Is there a resolution limit for sensors?

Messages
364
Edit My Images
Yes
Just crossed my mind that if we continue to see increases in sensor resolution, when will it start to affect the image quality adversely? What I mean is, larger sensor pixels in say full frame bodies are more sensitive to light as they have a larger surface area. But as you squeeze more MP into a sensor, these pixels will have to reduce in size, so will there be a point at which there is little to no value in increasing the MP?

If so, would FF cameras of today be relatively future proof having reached a standard/quality that probably won't be bettered by much?
 
what does full frame mean - what about medium format

Optically - yes - there are a number of issues

1. resolving power of lenses
2. pixel site size
3. thickness of anti-alliasing filter on sensor

Electically, There is the problem of cross talk (between pixel sites and the correspondining wiring and amplifiers), as things get closer, the effects are worse


Ideally, a much larger sensor, with larger pixel sites, which ultrasonically can be moved so multiple shots can be taken to reduce Moire (reducing the need for the Anti-aliasing filter)

Thats why MF cameras with Digital backs are pretty unbeatable optically
 
Personally, I think that 20odd MP in a 35mm size sensor or 50ish in a medium format sensor is more than enough for any application.

What we'll hopefully start to see is a continued improvement in noise handling and colour transitions. And of course the glass that all these big sensors need to really shine.

Sadly pixies are what sells cameras and this mad arms race between the different camps could carry on and on and on and on and..............
 
tbqfh i'd be happy with 6mp and lower noise :)

however i think the camera companies are somewhat keeping up with the megapixel race
 
Professional astronomers get around many problems that larger ccd chips suffer by cooling the actual chip and electronics to -20/40 degrees celsius. I use a dedicated astronomic camera that utilizes an electronic gizmo called a peltier device to cool the ccd chip to about -20.
These types of camera are used for long exposures so signal to noise ratios can be kept under control.

Alas it takes about 20 minutes to achieve so not really suitable for consumer dslr's yet.

Examples here

Even so there are many astrophotagraphers who use dslr's and get stunning results by taking many short exposures then stacking the images, subtracting dark frames and such.

Example here
 
Professional astronomers get around many problems that larger ccd chips suffer by cooling the actual chip and electronics to -20/40 degrees celsius. I use a dedicated astronomic camera that utilizes an electronic gizmo called a peltier device to cool the ccd chip to about -20.

Alas it takes about 20 minutes to achieve so not really suitable for consumer dslr's yet.

Examples here

that is some piece of kit!
 
I believe the most high tech direction is actually a single pixel sensor...
 
re density - yes absolutely there is a limit, because of the wavelength of light. For example it is impossible to photograph the surface of a CD because the variations are so small that the reflected light gets mixed up together (which actually is how they play in the first place).
 
I read somewhere that Canon have a road map which sets out how they will get to around 25 MP for crop sensor DSLRs and 55-60 MP for full frame DSLRs. So you'd think that if there are limits, they must be at or beyond that.

But ... point-and-shoots are already w-a-a-y beyond that. A full-frame sensor measures 36 x 24 mm, which is 8.64 sq cm; so their road map takes them to (very roughly) 7 MP per sq cm. But the newest P&S cameras already deliver 43 MP per sq cm!!! (For example the Canon SD960IS packs 12.1 MP onto a sensor which is 6.2 x 4.6 mm, or 1/30th the size of a full-frame sensor.

If Canon engineers could make a full frame sensor with the same pixel density as the SD960IS, it would have about 360 MP.
 
But ... point-and-shoots are already w-a-a-y beyond that. A full-frame sensor measures 36 x 24 mm, which is 8.64 sq cm; so their road map takes them to (very roughly) 7 MP per sq cm. But the newest P&S cameras already deliver 43 MP per sq cm!!! (For example the Canon SD960IS packs 12.1 MP onto a sensor which is 6.2 x 4.6 mm, or 1/30th the size of a full-frame sensor.

