[gazonk]

Maybe that aesthetic will come back

Good grief, the return of David Hamilton ?

But don't forget that the DOF of an 1.2 lens on APS-C isn't really that shallow. It's like 1.8 on FF - i.e. a 33mm f/1.2 would give me the same FOV but slightly less shallow DOF than the good old M-50/1.7 does on my LX.

[Christopher M.W.T]

The problem with faster lenses very shallow DOF (1.2) is it works against AF systems and leads to centre point framing to stay within the very small circle of sharpness. They reduce creative opportunities, especially with off-centre focus selection. Asking for really fast glass goes against expectations of sharpness, especially across the frame. This is a major reason why they are not being produced as they used to be. They reached their zenith in the "soft focus"look days when people wanted softness i their images. Maybe that aesthetic will come back

It also reduces contrast, and with today's digital sensors, the loss of colour vibrancy is visibly apparent where not so much with film. This is the other reason why faster glass has become difficult to justify manufacturing. That and higher ISO capabilities in the cameras themselves have reduced the need.

Interesting, I've often wondered why super fast primes are so hard to come by these days. You think though with advancements in lenses, manufacturing, coatings etc they could improve up on this?

[Aristophanes]

Interesting, I've often wondered why super fast primes are so hard to come by these days. You think though with advancements in lenses, manufacturing, coatings etc they could improve up on this?

With Canon there is the 50/1.2 at $1,400

If you look at Pentax, you see that th A 50/1.2 went out of production in the late 80's? There was a similar production stoppage with other brands. Why? Likely the rise of AF. The too-shallow DOF impeded accuracy for shallow DOF.

Also, it's a tiny niche market for a very expensive piece of glass with sub-optimal differences between slightly faster, less expensive models. Check out the review of the pricey Canon here:

Canon EF 50mm f/1.2 L USM Lens Review

Note the bokeh comparisons at 100% crop. Without the crop at normal viewing distance the aesthetic difference between 1.2 and 1.4 is almost nil, but Canon will gladly take your extra $1,200... a fool and his money...

I also suspect that digital sensors exacerbate the sharpness issue. All photo software can zoom, but in the film days, few of us had loupes. f/1.2 = image softness so apparent in digital. It's optically a given that too wide an aperture creates image softness. That's why most "pro" glass from any DSLR system starts at f/2.8. What do we hear? I want a smallish prime, f/1.2 fast, 20mm, and sharp across the frame.

Choose 1 of those 4. Can't get all.

With regards to the point that an APS-C sensor has an inherently deeper DOF, true. but the same # of photons is getting in regardless of sensor size or 135 film. So it it's low-light sensitivity you want, it's there. If you want an even shallower DOF, argue for an FF sensor. or retrofit the Leica 0.95. Obviously rather than put resources into glass, camera-makers are seeing more value in ISO advances. If you want prettier bokeh, Photoshop is your friend.

As for the David Hamilton aesthetic, well, interesting subject matter, not all bad. Not at all.

[ghelary]

Wrong. 1.2 is close to half a stop from 1.4, but 1.8 is only a little more than a quarter stop fro 2.0.

The half-stop between 2.0 and 1.4 is 1.7.

Before trying to say right or wrong, please get you math right, aperture is defined by the ratio of the diameter of the front element and the focal lens, a stop is defined by the variation needed to gather twice the light (or half) when speaking about aperture stop are defined by the square of the diameter, hence 1 - 1.4 - 2 - 2.8 - 4 which are approximates of the real digits.

root square
1 1
1.2 1.44
1.4 1.96
1.414213562 2
1.7 2.89
1.8 3.24
2 4
2.8 7.84
2.828427125 8

So neither 1.7 or 1.8 are half a stop between "1.4" and "2", if you add that real aperture varies with focus distance and that lens makers take liberties with the labelling, pinning out a 1/4 of stop between 1.7 and 1.8 is plain silly.

