Camera Labs

Hello, thank you for the very informative site!

I had two questions regarding resolution testing, see the following link:

Canon5D etc link.

You've profiled each camera's respective resolution capability with a resolution test chart, however I don't know how you picked the particular resolutions, as the lines get rather undefined. Is there a standardized protocol for how much "fuzzyness" is permitted before you demarcate the resolution?

Also, might you know how to relate lines per picture height (lpph) to line pairs per millimeter (lp/mm)? For instance, what would 2000 lpph be equal to in terms of line pairs per millimeter.

Thank you

Hello Capital, welcome to the forums - and good question!

Since there's no standard software to analyse these resolution charts, its down to the individual to read and interpret them.

There's three main things to watch out for: first the point where the lines are still definitely seperate, second the point where moire begins to blur them, and third where the lines can no longer be seperated.

Of course a lot of this interpretation is down to the individual, which is why two different people or websites using the same charts may conclude different results. The important thing is that each person is consistent so at least their own results can be compared.

I take the result from the point just before moire creeps in, so it's at the point where the lines are still discernably seperate.

There's a whole science behind resolving power and analysis and for more information you may wish to visit the site of the company which supplies our charts, Applied Image.

There was also an earlier thread about this here:

http://www.cameralabs.com/forum/viewtop ... =line+pair

Gordon

Gordon,

Thank you for your reply and link.

I guess to convert line pairs per picture height to line pairs per millimeter, one must know the height of the photo sensor?

So for a full frame sensor of 36x24 mm and supposing a resolution tested of 2000 line pairs per picture height, that the equivalent lp/mm is then:

2000 line pairs.....1 picture height
-----------------....x ---------------------....~....83 lp/mm
picture height..........24mm


However, if the test is performed on a camera/sensor rotated 90 degrees, wouldn't one also pull from the resolution chart a value of 2000 lpph?

Thereby yielding a significant decrease in apparent lp/mm? (i.e. 55.6 lp/mm versus 83 lp/mm) Something appears to be amiss.


Another question I had is at what distance do you place the focal plane from the test chart as you tested both 50mm and 85mm lenses.

Thank you.

Hi Capital, in answer to your distance question, we don't test at a fixed distance. We test at a fixed focal length, so since the chart stays the same size, we have to adjust the distance between the camera and it so it correctly fills the frame.

So the test with the 85mm was taken further away than the one at 50mm.

FYI we test most DSLRs at 35mm becuase with the crop factor that works out close to the old 50mm standard, and it's also roughly halfway through the typical 18-55mm kit ranges.

Gordon

I think the resolution charts are getting less and less relevant.
I'm always smiling when I turn to Gordons page on corner sharpness to see how smudged some lenses are in the edge at full open. This is much more critical than the last 100 lines in the resolution charts...
I would be content seeing example crops at infinity full open vs. F11 from corner/center plus example crops at close distance (10:1 or even 5:1) full open vs. F11. Because my experience is that the pic quality degrades quite a bit when focussing closer
The only really frustrating thing that comes up when you get near the limits of resolution is moire: So I would suggest when testing a body to show a shot at F8/center of a Siemens-star as this is purely a thing of the sensor and the processor in a body.

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Lens Reviews, My Pictures, My Photography Blog
D800+assorted lenses