Why is 4/3 "stuck" to 16mp when Sony 1" sensor is 20mp?

Why is 4/3 "stuck" to 16mp? The argument that it is not possible to get more real resolution falls flat: Sonys´1" sensor is 20mp and shows a great deal of resolution. Only the lens which cannot make the sharpness through out in the corners (RX100)
It isn't stuck. The MP will gradually increase for all sensor sizes up to the limit that can be tolerated for each sensor size using the existing sensor and in-camera processing technology.

The general pattern is that when your camera manufacturer of choice only makes a 10 MP camera owners of the camera say "I can't understand why any normal person needs more than 10 MP - I can make razor sharp 4 ft. x 3 ft. prints".

When the manufacturer goes to 16 MP these people then say "I can't understand why any normal person needs more than 16 MP".
Two major faults to the logic being applied : First, the " MP race" for the past decade or so since the transition to digital started from low mega-pixel digital cameras where they were truly inefficient to a point even human eyes can clearly spot faults on 6x4 prints to begin with, to which we have cameras like the D800s and A7Rs with 36 mp. Yes, technologies keep advancing everyday, but you are neglecting two simple facts that will not change over time regardless of technologies: Sensor sizes, a full framed sensor (35mm) will always be 36x24mm, an APS-c sensor will always be 23.5x15.6mm (canon is a tiny bit smaller), so on and forth for all the micro 4/3 and 1" sensors. No matter how many mega-pixels you are capable of packing into those sensors respectively, their sizes will never change like technologies itself does. The more you pack into the same sensor, the more signal-noise you get, you might be able to squeeze a bit less with advanced technologies, but you also can't change one fact: the more mp, the smaller individual pixel size also comes into play to hurt your overall performance of the sensor.

The second and foremost important fact you are missing out, if you are talking about prints from a 3mp camera from 10 years ago, people may not be satisfied I(Q wise, but as most have mentioned and experienced today, with 16 mp you can basically get razor-sharp wall-sized prints, ask yourself, even for professional uses, how often do you print bigger than that in real life besides some delusional theoretical desires by some pixel-peeping lovers typically live their lives on the forums ? Not to mention no matter how much more mp you squeeze into the sensor, our humanly eye sight will never grow like technologies does beyond its biological limitations to find faults of IQ at a certain point. ( Unless you hobby is to pixel-peep with micro-scopes, then no matter how technologies advances, it will never satisfy you.) The whole point is we have come to a point where "Enough is enough" both for practical everyday life usages and human eyes biological limitations have been met to find faults IQ wise among today's top performing cameras. Although greed will never be never satisfied for some, 16 mp isn't enough today, I'm sure 100 mp will not be enough for those neither even if we do get there at some point in the future.
 
Last edited:
The theoretical best linear difference in resolution between 20Mp and 16Mp is just 12%. Just 1/8 th more resolution.

That's not much at all.
And the linear difference between 16MP and 12MP is just 16%. So why did M43 bothered to move from 12MP in the first place?
 
So why did M43 bothered to move from 12MP in the first place?
I'd bet the engineers had no interest in moving to 16Mp from 12Mp and I'd say the only reason they did was because the marketing twits needed a headline number to promote a new camera.

People do the same thing with cars. They buy 200Hp hatchbacks when they need 75Hp. They buy cars with speed dials indicating speeds they'd be jailed for decades for driving at. People don't generally ask the question "what do I need ?". Most people just buy anything that sounds big.

And the marketing people know and exploit it.

And that's why the engineers keep being told to add more megapixels.
 
And Sony announces a 4k-capable camera...

Fine print: If you buy a $2000+ 4k recorder and lug it along.

At least they mention that detail in the press release. A lot of the time with Sony, you discover those things after you buy the product.

For most video users, Panasonic GH4 is likely to be better, cheaper, and hit the streets first.
 
Why is 4/3 "stuck" to 16mp? The argument that it is not possible to get more real resolution falls flat: Sonys´1" sensor is 20mp and shows a great deal of resolution. Only the lens which cannot make the sharpness through out in the corners (RX100)
Why do you think 'only' 16 mp is a handicap? Is the image quality any less? Not that I can see.

