camera... All this because I KNOW that all my portrait clients are clamoring to come to me because i have this camera and my competition doesn't and they all have space and a narcissistic desire to have a 30x45 or larger wall shot of their big fat mug on their walls. Go big or go home I always say!...

 

[RGF]

50 MP, 16 stops of DR. Canon must respond to that!!

:

Canon rumors need to respond to that.

New Canon 3D with 42 MP and 14 stops of DR, 10 FPS, and 4K videos.

 

[jrista]

For one, it could pretty much eliminate the reach gap between APS-C parts and FF, assuming you could maintain a high frame rate (and we know that's possible...Canon achieved 9.5fps at 120mp.) You could crop any part of a 50mp frame, and have the same kind of reach as a 20-24mp APS-C camera.

True and I spose you can argue an advantage for action shooting both in cropping from a larger area and that the AF points can cover a larger part of the crop. I don't see this being a big benefit from the FE system though given its AF performance, handling and lack of tele options though.

Don't get me wrong I certainly wouldn't say no to 50+ MP on an replacement for my D800 when I need one but even as mostly a landscape shooter who often prints large(A2 and up) its not something that would have be rushing out to buy a new camera.

Regarding the bolded...sure...but, you already have the D800. If your stuck below 25mp...a jump to 50mp is really, really quite nice.

 

[mkabi]

But its maybe because your shooting habits differ from mine and you don't have to have that safety net.

2 TB drives cost 70 euro, that is less what i have spend in a month for film.
harddrive space should not be a problem, especially not for a working pro.

i backup to 4 hard drives for each of the 3 storage drives in my PC (automatically to a NAS and a second PC).
then i have a sharkoon docking station to do extra backups to harddisks i store offline.

harddrive cost is nothing i worry about... as i have not worried about film cost.

processing power... well when i do the surface blur filter i will sure hate a 50 MP file.
but then, i keep my systems up to date.

but of course when someone spends 1000$ on a tripod, or 2000$ on a lens but still works with a 3 year old celeron, he should think about upgrading his PC too.
at least when he wants to edit 30+ MP files.

and let´s not forget, MF owners have to deal with such kind of data for a while.

Were not talking about the same thing you are working with consumer grade hardware, 3-4 drives of 2TB, I'm talking back ups of 15TB per year over a 5 year span not 15TB total, you don't just add another drive as and when you need it you have to have some sort of contingency and then double it for a safe backup. Im talking rackable server grade systems.

http://store.apple.com/uk/product/HE155ZM/A/promise-pegasus2-r8-32tb-8-x-4tb-thunderbolt-2-raid-system?fnode=1597ea08b4818b25cb21c5ba72b87b4d0d6ddab3e2699ec45f6da6e7ee44455736582815946e8c9747dae63d034ecfcc00256cd81ca2e3b77144751317d06785ac75fb316a0d7ed7f06da7bc61f7e4ed5d875e4aa144aa7b3973eab81e40b0fbc1a860980f4573a959f0adbece8b7cc6

Thats £7000 worth of back up storage on its own, to me thats not a small amount of money and if you overlooking this respect is a huge mistake and can ruin a photographers reputation if the worst should happen. Add a decent machine and 2 high quality colour correct monitors and you are easily looking at £15000 for the set up. Pros need proper storage solutions that large data is accessible quickly without a lot of small little raid drives knocking around, or spreading data across an old PC or random drives. These are very expensive, when your dealing with someones wedding and memories you need to ensure its safe incase of unforeseen events. If you don't take it seriously you shouldn't take on the responsibility.

That storage will only last 2 years with at 50mp with a commercial event and wedding photographer. With a camera with half the output your effectively doubling your timescale and saving money and still delivering incredible quality with the 5DMKIII.

Medium format isn't really suited to wedding work, the cameras are slow, heavy and expensive. Although not unheard of MF isn't overly popular in the professional event photography. The likelihood of a photographer shooting 1500 frames per wedding is fairly slim with MF. Medium format is much more suited to Landscape work and no chance they are shooting anywhere near that amount of data per shoot, its not just the data of one image import 1500 50mp files into lightroom and see how long it takes for it to create previews… good luck with that! Which is why full frame cameras are much more suited, smaller files, smaller system, high quality.

This is one of the reasons a lot of wedding photographers who picked up the D800 switched to the 5DMKIII after it came down in price, Data! It sounds ridiculous but it really is one of the biggest factors when you need to manage peoples events.

You're wasting your e-breath...
I made the same arguments in the past, some guy brought in some chart that google uses to monitor failure rate of their hard-drives (the same ones they use on their servers).

Fact of the matter is, when consumers want something... they really want something... like little babies crying at Wal-Mart b/c their parents didn't buy them chocolate... they will scream and shout... making a big spectacle of things. *face-palm*

Actually a better example would be a kid at a toy store screaming and shouting... and when the parent does get him/her that toy... they will probably play for it for a while, then throw it to the side and never play with it again. I guess some habits don't leave us... Don't get me wrong... I'm sometimes guilty of doing the same.

