Please help noob understand how MP's effect sharpness.

digitalman4242

Senior Member
Messages
1,782
Reaction score
372
Location
US
I keep reading posts regarding D800 vs D700. Some say the NR on the D800 isn't as good or that it may not be as sharp because it has 36 MP's instead of 12. Then some people will say that if you downsize the D800 photo to 12 mp the NR and the sharpness will be as good or better then the D700. When they say downsize do they mean in your software take the resolution down, or in camera set it to a lower resolution?
 
People say all kinds of crazy things.

When they're talking about "downsampling" or "downsizing" an image, they're doing that in post production.
 
I keep reading posts regarding D800 vs D700. Some say the NR on the D800 isn't as good or that it may not be as sharp because it has 36 MP's instead of 12. Then some people will say that if you downsize the D800 photo to 12 mp the NR and the sharpness will be as good or better then the D700. When they say downsize do they mean in your software take the resolution down, or in camera set it to a lower resolution?
It would be interesting to simply crop the 800 image to 700 frame size, and then compare the two.
I think that the 800 would win for sharpness.
 
It doesn't.

You're confusing potential resolution of the sensor with sharpness.

Sharpness is a quality the image as a whole has when viewed as a whole. This has nothing to do with pixel level detail, which will be largely invisible in even a high quality print viewed at normal sizes.

If I hand you two identically shot 8x10 prints, one from a camera with say 10Mp and one from one with 36Mp you won't be able to tell which print camera from which camera.

Forget pixels and pixel counts.

--
StephenG
 
Sharpness is a quality the image as a whole has when viewed as a whole. This has nothing to do with pixel level detail, which will be largely invisible in even a high quality print viewed at normal sizes.

If I hand you two identically shot 8x10 prints, one from a camera with say 10Mp and one from one with 36Mp you won't be able to tell which print camera from which camera.

Forget pixels and pixel counts.
Probably it would be better to understand pixels and pixel counts as they relate to how prints look. Generally people agree that greater than about 300 Pixels Per Inch pixel density does not improve {resolution/sharpness/detail} in a printed image, and if we accept that as the "standard", we can draw a few relatively safe conclusions...

A 10 MP image with a 4:5 aspect ratio is 2828x3535 pixels. At 300 PPI that will print out at 9.4x11.8 inches. Hence any print smaller than 9x11 will not appear sharper. But as the size is increased from 10x12 on up the sharpness will decline.

A 36 MP image with a 4:5 aspect ratio is 5373x6716 pixels. At 300 PPI that will print out at 17.8x22.4 inches Hence any print smaller than 17x22 will not appear sharper. But as the size is increased from 18x22 on up the sharpness will decline.

Put those two paragraphs together and it is clear that if 8x10 prints are made, then greater than 10 MP will not produce a sharper image, but for images significantly larger than 8x10, for example 16x20, the larger 36 MP image will be visibly sharper.

How much sharper depends on a number of things though, as of course the scene photographed may not have detail fine enough to make any difference! But assuming it does, and the quality of the entire process is high, at some point between 10 MP and 36 MP the differences will become quite visible in 16x20, or larger, prints.

And, for uncropped images, that is just about exactly the significant difference between shooting with a D700 (perfect for a bit larger than 8x10 or smaller) and a D800 (perfect for a bit large than 16x20 or smaller). Of course another consideration is that it requires a full frame D700 image to crop it to a 4:5 aspect ratio to get a high enough pixel density for 8x10 prints, while with a D800 there are about 4 each 8x10 prints that can be cropped out of each frame! That may not be significant at all to a studio photographer, but for many other subjects (sports and wildlife, for example) the ability to crop may be very significant.
 
Probably it would be better to understand pixels and pixel counts as they relate to how prints look. Generally people agree that greater than about 300 Pixels Per Inch pixel density does not improve {resolution/sharpness/detail} in a printed image, and if we accept that as the "standard", we can draw a few relatively safe conclusions...

A 10 MP image with a 4:5 aspect ratio is 2828x3535 pixels. At 300 PPI that will print out at 9.4x11.8 inches. Hence any print smaller than 9x11 will not appear sharper. But as the size is increased from 10x12 on up the sharpness will decline.

A 36 MP image with a 4:5 aspect ratio is 5373x6716 pixels. At 300 PPI that will print out at 17.8x22.4 inches Hence any print smaller than 17x22 will not appear sharper. But as the size is increased from 18x22 on up the sharpness will decline.

Put those two paragraphs together and it is clear that if 8x10 prints are made, then greater than 10 MP will not produce a sharper image, but for images significantly larger than 8x10, for example 16x20, the larger 36 MP image will be visibly sharper.