Do you know what you are buying when you shop for more mp? Few do, do you? You can make giant enlargements with 8 and 12 mb images. The technical need for more mp was surpassed several years ago

What does a few more pixels actually mean to YOUR photography? Why aren't you clamoring for quality lenses? Something that will always have some effect on your photography?
 
Why is 4/3 "stuck" to 16mp? The argument that it is not possible to get more real resolution falls flat: Sonys´1" sensor is 20mp and shows a great deal of resolution. Only the lens which cannot make the sharpness through out in the corners (RX100)
It isn't stuck. The MP will gradually increase for all sensor sizes up to the limit that can be tolerated for each sensor size using the existing sensor and in-camera processing technology.

The general pattern is that when your camera manufacturer of choice only makes a 10 MP camera owners of the camera say "I can't understand why any normal person needs more than 10 MP - I can make razor sharp 4 ft. x 3 ft. prints".

When the manufacturer goes to 16 MP these people then say "I can't understand why any normal person needs more than 16 MP".
Two major faults to the logic being applied : First, the " MP race" for the past decade or so since the transition to digital started from low mega-pixel digital cameras where they were truly inefficient to a point even human eyes can clearly spot faults on 6x4 prints to begin with, to which we have cameras like the D800s and A7Rs with 36 mp. Yes, technologies keep advancing everyday, but you are neglecting two simple facts that will not change over time regardless of technologies: Sensor sizes, a full framed sensor (35mm) will always be 36x24mm, an APS-c sensor will always be 23.5x15.6mm (canon is a tiny bit smaller), so on and forth for all the micro 4/3 and 1" sensors. No matter how many mega-pixels you are capable of packing into those sensors respectively, their sizes will never change like technologies itself does. The more you pack into the same sensor, the more signal-noise you get, you might be able to squeeze a bit less with advanced technologies, but you also can't change one fact: the more mp, the smaller individual pixel size also comes into play to hurt your overall performance of the sensor.

The second and foremost important fact you are missing out, if you are talking about prints from a 3mp camera from 10 years ago, people may not be satisfied I(Q wise, but as most have mentioned and experienced today, with 16 mp you can basically get razor-sharp wall-sized prints, ask yourself, even for professional uses, how often do you print bigger than that in real life besides some delusional theoretical desires by some pixel-peeping lovers typically live their lives on the forums ? Not to mention no matter how much more mp you squeeze into the sensor, our humanly eye sight will never grow like technologies does beyond its biological limitations to find faults of IQ at a certain point. ( Unless you hobby is to pixel-peep with micro-scopes, then no matter how technologies advances, it will never satisfy you.) The whole point is we have come to a point where "Enough is enough" both for practical everyday life usages and human eyes biological limitations have been met to find faults IQ wise among today's top performing cameras. Although greed will never be never satisfied for some, 16 mp isn't enough today, I'm sure 100 mp will not be enough for those neither even if we do get there at some point in the future.
Thanks for your response 35mmBluSky. Just a few points.

I hope we never come to the time when "Enough is enough". Yes, sensor sizes will probably remain fixed but sensor technology and processing speed will certainly improve. Film manufacturers were always improving their products and nobody said "enough is enough, don't improve your film further". I know that I tried many films and developers to get less grain and sharper images.

I don't think you are greedy or foolish if you want improved technology, increased dynamic range, more resolution and better high ISO performance. What is the downside to having these improvements.

How many pixels you would like depends on your post processing and printing needs. In post processing I am always trying to improve my images by cropping and sometimes action shots and outdoor macros need a fair bit! Others say they never crop.

If high ISO performance and camera responsiveness are not compromised, then I only see advantages in having more pixels. Computers have huge hard drives now so storage isn't an issue.

Ideally you would "bin" your 100 MP sensor to say 25 MP if you want to improve its high ISO performance. Then everyone is happy. 100 MP at ISO 100 for the croppers and 25 MP at ISO 12,800 for the dark churches. This can be done in camera or during post processing.
 
