Sharpness

[george013]

Pixels is a common unit of measure for edge spread response, George. QuickMTF is the same; easily converted into other units - as a matter of fact QuickMTF does it for you ...



That's life in the big city of resolution, sorry, George.





Nope. The sensor could be replaced by film or even a piece of toilet paper; the sharpness of the lens will not be affected one little bit!

This might help:

http://www.falklumo.com/lumolabs/art...ess/index.html

I refer to the DxO lens tests. https://www.dxomark.com/Lenses/Sigma...Nikon-D70__197
Just an example. Switch the camera and the sharpness of the lens is changed. DxO says somewhere they test the camera and that is lens and body. The lower the pixel density, the less sharp the lens will be.
In this example I did compare this lens with a D700,12.1Mpixel, and a D750,24.9Mpixel. About twice as much on a nearly equal sensor. In the above test the sharpness for this lens on a D700 is 9P-Mpix, on a D750 it's 19P-Mpix.


George

[xpatUSA]

I refer to the DxO lens tests. https://www.dxomark.com/Lenses/Sigma...Nikon-D70__197
Just an example. Switch the camera and the sharpness of the lens is changed. DxO says somewhere they test the camera and that is lens and body. The lower the pixel density, the less sharp the lens will be.
In this example I did compare this lens with a D700,12.1Mpixel, and a D750,24.9Mpixel. About twice as much on a nearly equal sensor. In the above test the sharpness for this lens on a D700 is 9P-Mpix, on a D750 it's 19P-Mpix.

George

To me, "P-Mpix" is an artificial single-number measure introduced by DxO to oversimplify comparisons. It has nothing to do with the sharpness of a lens which the rest of the world measures with MTF diagrams. Sorry.

[dje]

I tend to take a "holistic" view of the terms sharpness and resolution.

Acutance is often defined by the spatial "rise time" of a knife edge. This shape is referred to as the Edge Spread Function and can be measured using the Slanted Edge technique with a camera. The derivative of the ESF is the Line Spread function or LSF. Two dimensionally this becomes the Point Spread Function. The FFT (Fast Fourier transform) of the PSF is the MTF (Modulation Transfer Function). This function has x axis units of pixels/cycle cycles/pixel**, or if you use the pixel pitch of the sensor, it can be expressed as Line Pairs per mm. So it's all the same thing, you are just looking at spatial distance vs spatial frequency (similar to Electrical engineering analyses with time and freqency rather than spatial distance and frequency).

The MTF of lenses alone can be measured with other techniques using an optical bench but these are beyond the means of all but the specialised optics labs.

I have seen reference to the resolving power of a lens defined as the LPPMM for which the MTF is 5%. One common notion for the term "Sharpness" is how sharp an image looks to the human eye under certain viewing conditions. One measure of this is the MTF weighted and integrated over the MTF range. Weighting peaks around 30 lppmm if I remember correctly.This is the concept of SQF or subjective quality factor. I believe DXO Mark uses something like this and calls it acutance.

And on it goes...

Dave

Edit : Correction above, should be cycles/ pixel rather than pixels/cycle.

[Abitconfused]

I have been happy with the sharpness and light gathering power of the Sigma 20mm f/1.4



For some images, sharpness adds that certain savoir faire (593 KB) f/1.6, /1/125s, ISO 800

[Abitconfused]

And, of course, the apparent sharpness of an image is dependant upon post-production factors. Here DXO's Prime manages 800 ISO quite nicely on the 20mm at f/1.4, 1/160s

[Abitconfused]

Obvious over sharpening doesn't really help us...



I think...

[george013]

I tend to take a "holistic" view of the terms sharpness and resolution.

Acutance is often defined by the spatial "rise time" of a knife edge. This shape is referred to as the Edge Spread Function and can be measured using the Slanted Edge technique with a camera. The derivative of the ESF is the Line Spread function or LSF. Two dimensionally this becomes the Point Spread Function. The FFT (Fast Fourier transform) of the PSF is the MTF (Modulation Transfer Function). This function has x axis units of pixels/cycle, or if you use the pixel pitch of the sensor, it can be expressed as Line Pairs per mm. So it's all the same thing, you are just looking at spatial distance vs spatial frequency (similar to Electrical engineering analyses with time and freqency rather than spatial distance and frequency).

The MTF of lenses alone can be measured with other techniques using an optical bench but these are beyond the means of all but the specialised optics labs.

I have seen reference to the resolving power of a lens defined as the LPPMM for which the MTF is 5%. One common notion for the term "Sharpness" is how sharp an image looks to the human eye under certain viewing conditions. One measure of this is the MTF weighted and integrated over the MTF range. Weighting peaks around 30 lppmm if I remember correctly.This is the concept of SQF or subjective quality factor. I believe DXO Mark uses something like this and calls it acutance.

And on it goes...

Dave

I tend to think that what is meant by sharpness of a lens is actually the sharpness of a system, minimal body and lens as said in DxO. The pixel pitch of the D700 is twice the pixel pitch of the D750, 7.11E-5 and 3.46E-5.
If pixels/cycle is used and pixels is a natural number meaning a whole number and the minimum is 1 then there is a difference in max sharpness for those two systems using the same lens.

I might be wrong, but I thought sharpness with analogue films is related to the film quality.

George

[dje]

If pixels/cycle is used and pixels is a natural number meaning a whole number and the minimum is 1 then there is a difference in max sharpness for those two systems using the same lens.


George

Sorry George, I had that around the wrong way. Should be cycles/pixel, not pixels per cycle. But yes it is the sharpness of the lens/camera system that is being measured and the conversion from cycles/pixel to lppmm depends on the pixel pitch.

Dave

[george013]

Sorry George, I had that around the wrong way. Should be cycles/pixel, not pixels per cycle. But yes it is the sharpness of the lens/camera system that is being measured and the conversion from cycles/pixel to lppmm depends on the pixel pitch.

Dave

I didn't see that typo. I just meant to say that as soon pixels are introduced a non-optical factor is introduced. Pixels are no optical elements and can be different in size.
So the conclusion can be made that pixel density, pixel pitch, is a limiting factor of the sharpness of the lens. What I wrote in post 19.

That makes it difficult to compare lenses when the tests are done with different camera's. Not only for sharpness but also for ca.

George

[xpatUSA]

I tend to think that what is meant by sharpness of a lens is actually the sharpness of a system, minimal body and lens as said in DxO. The pixel pitch of the D700 is twice the pixel pitch of the D750, 7.11E-5 and 3.46E-5.

Just as your definition of the sharpness of a lens differs from the norm, it appears that your D700 is also different.

Your D700's pixel pitch: 71.1um - everyone else's: 8.45um.

https://www.dpreview.com/reviews/nikond700/2
.

[DanK]

I tend to think that what is meant by sharpness of a lens is actually the sharpness of a system,

That introduces needless confusion, since "X of a lens" refers to a property of a lens, and "X of a system that includes a lens" refers to a property of the system that includes the lens. If you insist on using "sharpness of a lens" to mean "sharpness of a system that includes the lens," you shouldn't be surprised that others either disagree or are confused.