Need help explaining the difference between resolution and pixel desnity

cwinhall

cwinhall

Suspended / Banned
Messages
571
Name
Colin
Edit My Images
Yes
Hi everyone,

So I'm struggling to explain the difference in digital resolution when comparing a single 30mp image and a 30mp image, stitched from multiple lower res images.

For example;

Say you take a single photo with a 30mp camera at 18mm of a square metre of grass.

Vs

Multiple images at 5mp with a macro lens (say 100mm) of the same square metre of grass. Then stitched together.

Having done this, I know that the stitched image produces a much higher quality image when zooming in as far as possible.

I am struggling to explain why this is the case though. Does it come down to pixel density and optical zoom over digital zoom? Also not neglecting the fact that there is less distortion on the stitched image.

I hope I have made sense in this? If not I will try again... lol

Thanks for the help.
 
If the final size of the two images is the same, both will have the same resolution and detail.


Steve.
 
Steve Smith said:
If the final size of the two images is the same, both will have the same resolution and detail.


Steve.
Click to expand...


Nope.



If you cover 100mm square with a 5MP image, you'll need 10 of these to cover the same square meter of grass you shot with a single 30MP image, so the quality is higher because your stitched image is 50MP. Also, as the grass it's made up from was shot closer, so probably was more detailed regardless of resolution due to being less lens quality dependent.

Hang on... wouldn't you need 100x 100mm/2 images to cover 1 metre/2? If so, then it will definitely be higher resolution, as that would make it 500MP.
 
Last edited:
So if you take a single 30mp image at wide angle or 30 1mp images with a macro lens. The final output will be the same quality?
 
the way i see it if we have a 30mp full frame sensor and a same size sensor with 5mp, the 5mp sensor would obviously have larger cell sites capable of gathering more information than each cell site on the 30mp sensor
maybe thats the reason why more detail can be seen
 
Convoluted question for a simple principal.

Think of a photosite as being a bucket that catches photons. A larger photosite will capture more photons, therefore a sensor with larger photosites gives more accurate representation of what's exposed to it. (All other things being equal)
 
lenses aren't infinitely sharp, lower density sensor is therefor easier :-)
 
Steve Smith said:
Yep.

By size, I meant image size in pixels.

If both are e.g. 6000 x 4000 pixels, both will show the same detail, all other things being equal.


Steve.
Click to expand...

That's not what he asked though. He asked if you make up a shot of a 1m/2 patch of grass from A) one 30MP exposure, and B) lots of 10cm/2, 5MP exposures, which would have the most detail. The answer is B. He didn't specify A final file size. The resulting image from B wouldn't be 6000x4000, so not sure where you're getting those figures from.
 
Last edited:
However, the OP stated:

"when comparing a single 30mp image and a 30mp image, stitched from multiple lower res images"

Which to me means the same size in terms of pixel size.


Steve.
 
cosmicma said:
the way i see it if we have a 30mp full frame sensor and a same size sensor with 5mp, the 5mp sensor would obviously have larger cell sites capable of gathering more information than each cell site on the 30mp sensor
maybe thats the reason why more detail can be seen
Click to expand...

This makes sense to me. Thanks.
 
This question has got far more to do with the effective physical size of the final image (at sensor level) than it has to do with pixels.

The question as posed doesn't quite stack up, as Pooks pointed out. It sounds more like the OP is creating, in effect, an image taken on a much larger format camera, but that isn't what's been said. But if that is actually what emerges in the final image, then the answer lies in lens MTF, ie bigger format = sharper.
 
HoppyUK said:
This question has got far more to do with the effective physical size of the final image (at sensor level) than it has to do with pixels.

The question as posed doesn't quite stack up, as Pooks pointed out. It sounds more like the OP is creating, in effect, an image taken on a much larger format camera, but that isn't what's been said. But if that is actually what emerges in the final image, then the answer lies in lens MTF, ie bigger format = sharper.
Click to expand...

Hi Hoppy,

What does Lens MTF mean? I probably havent worded it right, I'm well aware of that. Let me try again.

I am taking multiple photos of a small area of the ground on a unit that I have created. It uses http://www.ximea.com/en/products/usb3-vision-cameras-xiq-line/mq013mg-e2 this sensor with a 16mm lens attached. Essentially making the focal length 346.4mm (based off the 1/8" sensor format I believe?)

It takes 32 photos with about a 20% overlap. I should mention also that each image is shot at f/16 to get a high DOF.

Now I know from doing this, that it creates a far superior final image than a single shot dSLR, even if the dSLR produces higher resolution images.

I'm trying to work out the wording for WHY this is the case so I can tell my clients.
 
Last edited:
Steve Smith said:
However, the OP stated:

"when comparing a single 30mp image and a 30mp image, stitched from multiple lower res images"

Which to me means the same size in terms of pixel size.


Steve.
Click to expand...

In that case, you have a point, but I still reckon the stitched image would be superior because effectively it's using a "larger" sensor, so lens defects would be less noticeable. Larger sensors of the same resolution produce sharper images.
 
cwinhall said:
Hi Hoppy,

What does Lens MTF mean? I probably havent worded it right, I'm well aware of that. Let me try again.

I am taking multiple photos of a small area of the ground on a unit that I have created. It uses http://www.ximea.com/en/products/usb3-vision-cameras-xiq-line/mq013mg-e2 this sensor with a 16mm lens attached. Essentially making the focal length 346.4mm (based off the 1/8" sensor format I believe?)

It takes 32 photos with about a 20% overlap. I should mention also that each image is shot at f/16 to get a high DOF.

Now I know from doing this, that it creates a far superior final image than a single shot dSLR, even if the dSLR produces higher resolution images.

I'm trying to work out the wording for WHY this is the case so I can tell my clients.
Click to expand...

Lens MTF is Modulation Transfer Function. It's the way sharpness is measured, eg those squiggly graphs that Canon, Nikon, Sigma etc publish. Sharpness has two components, resolution (the fineness of detail) and contrast (how clearly those details are shown) and the two are inextricably linked. When resolution demands go up, so contrast goes down and the image looks less sharp (contrast is the most significant of the two in terms of visual sharpness). Note with those graphs that the plots at 10-lines-per-mm are always higher than those at 30-lpmm.

In practise, this means that smaller sensor formats makes the lens work harder (deliver greater resolution) because the image has to be enlarged more for same size final output. This is why FF is sharper than APS-C, and it has very little to do with pixels (within reason).

The information you've given is not complete, and some of the things you've said are not right. For example, what is this 30mp DSLR that your comparing to, and what is the physical image size? APS-C? Also, I think you mean your device has a 1/1.8in sensor which would give a crop factor of roughly 5x (relative to FF) so in those terms your 16mm lens has a field of view equivalent of 80mm-ish. Then you've cropped it...

Making a few assumptions, I think your final composited 32-frame image is close to FF in terms of area, and in the 30mp kind of zone, but you could easily work this out exactly by taking your final composited image pixel dimensions x pixel pitch (5.3 microns according to your link). If all that is right or thereabouts, then lens MTF will be a significant factor, but there are a couple of other things contributing. One is that when using these kinds of equivalence comparisons, it's okay when comparing say FF and APS-C or even M4/3 formats, but with a 5x crop factor that's really pushing it. The other is diffraction, and as you say you're using f/16 this will be capping sharpness.

Bottom line, if all this is somewhere right, then it's lens mainly lens MTF that is making the composited image appear sharper, plus a few unknowns. Basically as I and some other posters guessed at.
 
You must log in or register to reply here.
Back
Top