Why is 4/3 "stuck" to 16mp? The argument that it is not possible to get more real resolution falls flat: Sonys´1" sensor is 20mp and shows a great deal of resolution. Only the lens which cannot make the sharpness through out in the corners (RX100)
Increasing resolution causes A WHOLE LOT of problems, if everything else kept the same:

1) worse battery life

2) slower sequential shooting speed and/or higher buffer flush time

3) smaller buffer size (in frames)

4) worse vignetting in the corners

5) worse color shifts in the corners

6) fewer pictures per memory card

7) longer transfer and processing times

8) more line skipping or worse rolling shutter artifacts in video (if readout speed in mpix/s kept the same)

9) worse quality of live view (and EVF), especially in low light

10) worse low light performance

11) worse dynamic range

12) slower focusing (in CDAF and OSPDAF)

Probably something else too.

How many people would benefit prom sensor resolution even higher than 16 mpix? Very few - almost nobody prints, let alone prints poster-sized pictures and looks at them nuder microscope. How many people would suffer the consequences from one or more of 1-12? Almost everybody.

If fact I wish somebody would produce a m43 camera with slightly wider than m43 sensor with only ~12 mpix, for perfect UHD 16:9 crop but still with natively big 4:3 crop, and take all the advantages lower resolution brings. Sony almost done it on A7s, just managed to screw up enough things for the camera to be disappointing.

PS. I will explain every item if you don't understand why.

PPS. That said, I am sure that market ignorance will demand and m43 camera makers will produce higher res, 20-24mpix cameras, probably as soon as this Photokina (E-M5 Mark II?).
 
Last edited:
Why is 4/3 "stuck" to 16mp? The argument that it is not possible to get more real resolution falls flat: Sonys´1" sensor is 20mp and shows a great deal of resolution. Only the lens which cannot make the sharpness through out in the corners (RX100)
Your premise is wrong first of all. Nothing is... "stuck". Second, the differences are not as great as you seem to think. Most of the large prints I make are 16"x20" (on rare occasions I'll print a larger one). Taking an image and then cropping to the 5:4 aspect ratio with each camera would yield an imperceptible difference (240x192px) even if the sensors could produce a near equal picture. Even a 24MP camera cropped to 5:4 for a large wall type print doesn't produce an image all that much bigger. And 680x544px is not enough to even matter.


White area showing 5:4 crop and their respective image sizes from 3 different sensors.

It seems like you're just measurebaiting; please stop.

--
NHT
while ( ! ( succeed = try() ) );
 

Attachments

  • 2888592.jpg
    2888592.jpg
    53.3 KB · Views: 0
And Sony announces a 4k-capable camera...

Fine print: If you buy a $2000+ 4k recorder and lug it along.

At least they mention that detail in the press release. A lot of the time with Sony, you discover those things after you buy the product.

For most video users, Panasonic GH4 is likely to be better, cheaper, and hit the streets first.

Now that depends on what you use it for doesn't it.....

But if you insist, ok your as yet largely unavailable camera is better than the Sony as yet unavailable camera....there...happy?

I would love either the GH4 or the A7s ....A7s would be better for me (on specs at least) as I would use the better high ISO for both stills and video.

GH4 is not much better for stills than my GX7 and I leave that at home when shooting in low light with the A7.

Depends on what the user wants.
 
Why is 4/3 "stuck" to 16mp? The argument that it is not possible to get more real resolution falls flat: Sonys´1" sensor is 20mp and shows a great deal of resolution. Only the lens which cannot make the sharpness through out in the corners (RX100)
Why do you think 'only' 16 mp is a handicap? Is the image quality any less? Not that I can see.

Do you know what you are buying when you shop for more mp? Few do, do you? You can make giant enlargements with 8 and 12 mb images. The technical need for more mp was surpassed several years ago

What does a few more pixels actually mean to YOUR photography? Why aren't you clamoring for quality lenses? Something that will always have some effect on your photography?

