Home

Please support mflenses.com if you need any graphic related work order it from us, click on above banner to order!
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

How to calculate sensor spatial resolution?
Post new topic   Reply to topic
Home   Manual Focus Lenses
View previous topic :: View next topic  


PostPosted: Fri Mar 06, 2015 2:54 pm    Post subject: How to calculate sensor spatial resolution? Reply with quote
Hi folks

I'm trying to work out, in line pairs per mm the spatial resolution of a sensor.

This sensor produces images 4000 pixels wide and is 7.6mm wide.

4000 pixels/2 gives us line pairs, correct? That would be 2000lp.

Therefore, to get lp per mm, we would divide 2000 by 7.6, correct?

2000/7.6 = 263

So is the correct answer 263lp/mm?

Seems awfully high to me.


PostPosted: Fri Mar 06, 2015 3:25 pm    Post subject: Reply with quote
I think so.
http://photo.blogoverflow.com/2012/06/the-realities-of-resolution/


PostPosted: Fri Mar 06, 2015 3:38 pm    Post subject: Re: How to calculate sensor spatial resolution? Reply with quote
iangreenhalgh1 wrote:
Hi folks

I'm trying to work out, in line pairs per mm the spatial resolution of a sensor.

This sensor produces images 4000 pixels wide and is 7.6mm wide.

4000 pixels/2 gives us line pairs, correct? That would be 2000lp.

Therefore, to get lp per mm, we would divide 2000 by 7.6, correct?

2000/7.6 = 263

So is the correct answer 263lp/mm?

Seems awfully high to me.


Yes, your calculation is correct. The sensor you're describing has a 1.9um pixel pitch, which is pretty small, hence the high calculated resolution.

Note however that the actual resolution is less if the sensor has an anti aliasing filter. Typically this cuts resolution by 50% or more, sometimes 100%. Plus the demosaicing algorithms don't produce sharp images at pixel level so the algorithm itself limits resolution. Don't expect to be able to clearly see 263 line pairs on a test pattern with the sensor, it just won't happen.


PostPosted: Fri Mar 06, 2015 3:46 pm    Post subject: Reply with quote
calvin83 wrote:
I think so.
http://photo.blogoverflow.com/2012/06/the-realities-of-resolution/


That is an excellent article. Thanks for posting the link, I had not seen it before.


PostPosted: Fri Mar 06, 2015 4:41 pm    Post subject: Reply with quote
That resolution also only refers to vertical/horizontal lines/mm. For diagonals you should expect it to be less.


PostPosted: Fri Mar 06, 2015 4:50 pm    Post subject: Reply with quote
Thanks Ian & Calvin83


PostPosted: Fri Mar 06, 2015 6:36 pm    Post subject: Reply with quote
I had about 100lppm in memory as the max working resolution of a camera, seems to be correct from that article.
Well there still are lenses that outresolve that... not for long as sensor development seem to move faster than
lens development ... Wink


PostPosted: Fri Mar 06, 2015 7:03 pm    Post subject: Reply with quote
Cheers guys. The sensor in question is that of my Pentax Q7 which has no AA filter.

I was trying to work out what a lens would have to resolve to be able to outresolve the sensor.


PostPosted: Fri Mar 06, 2015 7:53 pm    Post subject: Reply with quote
kds315* wrote:
I had about 100lppm in memory as the max working resolution of a camera, seems to be correct from that article.
Well there still are lenses that outresolve that... not for long as sensor development seem to move faster than
lens development ... Wink


Klaus...none of the cameras discussed have CCD sensors with the kind of resolution Ian is working with. If you do the same type of analysis, using lenses designed to have large diffraction-limited apertures, you'll find that the 100lp/mm moves into the 180-200lp territory.


iangreenhalgh1 wrote:
Cheers guys. The sensor in question is that of my Pentax Q7 which has no AA filter.

I was trying to work out what a lens would have to resolve to be able to outresolve the sensor.


I prefer looking at the problem from the pixel pitch / Airy disk perspective. So what you need is a lens that is diffraction-limited at f/2.8 or larger to out-resolve your sensor. This is the DLA for 1.9um pixel pitch. In practice you won't see much degradation at probably f/3.3 or even f/4 depending on the quality of the Pentax demosaicing algorithm. The DLA is the point at which no visible degradation occurs because the the first null of the Airy disk falls on the center of the adjacent pixels (Rayleigh limit). You have to go further to see degradation.


PostPosted: Fri Mar 06, 2015 9:00 pm    Post subject: Reply with quote
Cheers Ray. Not sure I understand fully, but it's something to mull over.

This shot was one that got me thinking about the resolution. It was taken with a Computar 1.3/8.5mm lens wide open. Notice the hairs in front of the girl's face and how the lens has resolved them, so it must be a pretty sharp little lens.



PostPosted: Fri Mar 06, 2015 9:46 pm    Post subject: Reply with quote
iangreenhalgh1 wrote:
Cheers Ray. Not sure I understand fully, but it's something to mull over.

This shot was one that got me thinking about the resolution. It was taken with a Computar 1.3/8.5mm lens wide open. Notice the hairs in front of the girl's face and how the lens has resolved them, so it must be a pretty sharp little lens.



Ian...is this a 100% crop or is it a downsized full-sensor image? It appears to be a crop because it is an odd size, 1500x999. If so, then the resolution does seem very good.

Note that there is little advantage from lens perspective to shoot wide open vs stopping down to the DLA (in spite of what the resolution article says) especially for sensors without AA filters. So what I'm saying is that you will get essentially the same performance shooting at f/1.3, f/2, f/2.8, or possibly even f/3.3 from system resolution perspective. You'll get shallower DOF at f/1.3, but the parts of the image that are in critical focus will have about the same sharpness up to f/2.8 or possibly f/3.3. Unless you're trying for shallow DOF, it is usually better to stop-down to the camera DLA. I learned this empirically after having the AA filter removed from my T2i, and found that I could easily excite false colors if the aperture was wider than the DLA. Also, the lens itself is probably sharper at f/2.8 than it is at f/1.3 due to aberrations, so the final image may actually be sharper.


PostPosted: Fri Mar 06, 2015 10:45 pm    Post subject: Reply with quote
Well then thanks for confirming I'm not overthinking this whole photography thing. Very Happy


PostPosted: Fri Mar 06, 2015 11:16 pm    Post subject: Reply with quote
Hi Ray

It's a 4000x3000 RAW image downsized then saved as an 80% JPEG, no other processing.

It was shot on a dark train at ISO 3200, f1.2, just to see if it worked.
Post new topic   Reply to topic
Home   Manual Focus Lenses
Page 1 of 1