Sensor Size_difference?

[Ronnoco]

A number of pros here have suggested that the larger sensor size of full frame cameras such as the Canon 5D are sharper than DSLRs with smaller sensors, and DSLRs with smaller sensors are sharper than superzooms that have even smaller sensors and these in turn are sharper than most compacts again related to sensor size.

Test results from Popular Photography and Image Resource don't seem to bear that out.
The Nikon 200D which does not have a full frame sensor and only 10 megapixels tests out at slightly better resolution than the Canon 5D at 12 megapixels. The Samsung Pro 815: an extreme range superzoom tests out with the same resolution as the Canon 20D an excellent DSLR. The Minolta A200 superzoom also tests out with the same resolution as the Canon 20D, DSLR.

It would seem that sensor size does not make a difference in terms of resolution or sharpness on these cameras, or the test results would be drastically different.

Ronnoco

[Photo-John]

Resolution, as in resolving power, isn't the issue. The issue is the quality of information captured with each pixel. Larger pixels mean cleaner information. The bigger the pixel, the better less noise there will be and the more color depth is possible. Improvements in sensor-level noise reduction will help smaller sensors. But they'll also help larger sensors. Therefor, a larger sensor will always be able to delive better image quality. If best image quality is the goal, a larger sensor is always preferable. If flexibility is the goal, a superzoom compact might be a better camera.

[Sebastian]

Resolution, as in resolving power, isn't the issue. The issue is the quality of information captured with each pixel. Larger pixels mean cleaner information. The bigger the pixel, the better less noise there will be and the more color depth is possible. Improvements in sensor-level noise reduction will help smaller sensors. But they'll also help larger sensors. Therefor, a larger sensor will always be able to delive better image quality. If best image quality is the goal, a larger sensor is always preferable. If flexibility is the goal, a superzoom compact might be a better camera.

Physics. Can't beat it.

Great rundown of the issues PJ.

[Photo-John]

I forgot to mention my favorite example of how a larger sensor makes a difference. I have a 4-megapixel digital SLR, the Canon EOS 1D. I've taken lots and lots of photos with all kinds of cameras, including the Konica Minolta A200. And while the image quality of the Konica Minolta A200 is very good, it doesn't come close to that of the EOS 1D, which has half the resolution. And because the 1D produces cleaner files, I can enlarge them almost as much as I want without any artifacting problems. And with more color-depth, the 1D has a less of a tendency to blow out highlight details or block up in the shadows. That's not true of the A200. I'm not saying the A200 has poor image quality - I've had images from the A200 published as full-page magazine images. It's great - for a compact digital. But the 1D is definitely better.

[Ronnoco]

Resolution, as in resolving power, isn't the issue. The issue is the quality of information captured with each pixel. Larger pixels mean cleaner information. The bigger the pixel, the better less noise there will be and the more color depth is possible. Improvements in sensor-level noise reduction will help smaller sensors. But they'll also help larger sensors. Therefor, a larger sensor will always be able to delive better image quality. If best image quality is the goal, a larger sensor is always preferable. If flexibility is the goal, a superzoom compact might be a better camera.

Colour depth is still 12 bits for most DSLRs and well as superzoom compacts. Dynamic range may be 10 stops on the Canon Mark II but the print range is 6.6 stops for all cameras and Nikon compacts for example surpass that to 8.8 stops.

There is also a loss in levels when a 12 bit raw image is converted to an 8 bit image for printing as well. Better dynamic range may be a questionable advantage if it is much of it is lost in the conversion and the printing process.

Improvements in sensor level noise reduction are at the moment proprietary and very different in nature between for example, Canon Nikon and Sony. Whether these processes will be used on larger sensors depends on the indivdual companies and their marketing plans. Ten plus megapixel compact superzooms are coming out or on the drawing boards and some may have OLED viewfinders, stereo sound and improved video capability. Larger sensors in superzooms are also being considered.

Personally pocket cameras, superzooms, DSLRs, a good film camera or two, and a studio camera are all useful for different purposes. Nevertheless it will be interesting when the quality of superzooms and DSLRs comes closer together and flexibility becomes a bigger issue in some areas of photography.

Ronnoco

[OldSchool]

A sharp picture requires a sharp lens. Thus, a smaller sensor will need a sharper lens than a larger sensor to get the same precision.