I'm one of the ones that want 4K, but I know I won't use it much... and the storage on that will be insane.

 

[jrista]

Were not talking about the same thing you are working with consumer grade hardware, 3-4 drives of 2TB, I'm talking back ups of 15TB per year over a 5 year span not 15TB total, you don't just add another drive as and when you need it you have to have some sort of contingency and then double it for a safe backup. Im talking rackable server grade systems.

I think this is really becoming an archaic way of looking at the storage problem. No one buys that much PERSONAL storage space. It really isn't necessary anymore, not with secure cloud storage becoming so cheap.

Look at Amazon Glacier. If you need MASSIVE long-term data archival, something to the tune of 10+ terrabytes a year, put it in Glacier. It's all SSL encrypted transport, with 256-bit AES encryption on the data itself. For long term, totally secure and encrypted data storage, it costs practically nothing (relatively speaking) to actually put stuff into Glacier. The costs only really start to mount IF you need to pull something out.

For massive bulk data storage, it no longer makes any sense to buy your own hard drives. Especially not if your planning to never delete anything, thus literally requiring 15TB a year to store all the data.

You could always keep a NAS at home with 4 to 6 drives and some 16TB of storage space or so, which is actually relatively cheap. You can keep active projects along with "potentially active for some time" or "might be needed quickly" kinds of projects, and swap chunks of data into Glacier or some other cloud storage option as you find there is no longer a need for instantaneous access. (And there is simply no logical argument for needing instantaneous access to years and eventually decades of high resolution photography...at some point, ALL data becomes archival.)

If you still want to make sure you have access to certain projects more quickly than you might be able to get them off of something like Amazon Glacier, you can still always store some of that data on bluray disc. Eventually everything would end up on Glacier, but if you have certain customers that you know will probably be back for changes or whatever within a few months period, or even a year period, you can temporarily keep that stuff on bluray for quicker restoration.

There are other options out there for cloud storage as well. You could augment (or even replace) something like Glacier with more "real time" storage. I've seen a terrabyte of cloud storage for as little as $20 a year, which is about $1.67 a month. You could have 15TB of cloud storage for $25 a month. To get 16TB of storage (four 4TB drives), that would run you around $480 (cheapest price I could find on Amazon after several pages was $120 for a single internal 4TB drive...that's the CHEAPEST I could find, on NewEgg the cheapest was $150). Now, if you want your 16TB of storage to be RELIABLE, were talking RAID. Since you can't stick set after set of 16TB raided drives into a computer, were also probably talking NAS. Reliable drives are also more costly...at the moment, quality 4TB drives are closer to $200. We probably want reliable, high speed drives designed to operate in a redundant array, in which case the minimum price is probably $250. At $250 a drive, that's $1000 just for the drives.

Finally, we'll want an actual NAS device for these drives. I personally use a ReadyNAS NVX, which utilizes X-RAID for LIVE drive swapping. You can replace a bad drive, or swap out a smaller drive for a larger drive, and the system will either restore onto the new drive or expand onto the new drive while still operating. You get about 0.7x the actual storage space (so, for 16TB, we would get 11.2TB of usable space, the rest is parity), meaning to really get 16TB of usable space, were talking about even larger drives, or using a 6-drive NAS device. Either way, the cost goes up, to around $1500 just for the drives. The cost of the NAS is around $600 for a 4-bay, and over $1000 for a 6-bay.

That brings the total cost of reliable, recoverable 16TB home RAID storage to as much as $2500 on the low end. That's a lot of money. A LOT of money. You would spend $300 a year on 15TB of cloud storage if you go with the cheaper stuff (like MediaFire Pro, the one that costs around $20 a year). I think MediaFire even has an enterprise deal for $20 or $30 a month, and it offers a LOT more than 15TB of storage space. That may even be cheaper than Amazon Glacier (although, possibly not as reliable.) At those prices, you would break even on the cost of your ONE personal NAS setup in about seven years.

I spent $1200 on my older ReadyNAS NVX and four hard drives years ago, and back then that was still quite a lot of money (but, it was also before the cheap cloud storage age.) I honestly don't see any reason why storage has to be expensive, even if you have 90mb RAW files. It just doesn't make sense to fill your personal living space up with countless hard drives, or some kind of personal SAN. It's not the cheapest solution.

I think the problem of storage space for high resolution RAW images is really a simple problem. Just stick it in the cloud. It's secure, reliable, and super cheap.

 

[Canonicon]

Were not talking about the same thing you are working with consumer grade hardware, 3-4 drives of 2TB, I'm talking back ups of 15TB per year over a 5 year span not 15TB total, you don't just add another drive as and when you need it you have to have some sort of contingency and then double it for a safe backup. Im talking rackable server grade systems.

I think this is really becoming an archaic way of looking at the storage problem. No one buys that much PERSONAL storage space. It really isn't necessary anymore, not with secure cloud storage becoming so cheap.