If Canon engineers could make a full frame sensor with the same pixel density as the SD960IS, it would have about 360 MP.

Yes but what's the point? The P&S cameras pursue a different goal - they don't care about high ISO performance and image quality even at a base ISO is leaving something to be desired. A vast majority of them with these super dense sensors rely on in-camera processing software to make photos look OK but in RAW it is the whole other picture...

Of course the technology moves on so it might be possible one day that we have relatively noise free sensors with those densities available - but now is not the time yet ;)

Imho, Nikon showed everyone what can be done (with D3) when modern sensor making technology is applied to lower density sensors and larger pixels. Imagine what kind of camera would it be if they produced 6-8 Mp sensor in full frame - candlelight photography with stanadard kit lenses anyone?-)
 
There will be all manner of developments yet. Like people have said megapixels sell cameras! We are already using processors that defy the laws of physics as we knew them 10-20 years ago. The CCD is basically a computer chip and the minimum size of a site is going down just like the minimum size of a transistor is going down on processors. WE all know that megapixels are just PART of the total when it comes to quality but the "great unwashed" out there have no idea at all.

I took my D1x from the car into work to take some pictures using the 105mm macro and a SB600. I took the pictures which were perfectly good then the usual happened and someone came in and started asking questions about the camera (it is large in comparison to what most people use). The first question how many megapixels is that then. "Just under 6" I replied. "6!" came the response. "yes 5.6". "Well whats the use in that, you should have asked me to take the picture. My phone is more than that". He then took several pictures with his phone and emailed them to himself so they could be seen on the screen whilst I popped the memory card into the computer. Guess who looked very stupid :)

The thing is, this is what people actually think and it will carry on at least for a while. We thought that it was going to carry on with computers and the MHz thing but that was killed off by Intel and AMD when THEY came a cropper with it.
 
We thought that it was going to carry on with computers and the MHz thing but that was killed off by Intel and AMD when THEY came a cropper with it.

Yes they did - since electricity moves ~6cm per clock cycle at 3GHz, hence the need for multi-core instead.

Anyway re density, I have no idea how close sensors could be spaced but 1.5x wavelength seems like a likley limit - at 400nm that would be ~1.8MP per mm2 - well over 1GP for full-frame :D
 
Anyway re density, I have no idea how close sensors could be spaced but 1.5x wavelength seems like a likley limit - at 400nm that would be ~1.8MP per mm2 - well over 1GP for full-frame :D

Now that would be interesting! Imagine doing long exposures with a 1GP camera, you'd need a mainframe attached to it to process the shot. :p
 
What if this thread had been started 10 years ago? or even 20 years ago!!!

John
 
There'll be some new sensor technology along at some point that'll improve things vastly :)

I just with the phone makers would stop this stupid megapixel race. What do people want a cameraphone for? Snaps on a night out in low light. My C905 crams an 8mp sensor into something I can only assume is the size of a pea. Make it 2mp and vastly improve ISO performance and it would be far far better.
 
The problem we have is that MP is taken as a measurement of quality, which it isn't, a 6mp DSLR easily blows away an 8 mp phone camera etc etc. We really need a new standard measurement, that take into account pixle size and density, sensor size etc, and comes up with a simple comparrison of image quality rather than just the number of pixels.
Wayne
 
i always knew that MP never made the cameras quality but i was unaware (according to canons roadmap) that pizel densities are not reaching there limit! i understand that technology needs to improve to utalise the exta density and keep noise controlled and lens resolution usuable (could we see a new breed of lenses per chance), but wow 35mm FF DSLRS could become the digital backs on MF that exists now.
 
all i can add is the more pixels i have the easier my job is
before we got to 14mill plus transparencies were still superior
for magazine work at least

now we have cracked this threshold (1Ds mkII and mkIII) its a lot easier

the more the merrier i say

bring on 50mill plus!!!
 
Back
Top