I'll let others extrapolate from here....
Tim

[Photo-John]

Colour depth is still 12 bits for most DSLRs and well as superzoom compacts. Dynamic range may be 10 stops on the Canon Mark II but the print range is 6.6 stops for all cameras and Nikon compacts for example surpass that to 8.8 stops.

Good point about the technical color depth. However, the real color depth is different with different sizes of pixels. I like to compare the difference between a glass and a bucket. Which one is going to get better information? I say the bucket. More surface area means more light and cleaner, richer information can hit the pixel, before processing. That means more dynamic range in the RAW images

And even though the dynamic range of a print may be much smaller than the dynamic range of a RAW or 8 bit converted image, the higher dynamic range delivers images with more potential, should the photographer care to optimize. Most consumer photographers probably won't care. But serious enthusiasts and pros do care and will optimize their images either at the RAW level, in Photoshop, or both. Those photographers will benefit from the higher dynamic range and even though the dynamic range of the print may be lower, their optimized files will make better prints.

Again, the real issue here isn't image quality - it's flexibility. If I want to ride my bike 50 miles with a camera, which camera do I want to carry? I want the smallest, lightest camera, with the most features, that can deliver useable image quality. The Konica Minolta A200 fits the bill. I'd still rather have my SLR. But there are places I definitely don't want it on my back. Actually, I'm looking to the Four Thirds System for the best of both worlds. The pixels are larger than those on compact cameras and the system is very powerful in terms of features and lens reach. I think it's a great system and I'm hoping it continues to grow and gain in popularity.

[Ronnoco]

A sharp picture requires a sharp lens. Thus, a smaller sensor will need a sharper lens than a larger sensor to get the same precision.
....
Tim

Camera comparisons don't bear that out, Tim. The Samsun Pro815 long range superzoom for example has the same resolution in tests as the Canon 20D which is a DSLR with a larger sensor, so does the Minolta A200, which is also a superzoom.

The assumption is that the superzooms do not have as sharp a lens as DSLRs, but if that is true, how are they able to match the resolution and sharpness of DSLRs in photo lab tests?

Ronnoco

[OldSchool]

Camera comparisons don't bear that out, Tim. The Samsun Pro815 long range superzoom for example has the same resolution in tests as the Canon 20D which is a DSLR with a larger sensor, so does the Minolta A200, which is also a superzoom.

The assumption is that the superzooms do not have as sharp a lens as DSLRs, but if that is true, how are they able to match the resolution and sharpness of DSLRs in photo lab tests?

Ronnoco

Camera tests don't bear what out????

I have made no assumptions about the resolution (in general or in particular) of a superzoom or a DSLR. Further, I have offered no conclusion of what is "better". I just stated a fact of optical geometry that is often missed when discussing the benefits of sensor size and/or mega-pixel capacity.

If a particular camera (system) resolves better than another, it may be because of the particular sensor -- or it may because of the particular lens -- or it may be because of both.

Further, if only a single camera (system) was used in the test, then you have the statistical issue of possibly being on the tail of a distribution.

Sorry, no argument from me.
:^)
Tim
(someone who tests things)

[Ronnoco]

Camera tests don't bear what out????

I have made no assumptions about the resolution (in general or in particular) of a superzoom or a DSLR. Further, I have offered no conclusion of what is "better". I just stated a fact of optical geometry that is often missed when discussing the benefits of sensor size and/or mega-pixel capacity.

If a particular camera (system) resolves better than another, it may be because of the particular sensor -- or it may because of the particular lens -- or it may be because of both.

Further, if only a single camera (system) was used in the test, then you have the statistical issue of possibly being on the tail of a distribution.

Sorry, no argument from me.
:^)
Tim
(someone who tests things)

Well, from pros, the logic goes this way. Sharper lens makes a sharper photo as in greater resolution since greater resolution means greater sharpness. (That is in effect what you said.) With that, is that a longer zoom is usually associated with lower resolution and lower sharpness. A shorter focal length lens by Canon or Nikon is according to many photographers better glass than a longer lens from another company.
Optically and logically the reasoning makes sense.

My point however is that if you follow the logic, a superzoom with a long length like 28mm to 300mm+ should be much less sharp than a Canon 35mm to 55mm for example. That also makes sense and is logical.

My point is that the tests seem to be surprising in that they show that the resolution or sharpness of longer zooms with supposedly cheaper glass is equaling that of shorter length with supposedly "higher" quality glass from Canon and Nikon.