There's an ocean between pure science and engineering and marketing is in another planet.

[Aristophanes]

Interesting, I've often wondered why super fast primes are so hard to come by these days. You think though with advancements in lenses, manufacturing, coatings etc they could improve up on this?

Before trying to say right or wrong, please get you math right, aperture is defined by the ratio of the diameter of the front element and the focal lens, a stop is defined by the variation needed to gather twice the light (or half) when speaking about aperture stop are defined by the square of the diameter, hence 1 - 1.4 - 2 - 2.8 - 4 which are approximates of the real digits.

root square
1 1
1.2 1.44
1.4 1.96
1.414213562 2
1.7 2.89
1.8 3.24
2 4
2.8 7.84
2.828427125 8

So neither 1.7 or 1.8 are half a stop between "1.4" and "2", if you add that real aperture varies with focus distance and that lens makers take liberties with the labelling, pinning out a 1/4 of stop between 1.7 and 1.8 is plain silly.

There's an ocean between pure science and engineering and marketing is in another planet.

The charts are right here:

F-number - Wikipedia, the free encyclopedia

[Pål Jensen]

IMO the 28mm would have to be F/1.8 for me to consider it. F/2 just isn't fast enough.

Weird stuff. A 28 F:2 was regarded as fast in the time that high-speed Ektachrome was 160ISO. Before digital serious photography was rarely shot at more than 100ISO (in color at least). Now we can shoot at 3200ISO with great results. Ultra fast lenses aren't really needed anymore....

[yeatzee]

. Ultra fast lenses aren't really needed anymore....

Thats quite a bold statement

Now we can shoot at 3200ISO with great results

"Great results" is highly subjective.

-------------

Oh and I edited my original post to say F/1.7 instead of F/1.8. I meant to write that but read someone say F/1.8 and for whatever reason wrote the same.

[Aristophanes]

Thats quite a bold statement


"Great results" is highly subjective.

It's bold, but in market terms, largely accurate. That's why the premium for very fast glass is in the thousands of $$$ now, all the way up to $10,000 for the Leica 0.95 (which is no sharper at 2.8 than any other fast lens).

Subjectively AND objectively you would be hard pressed to tell the difference between 1.2's and 1.4's aesthetically. The market couldn't, so all manufacturers priced the product into the stratosphere, or they dropped production.

For Pros already doing PP, Photoshop is far cheaper for "great results" (meaning paying results). Plus, they can focus off-centre stopped down increasing sharpness, impossible on very fast glass due to the unavoidable nature of the optical design and the shallow DOF it creates. Start sharp and soften if needed in PS, but never start soft.

[Digitalis]

For Pros already doing PP, Photoshop is far cheaper for "great results" (meaning paying results). Plus, they can focus off-centre stopped down increasing sharpness, impossible on very fast glass due to the unavoidable nature of the optical design and the shallow DOF it creates. Start sharp and soften if needed in PS, but never start soft.

Have you ever seen what happens when you stop the pentax 50mm f/1.2 down to f/2? it is very sharp, noticeably sharper than a f/1.4 lens would be at f/2. And that is one of the many reasons why faster lenses like the leica noctilux 50mm f/0.95 exist so you can afford to stop the lens down to increase performance and DOF. but still retain high shutter speeds in dim lighting.

[Aristophanes]

Have you ever seen what happens when you stop the pentax 50mm f/1.2 down to f/2? it is very sharp, noticeably sharper than a f/1.4 lens would be at f/2. And that is one of the many reasons why faster lenses like the leica noctilux 50mm f/0.95 exist so you can afford to stop the lens down to increase performance and DOF. but still retain high shutter speeds in dim lighting.