How much sharper depends on a number of things though, as of course the scene photographed may not have detail fine enough to make any difference! But assuming it does, and the quality of the entire process is high, at some point between 10 MP and 36 MP the differences will become quite visible in 16x20, or larger, prints.

And, for uncropped images, that is just about exactly the significant difference between shooting with a D700 (perfect for a bit larger than 8x10 or smaller) and a D800 (perfect for a bit large than 16x20 or smaller). Of course another consideration is that it requires a full frame D700 image to crop it to a 4:5 aspect ratio to get a high enough pixel density for 8x10 prints, while with a D800 there are about 4 each 8x10 prints that can be cropped out of each frame! That may not be significant at all to a studio photographer, but for many other subjects (sports and wildlife, for example) the ability to crop may be very significant.
+1 on your general advice, with one addendum: Printing densities of greater than 300DPI are not visible when observed from a viewing distance of 1 foot based on the minimum resolvable angle of view of the human eye. As you step away from the image, this density drops; as you move towards it, this density increases. the rule of thumb is DPI=300/(viewing distance in feet).

The OP also had some questions pertaining to downsizing images and noise levels. The basic idea is this: overall image noise level is primarily related to the total sensor area for sensors built on the same processing technology. Intra sensel wiring will compromise this proportionality, but almost insignificantly. Therefore, there should be no difference in final image quality when images from two different resolution sensors of the same area are rendered for viewing at the same final resolution - for example, when a 36MP image from an FX sensor is downsized to 12MP and compared to a picture taken with a 12MP FX sensor. There are gotchas here and there having to do with the quality of the resampling algorithm used in the downsizing, but if one were to take a picture from a D800 and another from a D700 and downsize both to the 2MP resolution of your typical HDTV monitor, you should see no difference in the sharpness or noise levels other than that related to the differences in the bulk performance of the sensor process technologies and imaging chain quality used in each camera. Every aspect of the imaging chain has changed in the 5 years between the D700 and the D800. Ascribing image quality differences solely to sensor resolution is incorrect.
 
if one were to take a picture from a D800 and another from a D700 and downsize both to the 2MP resolution of your typical HDTV monitor, you should see no difference in the sharpness or noise levels other than that related to the differences in the bulk performance of the sensor process technologies and imaging chain quality used in each camera. Every aspect of the imaging chain has changed in the 5 years between the D700 and the D800. Ascribing image quality differences solely to sensor resolution is incorrect.

One shouldn't ever downsize to 2MP for viewing on an HDTV unless space is at a premium. About a 4MP jpg seems to be the sweet spot where going bigger gets you nothing. And yes, the difference between 12, 16 and 36MP is visible. Not night and day, to be sure, but it's there.
 
First of all sharpness, and resolution are not the same thing. I don't see why sharpness would be any better on either camera. Secondly if you cut the D800 sensor to the D7000's size your resolution should be roughly the same.
 
if one were to take a picture from a D800 and another from a D700 and downsize both to the 2MP resolution of your typical HDTV monitor, you should see no difference in the sharpness or noise levels other than that related to the differences in the bulk performance of the sensor process technologies and imaging chain quality used in each camera. Every aspect of the imaging chain has changed in the 5 years between the D700 and the D800. Ascribing image quality differences solely to sensor resolution is incorrect.

One shouldn't ever downsize to 2MP for viewing on an HDTV unless space is at a premium. About a 4MP jpg seems to be the sweet spot where going bigger gets you nothing. And yes, the difference between 12, 16 and 36MP is visible. Not night and day, to be sure, but it's there.
So, Reilly, I know you've recently switched from a D7000 to a D800e. I'm a little confused by your statement about downsizing, since when one views an image on the typical monitor it is usually downsizing and resampling the original to obtain the final image size if not viewed at 100%. It is true that a 30" photo editing monitor has a native resolution of 4MP, from which I can see the 4MP sweet spot claim for rendered originals. But my comment was by way of example in any case.

I don't doubt that you're seeing differences between downrezzed images taken with various cameras. I wonder to what extent you're seeing just the effects of pixel density vs. differences in sensor processing technology, AA filter strength, resampling algorithm performance, etc.. The D7000 and D800 have roughly similar pixel densities, but different sensor technologies; the D700 and D800e have both different densities and technologies. It begs the question of whether the higher resolution sensor provides more spatial data from which to derive cleaner edge transitions, which could translate into higher contrast ratio edges in the downrezzed rendering, yielding a higher perceptual sharpness.
 