--
I still like soup. . .
http://glenbarrington.blogspot.com/
16 may not be a handicap, maybe Sonys 20mp on the 1" sensor is. I don´t know. I have compared many photos between Canon s95 10mp against RX100 20mp and for ME the difference (10 vs 20mp) is big, both in noise and in detail (Yes, I like to study details in pixel level and like cropping also). And in HDR and video mode the difference is enormous.
 
Last edited:
but you also can't change one fact: the more mp, the smaller individual pixel size also comes into play to hurt your overall performance of the sensor.
That's not a fact.

--

Bob
Everything else being equal, it is a fact. At least from color-filtered and FSI sensors.
Let's see some evidence of this 'fact' then. I say it is bogus. I say that all else being equal in reality pixel size has little effect on the 'overall performance' - for instance, compare D800 and D610 or the two Sony A7s.



6085d34ca14347f69e82ec86bffb40b8.jpg.png



76ba8af8940c40b98ee83891847f001f.jpg.png



However, smaller pixels give more resolution and acuity versus a slight increase in shadow noise - all else being equal. When worked through, the theory suggests smaller pixels have no ill effects, and increase DR and resolution. In practice the theoretical improvements can't quite be delivered producing the tadeoffs above. The next generation, the tradeoffs work around a smaller pixel size, and thus more pixels.

--
Bob
 
I have compared many photos between Canon s95 10mp against RX100 20mp
Have you compared the size of those pixels and the age of those sensor technologies ?

The RX100 sensor is about 116mm² and the S95 just 43mm².

That means that even though the RX100 has twice as many pixels, each pixel is 34% larger so it will have less noise just on that basis.

Put another way, on an equal print size there is going to be about 2.7x the physical pixel area contributing to each pixel in the displayed image in the RX100's favor.

That's an overwhelming advantage.

And the age of those sensor technologies is a factor as well. There is no doubt that there was a significant improvement from the S95 ( 2010 ) to the RX100 ( 2012 ). Not to mention the fact that Sony have proven to be a maker capable of cutting edge sensor design.
 
I have compared many photos between Canon s95 10mp against RX100 20mp
Have you compared the size of those pixels and the age of those sensor technologies ?

The RX100 sensor is about 116mm² and the S95 just 43mm².

That means that even though the RX100 has twice as many pixels, each pixel is 34% larger so it will have less noise just on that basis.

Put another way, on an equal print size there is going to be about 2.7x the physical pixel area contributing to each pixel in the displayed image in the RX100's favor.

That's an overwhelming advantage.

And the age of those sensor technologies is a factor as well. There is no doubt that there was a significant improvement from the S95 ( 2010 ) to the RX100 ( 2012 ). Not to mention the fact that Sony have proven to be a maker capable of cutting edge sensor design.
Which said in a simpler manner, this means that the number of pixels isn't nearly as important as the technology behind them. And THAT brings us back to a restatement of my original question. Given roughly, the same level of technology, wouldn't improving the quality of lenses at this point do more for the resolving power and image quality of our photos than would stuffing more pixels into a photo using the same technology?

I suspect you are counting dots, when you should be measuring resolution and acuity.

--
I still like soup. . .
http://glenbarrington.blogspot.com/
 
Last edited:
I remember my first digital camera, a Casio, cost £800 in 1998 (I think) and produced just over 1mp, I thought it was great. I remember the other guys at the camera club debating that to achieve 35mm film quality would take years and have to be 15mp. Well here we are!

I think it's also no coincidence that when canon moved up the mp count from the 1DsII (16.7mp), they very quickly started to produce the (very expensive) Mk II L series lenses to counter the claims that users couldn't achieve the expected gains in resolution due to the lenses being the limiting factor. Interesting when the flagship 1Dx was introduced with fewer pixels than it's predecessors.
 
but you also can't change one fact: the more mp, the smaller individual pixel size also comes into play to hurt your overall performance of the sensor.
That's not a fact.

--

Bob
Everything else being equal, it is a fact. At least from color-filtered and FSI sensors.
Let's see some evidence of this 'fact' then. I say it is bogus. I say that all else being equal in reality pixel size has little effect on the 'overall performance' - for instance, compare D800 and D610 or the two Sony A7s.
Who said that between D800 and D610 all else is equal?