Look at Amazon Glacier. If you need MASSIVE long-term data archival, something to the tune of 10+ terrabytes a year, put it in Glacier. It's all SSL encrypted transport, with 256-bit AES encryption on the data itself. For long term, totally secure and encrypted data storage, it costs practically nothing (relatively speaking) to actually put stuff into Glacier. The costs only really start to mount IF you need to pull something out.

For massive bulk data storage, it no longer makes any sense to buy your own hard drives. Especially not if your planning to never delete anything, thus literally requiring 15TB a year to store all the data.

You could always keep a NAS at home with 4 to 6 drives and some 16TB of storage space or so, which is actually relatively cheap. You can keep active projects along with "potentially active for some time" or "might be needed quickly" kinds of projects, and swap chunks of data into Glacier or some other cloud storage option as you find there is no longer a need for instantaneous access. (And there is simply no logical argument for needing instantaneous access to years and eventually decades of high resolution photography...at some point, ALL data becomes archival.)

If you still want to make sure you have access to certain projects more quickly than you might be able to get them off of something like Amazon Glacier, you can still always store some of that data on bluray disc. Eventually everything would end up on Glacier, but if you have certain customers that you know will probably be back for changes or whatever within a few months period, or even a year period, you can temporarily keep that stuff on bluray for quicker restoration.

There are other options out there for cloud storage as well. You could augment (or even replace) something like Glacier with more "real time" storage. I've seen a terrabyte of cloud storage for as little as $20 a year, which is about $1.67 a month. You could have 15TB of cloud storage for $25 a month. To get 16TB of storage (four 4TB drives), that would run you around $480 (cheapest price I could find on Amazon after several pages was $120 for a single internal 4TB drive...that's the CHEAPEST I could find, on NewEgg the cheapest was $150). Now, if you want your 16TB of storage to be RELIABLE, were talking RAID. Since you can't stick set after set of 16TB raided drives into a computer, were also probably talking NAS. Reliable drives are also more costly...at the moment, quality 4TB drives are closer to $200. We probably want reliable, high speed drives designed to operate in a redundant array, in which case the minimum price is probably $250. At $250 a drive, that's $1000 just for the drives.

Finally, we'll want an actual NAS device for these drives. I personally use a ReadyNAS NVX, which utilizes X-RAID for LIVE drive swapping. You can replace a bad drive, or swap out a smaller drive for a larger drive, and the system will either restore onto the new drive or expand onto the new drive while still operating. You get about 0.7x the actual storage space (so, for 16TB, we would get 11.2TB of usable space, the rest is parity), meaning to really get 16TB of usable space, were talking about even larger drives, or using a 6-drive NAS device. Either way, the cost goes up, to around $1500 just for the drives. The cost of the NAS is around $600 for a 4-bay, and over $1000 for a 6-bay.

That brings the total cost of reliable, recoverable 16TB home RAID storage to as much as $2500 on the low end. That's a lot of money. A LOT of money. You would spend $300 a year on 15TB of cloud storage if you go with the cheaper stuff (like MediaFire Pro, the one that costs around $20 a year). I think MediaFire even has an enterprise deal for $20 or $30 a month, and it offers a LOT more than 15TB of storage space. That may even be cheaper than Amazon Glacier (although, possibly not as reliable.) At those prices, you would break even on the cost of your ONE personal NAS setup in about seven years.

I spent $1200 on my older ReadyNAS NVX and four hard drives years ago, and back then that was still quite a lot of money (but, it was also before the cheap cloud storage age.) I honestly don't see any reason why storage has to be expensive, even if you have 90mb RAW files. It just doesn't make sense to fill your personal living space up with countless hard drives, or some kind of personal SAN. It's not the cheapest solution.

I think the problem of storage space for high resolution RAW images is really a simple problem. Just stick it in the cloud. It's secure, reliable, and super cheap.

You miss that upping RAW filles needs a lot of bandwith.

Here in Europe where i live 60KB/s upload speed is the norm outside big cities.
That´s DSL6000 upload speed.

 

[Lawliet]

You miss that upping RAW filles needs a lot of bandwith.

Just send a HD to amazon, maybe once per month.
Seriously, thats part of the glacier service, it's designed for multi-TB of data on a regular base after all.

Here in Europe where i live 60KB/s upload speed is the norm outside big cities.
That´s DSL6000 upload speed.

LTE spreads fast, it's just a matter of time

 

[dtaylor]

You said it is always lower than the weakest component. So therefore, the weakest component has to be a hard limit on resolution. Perhaps increasing the resolution of another component causes the system resolution to asymptotically approach the weakest components resolution?

Essentially yes. By "hard limit" I meant that you can't say a 120 lpmm sensor is useless because of a 100 lpmm lens. You're never going to actually achieve 100 lpmm with the system, but pushing up the sensor resolution actually does get you closer even if the sensor is already >100.

I have the formula in a text book some where...I think ???

 

[Canonicon]

LTE spreads fast, it's just a matter of time

LOL your voice in the carriers ears.. im waiting since 2007 for faster internet.

It´s just one year that i was upgraded from DSL2000 to DSL6000.