Pages: [1]   Bottom of Page
 Author Topic: "resolution 200 lp/mm" means what?  (Read 2885 times)
Dynszis
Newbie

Offline

Posts: 15

 « on: September 03, 2012, 12:44:53 PM » Reply

Hello, all.

If a lens is advertised as having a resolution of 200 lp/mm, does this mean: 200 line pairs per millimeter sensor height?

I'm asking because today I stumbled upon C-mount lenses that were advertised as such, and that in principle, might be interesting to mount on to a Pentax Q via adapter.

However, if the above conjecture is correct, then 200 lp/mm would mean less than 1000 line pairs per picture height on a 1/2.3" sensor, and that resolution would seem anything but seductive.

Does anyone know whether I got this right?
 Logged
Ray
Sr. Member

Offline

Posts: 8947

 « Reply #1 on: September 03, 2012, 07:57:15 PM » Reply

Here's may take on it, for what it's worth.

A millimetre is a millimetre and a line is a line, whatever the lens and whatever the sensor.
However, lines on test targets can vary in width and contrast, often being different shades of grey ranging from black against white to a barely discernible grey against a slightly paler shade of grey. It is therefore necessary to know the MTF (or contrast) of the resolution specified, if the resolution figure is to be meaningful.

Because a line tends to alternate between black and white, it's possible to get the impression that a line is a black marking on a white background and that the space in between the black lines are not lines. For this reason one should distinguish between "line pairs" and "lines".  A "line pair" consists of one dark line and the adjacent lighter line, whereas a single line can be either light or dark.

It is possible to extrapolate such a result to the full height of a sensor. Such a result is usually referred to as "Line Widths per Picture Height", or LW/PH. However, a 'line width' is a single line, not a line pair. The specification you quote of 200 lp/mm (ie. 200 line pairs/mm) is 400 line widths per mm. Therefore a 1/2.3" sensor with dimensions of 6.16x4.62mm would have a resolution of 1848 LW/PH with such a lens.

However, as I understand, such a result is an extrapolation. If you were to photograph a test chart consisting of 1848 lines, top to bottom, and fill the whole chart in the viewfinder, you wouldn't be able to count all the lines. Many of the lines, towards the top edge and bottom edge would be too blurred to discern. This is why lens test sites such as Photozone usually provided three sets of figures for LW/PH. One for the centre resolution, one for the borders, and one for the extreme edges, all at a 50% MTF contrast.
 Logged
Dynszis
Newbie

Offline

Posts: 15

 « Reply #2 on: September 04, 2012, 03:24:01 AM » Reply

Hello Ray,

thank you for taking the time to reply.

I guess the upshot is that, since I confused line width and line pairs, it's literally not half as bad as I thought

Of course a figure of 1848 LW/PH is ultimately meaningless as long as we don't know whether that refers to MTF50 or something else, but it surely sounds reasonable enough to try to find out what that looks like in real life.
 Logged
viking79
Newbie

Offline

Posts: 1

 « Reply #3 on: January 23, 2013, 12:29:32 PM » Reply

Did you ever get a chance to try one of these 200 lp/mm lenses?  I have been looking at them for the Q as well.  Too bad the Q isn't 2/3" sensor

I want to try one of them but they are \$800 to \$1000 for an f/1.8 to f/2.8 lens depending on focal length.

I want to say that most of those are listed at MTF50, so MTF10 or MTF20 might be better than that, but don't know what the physical limits are for lenses.

Eric
 Logged
bjanes
Sr. Member

Offline

Posts: 2882

 « Reply #4 on: January 23, 2013, 01:24:33 PM » Reply

Hello, all.

If a lens is advertised as having a resolution of 200 lp/mm, does this mean: 200 line pairs per millimeter sensor height?

200 lp/mm means just that. Resolution per picture height is the resolution in lp/mm times the picture height in mm. For a larger picture height, the lens must have sufficient covering power.

Resolution by itself is not that helpful, and the contrast at that resolution must be stated. Roger Clark has posted a table showing resolution limits for a perfect lens at various levels of MTF. If one uses 50% contrast as a criterion for resolution, 200 lp/mm can be obtained only at apertures larger than f/4.

Regards,

Bill

 Logged
ErikKaffehr
Sr. Member

Offline

Posts: 8028

 « Reply #5 on: January 23, 2013, 11:16:09 PM » Reply

Hi,

Actually, quite a few lenses reach 200 lp/mm but at very low contrast.

MTF measurements with a camera/lens system highly depend on test methodology. I can produce 1400 or 4500 LWPH (at 50% MTF) from my lenses, depending on sharpening, I'm mostly using either LR 4.3 standard sharpening (3000 LW/PH) or sharpening set to zero (1400 LW/PH).

As long as all parameters are the same MTF is a valid comparison. The MTF normally given by vendors is either measured, calculated or designed by marketing departement. Any MTF that is 100% is not measured om MTF equipment.

Photozone publishes a lot of comparable MTF measurements, best source of information. I would say Photozone figures are somewhat comparable between different systems.

Best regards
Erik

Hello, all.

If a lens is advertised as having a resolution of 200 lp/mm, does this mean: 200 line pairs per millimeter sensor height?

I'm asking because today I stumbled upon C-mount lenses that were advertised as such, and that in principle, might be interesting to mount on to a Pentax Q via adapter.

However, if the above conjecture is correct, then 200 lp/mm would mean less than 1000 line pairs per picture height on a 1/2.3" sensor, and that resolution would seem anything but seductive.

Does anyone know whether I got this right?

 Logged

xpatUSA
Sr. Member

Offline

Posts: 305

 « Reply #6 on: January 24, 2013, 10:20:33 AM » Reply

The lens resolution's meaning is as stated in the previous posts and is also 2.5um, implying next to no diffraction! But manufacturers are said to be coy about reveal the the MTF at that resolution. Falk says it can be anywhere between 3 and 10% - see here: http://www.falklumo.com/lumolabs/articles/sharpness/ and he recommends that you assume 3%. To get an idea of what 3% means, 200 lp/mm lines which should show as the equivalent of 0 and 255 image levels would instead appear as 124 and 134 (neglecting gamma correction for the minute). At 10%, maybe 115 and 140. That estimate does not include the sensor's AA filter, BTW.

 Logged

best regards,

Ted
 Pages: [1]   Top of Page