However, smaller pixels give more resolution and acuity versus a slight increase in shadow noise - all else being equal. When worked through, the theory suggests smaller pixels have no ill effects,
They have plenty of ill effects stemming from reducing the ratio between light-gathering area of the pixel and the area taken by electronics and barriers between pixels, and also from the filtered nature of most modern color sensors and the fact that most light from a lens does not fall straight down.
and increase DR
How smaller pixels increase DR? Unless the camera uses technique when some pixels are exposed less than others (which is theoretical as no modern camera actually does this), it decreases DR, because smaller pixels have smaller full well capacity, so the ratio between the amount of light which overfills the pixel and the amount of light which does not produce stronger signal than read+light noise - DR - is smaller.
 
RX10/100 20mp sensor is about 116mm² = 172 000 pixels per mm²

E-PL5 16mp sensor is about 225mm² = 71 000 pixels per mm²

If the RX10/100 would have the same pixel density as E-PL5 it would land around 8MP.

So what you guys suggest is that RX10/100 should have better image quality with 8MP?
 
Last edited:
I have compared many photos between Canon s95 10mp against RX100 20mp
Have you compared the size of those pixels and the age of those sensor technologies ?

The RX100 sensor is about 116mm² and the S95 just 43mm².

That means that even though the RX100 has twice as many pixels, each pixel is 34% larger so it will have less noise just on that basis.

Put another way, on an equal print size there is going to be about 2.7x the physical pixel area contributing to each pixel in the displayed image in the RX100's favor.

That's an overwhelming advantage.

And the age of those sensor technologies is a factor as well. There is no doubt that there was a significant improvement from the S95 ( 2010 ) to the RX100 ( 2012 ). Not to mention the fact that Sony have proven to be a maker capable of cutting edge sensor design.
Which said in a simpler manner, this means that the number of pixels isn't nearly as important as the technology behind them.
That is why I am saying bigger pixels are better (for most people and from a certain resolution) ALL ELSE BEING EQUAL. Sure, technology improves - but then companies turn around and squander the improvements by increasing pixel counts because megapixels is the one thing which all customers understand. Of course resolution mattered a lot when sensors improved from 640x480 to 800x600. Not anymore, yet the simplistic measure of camera quality is stuck in people's minds.
And THAT brings us back to a restatement of my original question. Given roughly, the same level of technology, wouldn't improving the quality of lenses at this point do more for the resolving power and image quality of our photos than would stuffing more pixels into a photo using the same technology?
It would, but again, for most people the resolution of modern ILC lenses is enough too (of course, older designs are mostly terrible, especially wide open or close), but aperture is what improves low light pictures the most.
 
Last edited:
RX10/100 20mp sensor is about 116mm² = 172 000 pixels per mm²

E-PL5 16mp sensor is about 225mm² = 71 000 pixels per mm²

If the RX10/100 would have the same pixel density as E-PL5 it would land around 8MP.

So what you guys suggest is that RX10/100 should have better image quality with 8MP?
No, it would have about the same image quality (although better DR), maybe slightly better in RX100 with FSI sensor (RX10 and RX100II have BSU sensors where it does not matter that much), but it would have about 2.5 times battery life (as measured in still shots without flash), 2.5 faster sequential shooting, 2.5 times less noise in viewfinder/LCD (which especially visible in low light) and faster, more precise focusing in low light.

I would prefer its 3:2 sensor be about 10 mpix so 16:9 crop would produce exactly 8.3 mpix UHD frames. Without downsampling, it would not require as much power to shoot 4k 4:1:1 or 4:2:0 video natively - which now it cannot do. And downsampling for 1080p video would also be much simpler (primitive 4:1 binning), reducing power requirements for shooting full HD video too, and maybe even raising the quality of produced FHD video to 4:2:2 or even 4:4:4.

The readout of the sensor would also be twice as fast, reducing rolling shutter effects which plague the cameras (and their 4k-capable AX100 sibling) now.

That is what has been squandered so the number of megapixels looked more impressive for ignorant customers.
 

Keyboard shortcuts

Back
Top