Super slow, super IQ lenses - an idea?

So, just an idea I had.

Most of the time, a lens produces better shots stopped down. The bigger the aperture, the harder to retain nice IQ. That's why big aperture lenses get so much more expensive exponentially.
Now, I picked up an Olympus OM Zuiko 28mm f/3.5 the other day (maybe you saw the topic, anyway I LOVE it [for the price]). It's super sharp an produces amazing colors, the only 'downside' is, is that it's so slow. And that' where my idea popped up.

The rear element is tiny, not a whole lot of glass involved in the first place. Wouldn't it be an idea to make a 28mm f/8 lens - it's very slow, but the tiniest bit of glass involved would be really easy to make so that it works flawlessly, right?
A lens just for outdoors and landscapes. It'd be much easier not to have distortion, soft corners and all that nasty stuff...

Wouldn't it be possible to make such a lens, relatively, really cheaply, but with outstanding optical performance. I'm talking blowing-L-out-of-the-water quality.

Any thoughts why this would or wouldn't work?

Landscaper, cityscape r, and outdoor photographer will love it. I think it will be cheap, sharp, very good IQ, and light. But a fast lens will produce generally same result at those aperture. Waiting other's opinion.

Half of Canon's lenses already outresolve current sensors - it'd be difficult to "blow L out of the water"

The other problems are f/8 lenses would be very dark in viewfinder, and impossible to autofocus for non 1 series cameras

It's likely that the next generation of DSLR sensors will run into diffraction issues at around f/5.6. It would be a great pity if your super-sharp lens was beaten by the laws of physics.

Something else to consider - there comes a point when the cost of glass is insignificant in the cost of the lens manufacture - it might even get more expensive because making and mounting tiny lenses with sufficient accuracy becomes expensive... There are also commercial issues - at some point companies are not going to want to make or stock them because the margin made on each lens wouldn't really warrant the capital investment or keeping it in stock....

Smorter's comments on focus are right on too.... can't focus something if you can't see it... same rule applies to the camera's autofocus sensor.

From a lens designer perspective: lenses are already created optimized for closed down aperture (it's a designer's parameter - how much importance to give to stopped-down performance vs. wide open performance). So I'd say many of them are as good at f/8 as any specially designed slow lens would be. Regarding the cost - it may actually become more expensive to create smaller lenses once you reach certain size. So the current lenses of say f/2.5 (at 50mm) or f/3.5 (at 28mm) are probably as cheap as you can make them, and perform as well as any lens can at f/8. The only drawback is that they are heavier, and probably focus a bit slower than a f/8 lens (on the other hand - an f/8 lens would not even autofocus on DSLRs, which require at least f/5.6 to AF). But as a bonus you get faster apertures (with lower quality).

I think an f2.8 or f4 prime is already a "super slow" prime and probably already has as much advantage as it is going to get.

I might also consider that when your OLympus 28mm came out (~1972), 28mm was really very wide and f3.5 was relatively fast for those wide focal lengths. It was very likely the fastest aperture they could reasonably do at the time.

Sounds like your already acclimated to the vintage MF lenses and diffraction starts creeping in above f8.0 on most all of them that i have tried & tested with the exception of some of the Leica true APO designs that im sure you wont deem economical.

Commence doing some research on some of the MF lenses and by all means dont discount the modern AF prime lenses that produce very nice IQ at f2.8 to F8.0 at relatively cheap prices and there not L lenses.

I dont recall which format your using but below is a relatively inexpensive $250 Canon adapted Russian MIR24 35/f2 MF lens that does rival some of the L glass i have at present and in the past, and i guarantee surpasses what you have now that your happy with. The below was shot using a crop body and the IQ is better with my 5D MKII sensor:

https://photography-on-the.net …?p=12053636&pos​tcount=279


Regards,

I have the Mir24M too and it's quite nice if used correctly, I also have the Mir20M 20/3.5 but it's not on the same level in the extreme corners, may work better on crop or for video.

One thing some people need to get is that, at least in theory, the wider the aperture of the lens, the sharper it is, assuming perfect optical design. The reason is that as diffraction sets in at narrow apertures because the airy disk (smallest resolvable level of detail) enlarges, at wide apertures it becomes smaller, letting smaller/denser sensors resolve detail that would fall between the cracks of larger/lower-density sensors. Some lenses used in chip manufacture have apertures as wide, or wider than f/0.5 (zero point five) to resolve detail on the nanometer level.

On the 5D2 the diffraction-limited aperture is around f/11, which means that regardless of lens, f/11 is when softness will start to hit, but on the 7D this may be as low as f/8, and some people even claim this to be as low as f/5.6

The above points mean that there is a reason for Canon or other camera manufacturers to continue making wide-aperture lenses, and particularly making them as sharp as possible at those wide apertures, as it gives them more headroom for resolution increases. A native f/8 aperture lens would be largely unfeasible for use on anything other than on a format bigger than 35mm and with mirror lenses; in large format photography wide apertures of f/5.6~f/9 are not at all uncommon.

smorter wrote in post #12579116
Half of Canon's lenses already outresolve current sensors - it'd be difficult to "blow L out of the water"

The highest resolution that Canon offers on a DSLR is 21.1 Megapixel, or to be precise, it produces images that are 5616 × 3744. That means that theoretically, ignoring such things as the AA filter, non-square pixels etc a 5DII or 1DsIII can resolve 5616/(36*2) = 78 lp/mm. A 7D is currently the highest resolving camera Canon produces with a theoretical limit of 5184/(22.3*2) = 116 lp/mm.

In practice those number are 20-30% less primarily because of the antialiasing filter but there are other causes for degradation.

Can a Canon prime, say a nifty fifty resolve 78 lp/mm at say f/5.6? Sure, but the question is - at what MTF? What will the contrast between the line pairs be? MTF50 (50% contrast spatial frequency) is a standard measure.

The world record for a production lens is held by a rangefinder lens (A Zeiss 25/2.8 Biogon ZM) with around 400 lp/mm for MTF10 (or 200 lp/mm for MTF50) in the center at f/4 and it's diffraction limited at that point. At MTF10 you can just differentiate between the line pairs (another name is for it is the "vanishing resolution").

Very few Canon (or other SLR lenses) are likely to resolve over 100 lp/mm at MTF50 - and even if they did they would not be as good as they can be. A theoretically ideal lens that completely maxed out the current generation of sensors would have an MTF100 of 116. That would mean 100% contrast at maximum resolvable spatial frequency. No lens in the world can do that.