Sensor Resolution

[jdemott]

That is what most folks refer to as the multiplier effect.  If you put the same lens (for example, a 50 mm lens) on both cameras, then the APS sensor camera (Nikon) would show a smaller portion of the field of view but with greater "resolution", i.e., more detail.  If you put a 75 mm lens on the full frame (Canon) camera, it would then show the same field of view and same degree of detail as the Nikon does with the 50mm lens, assuming an identical number of pixels on the sensor and no differences resulitng from lens quality etc.  Rather than speaking of increased resolution, differences in sensor size are generally measured by the multiplication factor (1.5x) that would be applied to the lens focal length to give an equivalent field of view.  Resolution then is generally described as a property of the number of pixels and the optical properties of the lens and sensor filter.

Strictly speaking, I think you are right to speak of the increased resolution of smaller pixel sizes, but that doesn't seem to be the practice.  I recall one case where a very careful reviewer compared a full frame camera to an APS sensor camera and included some comparison shots taken with identical lenses (which of course showed greater resolution for the smaller sensor).  A number of people strongly criticized the reviewer for even bothering to do such an "unfair" comparison.

If you do a search on this and other forums you will find innumerable discussions about the advantages and disadvantages, real and imagined, of the different sensor sizes.  Among other things, those discussions get into the fact that the smaller pixel sizes of smaller sensors demand a higher degree of optical resolution from a lens in order to deliver the full resolution of the sensor.

[AJSJones]

It seems to me that if the Nikon and Canon both have the same number of pixels, and the Nikon sensor is smaller then the Nikon should have smaller sensors translating at least theoretically into greater resolution. (talk about pixel peeping!)

DrGary

That is completely incorrect.

Jonathan,
I agree with the rest of your response, even though it didn't directly address the question.   However, the first sentence, even without the "completely",  I respectfully disagree with.  I expect the OP was referring to spatial resolution where more closely spaced sensor sites will yield a higher resolution "theoretically" (aka , all other things being equal).
The 10D "outresolved" the 1Ds; for example  (see here) so the "quality" issue of the smaller sites did not offset their closer spacing in the fineness of the details captured (from a given image falling on the sensors).  Not so many details were captured, of course, and for many, seemingly exclusively since the advent of digital, more pixels is what is meant by "higher" resolution.  Would you have a problem with a generalization that "modern fine grained film has higher resolution than coarse-grained film from 50 years ago"?  Is that not a valid analogy?
Andy

[Graeme Nattress]

The more pixels, the more resolution. Smaller sensors, for the same resolution, have smaller pixels, but do not have greater resolution. They probably end up having lower resolution, as smaller pixels means you need a sharper lens to fully max out of the detail you can get from that sensor.

This is a bigger problem in the video world where sensors can be 1/3", and if the camera is high def, that means very, very small pixels. This can lead to poor low light, lack of dynamic range, and fuzzyness on small details. For years video people have paid a premium for large chip 2/3" cameras which do tend to produce very much better images.

Graeme

[AJSJones]

I suppose it depends on what the meaning of "is" is.

It would appear to depend on what the meaning of "resolution" is and it's clear that different folks use the term differently.
I like the way Lester distilled the issues that come up in discussions like these and agree the "system" resolution is best seen as system MTF at a particular S/N.  

1/(system resolution) = {1/(sensor resolution)} + {1/(lens resolution)}

BTW  The extra pixels and stuff that the 1Ds has over the 10D don't mean much if you are only printing 4x6 prints   so I totally agree it depends on : what you want to do, what tools you have and what tools are available.  So I shoot pictures of birds with the finest grain "film" available in the digital world (often cropping) and I shoot landscapes with a 4x5 with Provia100F arguably the finest grained slide film available today (for image size)- in case it's not clear where I'm coming from  :laugh:

[Dr. Gary]

It seems to me that if the Nikon and Canon both have the same number of pixels, and the Nikon sensor is smaller then the Nikon should have smaller sensors translating at least theoretically into greater resolution. (talk about pixel peeping!)

DrGary

[Jonathan Wienke]

It seems to me that if the Nikon and Canon both have the same number of pixels, and the Nikon sensor is smaller then the Nikon should have smaller sensors translating at least theoretically into greater resolution. (talk about pixel peeping!)

DrGary

That is completely incorrect. There are two factors involved in the resolution image: the quantity of pixels, and the quality of the pixels. The first is easily measured, the second is not, so marketing tends to focus on quantity rather than quality. Sensor size has an effect on pixel quality; all else being equal (which is practically never the case given the rapid pace of sensor performance improvement) physically larger pixels will capture more dynamic range at lower noise levels more accurately than physically smaller ones. But size has other effects as well. As previously stated, a smaller sensor is more demanding of lens performance than a larger one; but on the other hand, it is easier to design a high-resolution lens if it doesn't need to cover a large image circle. Smaller sensors mean you can fill the frame with smaller subjects with a lens of a given focal length than with larger sensors. That is a good thing if you shoot a lot of telephoto stuff, and bad if you use wide angles. Smaller sensors are also more conducive to wide DOF than larger ones, for reasons explained in another thread. This is a good thing for macro shooters that struggle to get a usable amount of DOF, and a bad thing for portrait photographers that like to shoot wide open and creatively use narrow DOF.

The bottom line is that there are advantages and disadvantages to the various formats of DSLRs, just like there are advantages and disadvantages to using 35mm film vs 4x5 film. Evaluate what you shoot and which format's strengths play to the intended task, and choose accordingly. Use the correct tool for the job.

[Jonathan Wienke]

I suppose it depends on what the meaning of "is" is. "Greater spatial resolution of the sensor" is meaningless unless you don't have glass long enough for the job and are trying to crop the least amount possible. In that special case it is possible to get a better print from a 10D than a 1Ds, but in any instance where one can achieve the desired composition in-camera, the 1Ds will completely outclass a 10D by resolution or any other criteria you care to name. And as I pointed out in my previous post, the increased spatial resolution of a smaller sensor is a disadvantage when shooting wide-angle stuff.

Yes, a physically smaller sensor has greater spatial resolution than a larger sensor with an equal pixel count. But that doesn't necessarily translate into greater resolution in the final image, which is ultimately the only resolution that really matters. Ultimately the other factors like signal-to-noise ratio and dynamic range and the things I mentioned in my previous post are far more relevant factors to consider than just the sensor's pixel pitch.

[lester_wareham]

It seems to me that if the Nikon and Canon both have the same number of pixels, and the Nikon sensor is smaller then the Nikon should have smaller sensors translating at least theoretically into greater resolution. (talk about pixel peeping!)

DrGary

Best to talk about:

a) sample rate, pixels per mm; important if you have to crop.

 image size: number of pixels; important if you don't need to crop.

c) sensor size; effects noise and required lens optical resolution.

Mixed in with all this is the trade-off with pixel size/noise and sample rate/anti-alias filter strength (the stronger AS filter the more sharpening artifacts).

To me resolution is the system MTF and noise characteristics not how many pixels.