Printable View
Hi all,
As i couldnt find any satisfactory information on this, maybe someone here knows...
How do image circle and lens resolution relate to each other? Do lenses with very large image circles necessarily have poor resolution?
Also, in digital application, do the digital LF lenses make much of a difference compared to high quality LF film lenses (mostly around F11 to F22)? Does anyone have experience with this?
Dave
You are probably speaking about the new "digital" view camera lenses versus the classical versions?
Well, resolution is not the only issue here.
Some residual defects named lateral color are enhanced by silicon sensors using the bayer style of color filter pattern.
Hence, manufacturers had to re-compute their lenses to reduce those defects for digital use.
When computing a lens, you have a trade-off between wide angular coverage and all other parameters that you want to optimize.
So it can be saif that new "digital" view camera lenses improve the performance of FTM curves and reduce residual color aberrations at the expense of a smaller image circle.
For example if you look at the Lino-Rodenstock line of digital lenses, you have two categories, the apo-sironar-digital and the apo-sironar-digital-HR. The HR versions have a smaller image circle but improved performances within this image circle.
Just a comment about resolution : film users are familiar with resolution tests made on fine-grain film with a test target like the beloved 1951 US Air Force target.
Top-class 'film' lenses for medium 6x6 and and 6x9 formats can exceed 80 cycles/mm in this kind of resolution test, whereas a digital sensor with a pixel pitch of 7 microns has a theoretical resolution limit of 70 cy/mm.
So if resolution in the classical sense of the 1951 USAF test target on fine-grain B&W film was the only issue, there would have been no need for developing new "digital" lenses.
The edges of the circle will have poorer resolution than the center; the farther out you go, the worse it gets.Quote:
Originally Posted by adooropens
Not necessarily.
Image circle is primarily a function of focal length and the design of the lens (angle of coverage). Lens resolution is a function of lens design and engineering/assembly etc but doesnt primarily revolve around focal length.
For example of image circle tessar lenses have an angle of coverage by virtue of their design of round about 60 degrees. A given tessar will cover a radius of f.tan(theta/2). As theta is the same the longer the lens the bigger the circle of coverage.
If you look at some real world examples at http://www.hevanet.com/cperez/testing.html and compare tessars eg nikkor M series 200mm and 300mm and Fuji L series 210mm you see broadly similar resolution given sample to sample variation (these are also lenses of some age so their care and history will have a bearing).
The coverage of these lenses is quite different: 200mm M 210mm versus 300mm M 385mm
Brian, I used to think, like you, that attainable resolution wasn't affected by focal length, given that the two lenses in question have the same prescription. But Brian Caldwell, a lens designer, has pointed out that within a prescription everything including the magnitude of the aberrations scales with focal length.
Comparisons within a design family, for example, tessar types, are dangerous because there's no reason to believe that one maker's design has the same prescription as another's. If you want to see what lens prescriptions look like, look at some lens patents or visit Eric Beltrando's site www.dioptrique.info. While there, look at, say, Eric's calculated curves for a variety of different f/4.5 Tessars.
Cheers,
Dan
Dan thanks for the very informative link.
If I understand you correctly, then within a lens design, in general as focal length increases circle increases and resolution decreases? Again if I understand correctly this would imply that as focal length increases one would need a 'better' design to achieve the same resolution as achieved at the shorter f with the 'simpler' design?
Within a lens design, the angular resolution is constant. Which means, of course, that linear resolution at the film plane decreases linearly with the distance between the lens and the film.
Brian, better doesn't aways mean more complex. It can simply mean different.
Most of the prescriptions that Eric shows are scaled for a 100 mm lens. To use one of his prescriptions to make, say, a 200 mm lens, all dimensions (each element's diameter and thickness, each surface's radius of curvature, distance between groups) would be doubled. If you look at, for example, at all of the tessar types for which Eric gives prescriptions you'll see a lot of variation in the glasses used (index of refraction, dispersion) and in element's thicknesses and surfaces' radii.
Not all lenses of the same general type were created equal.
Ole said it better than I did. But we both neglected a point Brian Caldwell made. Diffraction doesn't scale linearly with focal length, so at smaller apertures longer lenses are better that Ole's clear explanation would predict.
Cheers,
Dan
> How do image circle and lens resolution relate to each other? Do lenses with very large image circles necessarily have poor resolution?
Of course, this can be a loaded question, cause you have not set any boundaries for the lens vintage, cost, size, design, etc. etc. But in general, assuming we keep as many variables constant as possible, mainly fl and modern design, your assumption above is correct. To create larger image circles, lens designers are forced to design with higher f ratios. As an extreme example, a 200mm f2.0 lens for 35mm format, will deliver an image circle maybe 50mm in diam., but at near diffraction limited resolution, 1500/2 = 750 lp/mm of aerial resolution.
However, a 200mm LF lens for 8x10 must be designed at much higher f ratios.... as lens design has improved through the years, the optimized f stops have fallen. A newer 200mm fl range lens today can be optimized at approx. f16. So assuming the lens is diffraction limited at this f stop, the best aerial resolution it can deliver 1500/16 = 94 lp/mm, about 1/7 th of the 35mm lens. But as a general rule, LF lenses are never fully diffraction limited, as its impossible to control all the abberations in lens designs under 8 - 10 elements. The 35mm lens I quoted above, for an example, has 18 elements.
I imagine it is theoretical possible, to design a 200mm f2.0 lens to cover 8x10, but it would take some serious technological breakthroughs in optics design, and you would probably end up with a 30 lb lens that has one usable f stop, not very practical.
> Also, in digital application, do the digital LF lenses make much of a difference compared to high quality LF film lenses (mostly around F11 to F22)? Does anyone have experience with this?
the answer is YES. although some of this depends on the type of film you are comparing it against, and also, the density of the sensor you plan to use.....but I won't digress there. LF digital lenses are optimized for smaller image circles, therefore their designs are optimized at lower f ratios, often in the f5.6 - f8 for the shorter fl's and f8 - f11 for the normal to slightly long fl's. This represents a huge benefit when shooting digital vs. using LF film lenses of the same fl, whose designs are optimized around f11-16 and f16 - f22 respectively. This assumes you are trying to squeeze every bit of resolution from your image capture system.... if that is not the case, and you want / need to shoot at f22, as you suggested, then you have pretty much negated the value of the digital lenses. In which case, you can still use LF film lenses with digital backs and get comparable results. The contrast recorded will be much less with the LF film lenses, but since everything is digital, you can pump some of that back in during the tweaking session.