Lens ISO 12233 Crops

[Dr Croubie]

yeah, the ISO just randomly assign numbers to things in the order they get them. ISO 9660 for example governs how files are written to CD-Rs, ISO 12232 governs the

[Raid]

I was just wondering how useful these charts are in the digital era. Are we now at the stage where the pixel density and lens quality are getting to the point where this cart no longer has the resolution required?

[Daniel Browning]

I was just wondering how useful these charts are in the digital era. Are we now at the stage where the pixel density and lens quality are getting to the point where this cart no longer has the resolution required?

Right now, most cameras cannot resolve the full resolution of the chart (4000 L/PH), but the 24 MP D3X comes closest, and it's just an aliased mess. When we get cameras above 24 MP, we'll start to lose the ability to see how the aliasing artifacts look, but there's a long way to go before we'll lose the ability to determine the alias-free resolution. And even when that happens, this chart will still be useful for showing differences in contrast and aberrations, even if it can't show the ultimate limit in resolution.

[neuroanatomist]

Right now, most cameras cannot resolve the full resolution of the chart (4000 L/PH), but the 24 MP D3X comes closest, and it's just an aliased mess. When we get cameras above 24 MP, we'll start to lose the ability to see how the aliasing artifacts look, but there's a long way to go before we'll lose the ability to determine the alias-free resolution. And even when that happens, this chart will still be useful for showing differences in contrast and aberrations, even if it can't show the ultimate limit in resolution.

Note that the "full resolution of the chart (4000 LW/PH)" applies to the visual resolution features, i.e. the hyperbolic curves that are shown in Bryan's ISO 12233 crops. The standard ISO 12233 chart allows visual estimation up to 2000 LW/PH, while the derivative chart that Bryan uses doubles that, but only for the center hyperbolic curves (the mid-frame and corner ones go only to 2000 LW/PH).


However, the ISO 12233 standard includes slant-edged features intended for spatial frequency response (SFR) analysis - that's an image analysis approach which provides greater sensitivity and accuracy for measurement of spatial resolution than the visual 'estimation' method also available with the charts. Imatest is a commonly-used piece of software that performs a variety of relevant analyses, including SFR analysis (I infer that Bryan has Imatest, since his vignetting images look like they come from that software). Although Bryan doesn't report SFR analysis data, his charts have the ISO 12233 standard slanted-edge features, and also have a lower contrast set of slant-edge features for better correlation with grayscale images from dSLR sensors. Some other sites, e.g. photozone.de, derive their MTF data from SFR analyses performed with Imatest.


For those who are interested, I believe the charts Bryan uses are the Applied Image QA-77. Ihave the -3 version, and it's just shy of being large enough for proper framing at 16mm (on FF) at the MFD of the 16-35mm II, but fine for all my other lenses. From the Specs & Measurement tool he provides, it seems Bryan has all the sizes, as smaller ones are better for supertele lenses unless you have a football field-sized testing area).

[Daniel Browning]

Ah, very informative. Thanks John (neuro).