In fact the question is more generic than that. Smaller format lens designers try to compensate for their imaging system geometric resolution penalty (compared to a larger format when viewing final images at the same size) by designing ‘sharper’ lenses specifically for it, rather than recycling larger formats’ designs (feeling guilty APS-C?) – sometimes with excellent effect. Are they succeeding? I will use mFT only as an example here, but input is welcome for all formats, from phones to large format.
Since the diagonal of micro four thirds sensors is about half that of full frame sensors, we know thanks to geometry that the unprocessed linear spatial resolution (which for convenience we will call ‘sharpness’ in this post) captured by a mFT system needs to be about twice that captured by a FF system in lp/mm in order to show the same ‘sharpness’ on images displayed at the same final size.
In practice that’s fairly hard to accomplish at an imaging system level, in part because of the more stringent requirement on the lenses and in part because larger formats have more space to pack more pixels into the sensing area. The effect is fairly easy to see when data is presented on a per-picture-height basis. Here for instance is an older study (circa 2013) measuring system MTF50 performance obtained from the center of dpreview.com’s new studio scene raw captures in units of lw/ph.
Most captures were taken at f/5.6 and not set up equivalently. If you are interested, you can read camera/lens parameters as shot in dpr’s comparison tool. I have tried to eliminate obvious outliers but keep in mind that in some cases technique or some other variable may have been sub optimal so the captures may have been re-shot at a later time. We are not interested in the performance of a specific camera on a specific occasion, however: we are interested in the general trend, which appears fairly clear. On average the larger the format, the ‘sharper’ a final image will be when displayed at the same size. So this is the story at the imaging system level.
But for this article let’s leave the sensor aside and concentrate on lenses alone. Sensors with CFA filter arrays are fairly easy to evaluate on their own once pixel pitch and anti-aliasing filter strength are known.
So leaving weight, size and sensor resolution aside, have mFT lenses on their own become competitive with FF’s – are they typically twice as ‘sharp’ in lp/mm? There are two opposing forces at work when manufacturing such lenses: it’s harder to make smaller lenses sharper; on the other hand sometimes they can get away with a smaller surface, which may help produce better performance. Therefore the answer may depend on the lens and the application.
The last time I looked into it more than a couple of years ago most sensors had AA filters and the answer for my purposes (landscapes) was ‘mostly not’. Things may have progressed since then so I would like to explore the question again, this time with your help. It’s not a leading question, I am genuinely interested in assessing the situation today.
This type of exercise generally loses focus quickly because there are too many variables involved, from hardware variability to testing system, to tester ability – so we are going to narrow it down even further to just the bare essentials: only data from manufacturer MTF charts. This is less than ideal for a number of reasons, not least of which the fact that MTF curves only tell part of the story and many manufacturer MTF charts are not actually representative of measured data. Some say they cheat. Let’s acknowledge that and call it the manufacturer’s take. And if someone has a better suggestion (short of buying an optical bench) on how to compare lens-only performance objectively and quantitatively from data available online I would be happy to explore it.
In the meantime here are the rules of the spec comparison game for a first set of ‘normal’ lenses, 50mm primes:
- prime lenses of equivalent focal length +/-15% the spec
- data at about f/8 equivalent aperture (anything up to f/8 equivalent accepted)
- lenses will be compared in two categories: <$300 and <$1000, prices at amazon.com (new lenses, with warranty, no grey market)
- only 30 and/or 40 lp/mm equivalent sagittal and meridional curves will be compared directly off manufacturer MTF charts (no interpolation)
- source of data must be linked and the units of the curves easily determined
Here are my two 50mm prime candidates for Full Frame to start things off. First the Canon EF 50mm f/1.8 II for the budget category. The lower set of blue curves oscillating around 0.9 below is what we are interested in, the 30 lp/mm Sagittal (solid line) and Meridional (dashed) measurements at f/8.
The Sony FE 55mm F1.8 ZA is $998 on Amazon.com today.
Your mission, should you be willing to accept it, is to post in the dpr thread related to this post a manufacturer MTF chart conforming to the rules that has a curve directly comparable to the ones above, preferably with better performance in the same category. As an example, you could provide manufacturer MTF charts of 25mm mFT lenses showing a 60 and/or 80 lp/mm curve, costing less than $1000.
Let’s see if we can determine what the state of the art is, at least according to the manufacturers.
The thread has been going on for a few days now, with some good discussion but nary a chart of the desired type. I was hoping that some gurus would have had a number of manufacturer MTF curves stashed away in their drawers for favorite lenses, but apparently not. It turns out that while some manufacturer information is available on FF lens performance in the working f-range above wide open, there doesn’t seem to be any for mFT or other systems. Pity. Should anyone come across such charts, please let me know.