2 sensors in theory?

Elly

TPF Noob!
Joined
Oct 12, 2009
Messages
25
Reaction score
0
Location
Foggia
Can others edit my Photos
Photos OK to edit
in theory if there were two sensors identical in everything but pixels.one is 20mp and one is 10mp and i take the same photo with both.if i then resize the 20mp photo to 10mp would the two photos be identical and show the same amount of detail?
 
Typically, things don't actually work out that way. In the real world there have been and still are instances where sensors of 2x larger can be shot at full resolution size and their files down-sampled to equal the capture size of lower-MP count and therefore lower-resolution sensors.

In decent light, if one takes a studio flash shot with a Nikon D2x and then a Nikon D40, when resampled downward in size, the D2x file still looks a little bit better. Back when the Kodak 14n was the only F-mount full-frame at 13.8 megapixels, I test shot one and was literally *astounded* at the resolving power from the 14n,even when images were down-sized to what was then the current hobbyist/semi-pro standard size of 6 megapixels.
 
No.

In an exposure time, the larger pixels would record, say, 100 photons. Noise is SQRT(N) = 100±10. In the same amount of time, the smaller pixels would record 25 photons, the noise would be 25±5. That's a raw noise level of ±10% vs. ±20%.

However, when you bin the smaller pixels 2x2, if you AVERAGE you're still going to have a level of 25, and the noise will be SQRT(4x5^2)=10. So now the "macropixel" will have a value of 25 with an overall uncertainty of ±10, or ±40%. Scaling it up to match the exposure of 100 photons would correspondingly scale up the noise to ±40% or 100±40.
 
you mean that the resized larger sensor photo would still show more fine detail by a bit?
 
posted at the same time astrostu :) thats a bit beyond my technical knowledge to easily understand
 
posted at the same time astrostu :) thats a bit beyond my technical knowledge to easily understand

It's all about sensor noise in counting statistics. Basically, with light, your uncertainty is the SQRT() of how many photons you collect. Since you're collecting more photons over a larger pixel, your noise is less. Therefore it's not identical.
 
is that still true if you are shooting at iso 100 with no noise?
 
by the way i'm not talking about different sensor size.its the same size just one has more megapixels.so the one with more mp will have smaller pixel size.
 
You can't have no noise. It's physically impossible. The math I showed assumes you're at "ISO 0" -- no noise introduced by anything other than the raw statistics of photon counting. It's why when you shoot at higher ISOs, you get more apparent noise because you have the same photons , you're just amplifying both it and the ±10% noise (on the 100). As opposed to if you shot for a longer period of time and got, say, 10,000 photons, then the noise would be SQRT(10,000)=100 -> 10,000±100 = ±1% instead of ±10%.

I assumed you're talking about the same physical sensor size, say 1"x1". It's just how many pixels it's divided into.
 
astrostu.i think you say that the one with larger pixel size (the 10mp sensor) will have less noise so more detail?
 
astrostu.i think you say that the one with larger pixel size (the 10mp sensor) will have less noise so more detail?

Detail is a completely different issue. If both sensors have the same lens attached, then the one with more pixels will have more detail (assuming you're not diffraction limited (do a Wikipedia search on that)). But, each individual pixel with more detail will have more noise. Note that by "detail" I assume you mean, effectively, resolution, or the ability to differentiate the distance between two closely spaced dots/lines/whatever.

Edit: The only time that detail comes into play in conjunction with noise is when the noise is so high relative to the signal that you can't tell two pixels apart. For example, say you have a pixel that was hit by 9 photons, it could have recorded a value anywhere from 6-12 (9±3). The pixel next to it was hit by 16 photons, which ideally would easily be different from 9, but its noise is ±4 so it could have recorded anything from 12-20. If they both happen to record 12, then the detail between the two is lost.
 
Last edited:
when i read my original question it still looks like a simple question to me:sillysmi:
 
Elly,
THere was a post done just yesterday of photos made in south central Iowa. And on the reader board outside of what looked to be an auto repair garage was this message:

"Imagine if there were no hypothetical situations."

Words to live by?
 
It was a simple question. Here's a simple answer: no ;)

:scratch:and what one would show more noticable detail? not hypothetical just easily noticable when someone looks at the photo.derrell answered my question in the first post but i want to see if others agree.
 

Most reactions

New Topics

Back
Top