Ronnoco

[OldSchool]

(snip -- all good points)

My point is that the tests seem to be surprising in that they show that the resolution or sharpness of longer zooms with supposedly cheaper glass is equaling that of shorter length with supposedly "higher" quality glass from Canon and Nikon.

Ronnoco

And, that's my point...

The key word here is "supposedly". If the supposedly cheaper glass is ACTUALLY better then the supposedly higher quality glass, then one cannot make a conclusion about sensor performance.

Really, it gets down to the purpose of the study. If the test was to qualify sensors, then all other variables should be controlled for (like using the same lens for all sensors in the test). Note that the title of this thread is "Sensor Size Difference".

Cheers,
Tim

[Anbesol]

And, that's my point...

The key word here is "supposedly". If the supposedly cheaper glass is ACTUALLY better then the supposedly higher quality glass, then one cannot make a conclusion about sensor performance.

Really, it gets down to the purpose of the study. If the test was to qualify sensors, then all other variables should be controlled for (like using the same lens for all sensors in the test). Note that the title of this thread is "Sensor Size Difference".

Cheers,
Tim

ah yes i see what youre saying.

That brings up an interesting point ive observed - is people making absolute final judgements between a couple cameras based on one minor difference. Particularly the (morons) that say "oh this is one megapixel better, it must be better"! I try to inform people (one of my many jobs is as a camera salesman) but some people are too stubborn to realize that cameras arent all megapixels.

on the same token - i can see why sensor size doesnt create an incredible difference in image quality and noise performance with said case in point 20D vs 5D. I think the biggest, TRUE difference sensor sizes create is physical resolution that the lens offers, as in the glass is placing a much larger image onto the image sensor, as much as this world loves things that are small - bigger can often times be better.

I would like to say about the 20D vs 5D thing, Ive heard many people say that the 20d is nice because it converts you more telephoto. but heck, thats like saying that a regular size printer is good because it prints A4. The 5D, per pixel is still capable of achieving everything that the 20D does, heck, you put a digital zoom on that 5D (if it were possible) and youre basically using a slightly slower 20D without a built in flash.

[SmartWombat]

My point however is that if you follow the logic, a superzoom with a long length like 28mm to 300mm+

I think the superzoom has a focal length of more like 7mm, because it has a smaller sensor (my A1 is 7.2-50mm).
So you're comparing totally different lenses.
And if as you say it's easier to make sharp short focal length lenses, that's just what your superzoom has.
No surpsise there.
But it's yet another apples/oranges comparison.

[Ronnoco]

And, that's my point...

The key word here is "supposedly". If the supposedly cheaper glass is ACTUALLY better then the supposedly higher quality glass, then one cannot make a conclusion about sensor performance.

Tim

That is not logical. Anyone will tell you that because of the number of elements and the precision required that short focal lengths and shorter zooms are better quality than larger zooms and the longer the zoom range, the greater the compromise in quality.

So if a superzoom with a smaller sensor and longer zoom range (compromised in lens quality?) resolves the same number of lines as a DSLR with a shorter focal length (and better quality?) lens as well as a larger sensor, then the only possible conclusion is that neither glass nor sensor make a noticeable difference in sharpness or resolution.

Ronnoco

[Ronnoco]

I think the superzoom has a focal length of more like 7mm, because it has a smaller sensor (my A1 is 7.2-50mm).
So you're comparing totally different lenses.
And if as you say it's easier to make sharp short focal length lenses, that's just what your superzoom has.
No surpsise there.
But it's yet another apples/oranges comparison.

O.K. good point! That would explain why glass does not seem to make a big difference. Focal length is longer in a superzoom but only 7.2 - 50mm versus 35mm to 55mm on one DSLR kit lens.

So, bottom line seems to be that sensor size does not make a difference in sharpness or resolution either, but does make a difference in noise level and dynamic colour range but the dynamic colour range of large sensors is restricted by the print capacity of 6.6 stops.
Since superzooms with smaller sensors are past that at 8.8 stops for example, the value of dynamic colour range by the Canon DSLR Mark II at 10 stops for example is rather limited to say the least.

So we get back to noise as the only real difference related to sensor size that shows up on a print.

Ronnoco

[xsport652]

wow very informative. Someone should make this a sticky because it clears up many myths and I just learned a lot from these posts. You guys put it in a very simple way.
Thanks.