It begs the question of whether the higher resolution sensor provides more spatial data from which to derive cleaner edge transitions, which could translate into higher contrast ratio edges in the downrezzed rendering, yielding a higher perceptual sharpness.
I don't have a good handle on any of that, just what I see experimenting with different renderings from the different files on a 55" LED, my preferred delivery system. The FX sensor without standard filter definitely provides airy, unforced detail unadulterated by sharpening. 100% quality setting seems necessary to avoid compression artifacts.
 
Sharpness is a quality the image as a whole has when viewed as a whole. This has nothing to do with pixel level detail, which will be largely invisible in even a high quality print viewed at normal sizes.

If I hand you two identically shot 8x10 prints, one from a camera with say 10Mp and one from one with 36Mp you won't be able to tell which print camera from which camera.

Forget pixels and pixel counts.
Probably it would be better to understand pixels and pixel counts as they relate to how prints look. Generally people agree that greater than about 300 Pixels Per Inch pixel density does not improve {resolution/sharpness/detail} in a printed image, and if we accept that as the "standard", we can draw a few relatively safe conclusions...

A 10 MP image with a 4:5 aspect ratio is 2828x3535 pixels. At 300 PPI that will print out at 9.4x11.8 inches. Hence any print smaller than 9x11 will not appear sharper. But as the size is increased from 10x12 on up the sharpness will decline.

A 36 MP image with a 4:5 aspect ratio is 5373x6716 pixels. At 300 PPI that will print out at 17.8x22.4 inches Hence any print smaller than 17x22 will not appear sharper. But as the size is increased from 18x22 on up the sharpness will decline.

Put those two paragraphs together and it is clear that if 8x10 prints are made, then greater than 10 MP will not produce a sharper image, but for images significantly larger than 8x10, for example 16x20, the larger 36 MP image will be visibly sharper.

How much sharper depends on a number of things though, as of course the scene photographed may not have detail fine enough to make any difference! But assuming it does, and the quality of the entire process is high, at some point between 10 MP and 36 MP the differences will become quite visible in 16x20, or larger, prints.

And, for uncropped images, that is just about exactly the significant difference between shooting with a D700 (perfect for a bit larger than 8x10 or smaller) and a D800 (perfect for a bit large than 16x20 or smaller). Of course another consideration is that it requires a full frame D700 image to crop it to a 4:5 aspect ratio to get a high enough pixel density for 8x10 prints, while with a D800 there are about 4 each 8x10 prints that can be cropped out of each frame! That may not be significant at all to a studio photographer, but for many other subjects (sports and wildlife, for example) the ability to crop may be very significant.
Im using a D3200 (History notes: we also have a D800, and I've just sold my Canon 1Ds3).
Its as most people know the 3200 is a 24mp camera (not full frame).

I get a (approx) 13in x 20in @ 300 dpi which most of my clients are very happy with, it also gives me a 68mp file size.....which is bigger than my Canon 1Ds3 did.
How can this be so.
 
Im using a D3200 (History notes: we also have a D800, and I've just sold my Canon 1Ds3).
Its as most people know the 3200 is a 24mp camera (not full frame).

I get a (approx) 13in x 20in @ 300 dpi which most of my clients are very happy with, it also gives me a 68mp file size.....which is bigger than my Canon 1Ds3 did.
How can this be so.
The D3200 produces a 6016x4000 pixel image. As you note (and Nikon lists on page 46 of the D3200 user manual) that works out to about 13.3"x21.1" when printed at 300 PPI.

I'm not sure what kind of a file you are referencing as "a 68mp file size", so it's hard to compare that to something else.
 
Image Resolution: Number of Pix per Inch, not related to

Image File size on disk or...

Sensor Resolution: Number of photonic wells (pix say) on the sensor

More sensor Pix = sharper photo in perfect world scenario but not a higher dynamic range (generally)

However, that scenario is rare and in some cases the higher resolution will cause image to be worse if the original taking of the photo was not optimal in the first place.
 
Im using a D3200 (History notes: we also have a D800, and I've just sold my Canon 1Ds3).
Its as most people know the 3200 is a 24mp camera (not full frame).

I get a (approx) 13in x 20in @ 300 dpi which most of my clients are very happy with, it also gives me a 68mp file size.....which is bigger than my Canon 1Ds3 did.
How can this be so.
The D3200 produces a 6016x4000 pixel image. As you note (and Nikon lists on page 46 of the D3200 user manual) that works out to about 13.3"x21.1" when printed at 300 PPI.

I'm not sure what kind of a file you are referencing as "a 68mp file size", so it's hard to compare that to something else.
When i look at a picture on my screen, ...at the bottom (left hand side) it sez Doc: 68M...i presume that means the file size.
Cheers
George
 

Keyboard shortcuts

Back
Top