Yes, I have stopped a 1.2 down and a 1.4 and 1.7 to f/2 and frankly, no, I see little difference in sharpness. Hard to compare with Pentax as most lenses pre-date MTF, but Canon has par examples discussed extensively on their forums:

Canon EF 50mm f/1.4 USM - Review / Test Report - Analysis

Canon EF 50mm f/1.2 USM L - Review / Test Report - Analysis

As one can see, the 1.4 is measured at 1.8, not 2.0, but if you follow the curve, by f/2.0 they are essentially equal. In fact, at 1.7 to 1.8 they are essentially equal. At 2.8 the 1.2 is actually inferior on the borders by a very small amount. Both are soft at 1.4.

The 1.2 has 5x the CA. Equal on the vignetting. The verdict says it better than I could.

The Pentax 50/1.4 seems to be a little behind the Canon, but then again, it's a much older lens. Maybe the 25 year-old A 1.2 is that much better, but we'd need an MTF comparison which is not available. I cannot see how it is much superior to the Canon 1.2; in fact, one could reasonably assume it is not, and I doubt Pentax could make a better 1.2 than Canon. The DA*55 fares a little better than the venerable FA 50/1.4. I do not see how the A 1.2 is "noticeably sharper" than a Pentax 50 1.4 by f/2.0.

Back to the Canon where there is objective data. Is the extra stop worth $1,150? Some of that price difference is other quality aspects to the 1.2 lens, but essentially that's gravy. It's a cool grand to get a stop softer meat.

Also, I would take issue with the "increase performance" aspect. You lose contrast wide open. That's lost colour and shadows data you never get back. Fast glass is a tradeoff. I agree you get headroom by starting lower, but the tradeoffs are still there until about f/2.0 to 2.8, even with the Leica. Look at the MTF's in this review:

diglloyd: Leica 50mm f/0.95 Noctilux-M ASPH

I like the line "hardly anything will be in focus". Again, a tradeoff. Same for the 6-stop ND filters! Read the other compromises inherent to the optics. For every increase in the aperture size, something has to give. The "too bright" comment is telling. You'd have to stop down if there's a single candle in the frame in a dark jazz club, by which point you're probably in the price range and quality of a lens worth $7,000 less, like the 1.4 Summilux where they are identical in performance. (In fact, it is highly recommended to have the Noctilux fixed with a factory 3-stop ND filter!). You don't just get a faster glass and everything else stays constant. Your wallet is certainly lighter. Don't get me wrong, fast glass is fun. But at a certain point it's like driving a car capable of going 600 MPH. Are you really going to push it that far often? No, and it makes a lousy vehicle for just going to the store for a jug of milk. The Noctilux largely about bragging rights. It's not about making someone a better photographer in a dimly lit, smoky blues bar. Sad for the guy whose Noctilux was mistaken for a beer bottle to be thrown at the chicken wire.

[mattdm]

Before trying to say right or wrong, please get you math right, aperture is defined by the ratio of the diameter of the front element and the focal lens, a stop is defined by the variation needed to gather twice the light (or half) when speaking about aperture stop are defined by the square of the diameter, hence 1 - 1.4 - 2 - 2.8 - 4 which are approximates of the real digits.

root square
1 1
1.2 1.44
1.4 1.96
1.414213562 2
1.7 2.89
1.8 3.24
2 4
2.8 7.84
2.828427125 8

So neither 1.7 or 1.8 are half a stop between "1.4" and "2", if you add that real aperture varies with focus distance and that lens makers take liberties with the labelling, pinning out a 1/4 of stop between 1.7 and 1.8 is plain silly.

f/1.7 is actually very, very close to have a stop between f/1.4 and f/2. Specifically, sqrt(2)^1 ≈ 1.4142, sqrt(2)^2 = 4, and sqrt(2)^2.5 ≈ 1.6817.

Likewise, 1.2 is very close to half a stop between f/1 (sqrt(2)^0 = 1) and f/1.4: sqrt(2)^0.5 ≈ 1.1892.

(And f/1.8 is a third-stop less than f/2 — sqrt(2)^(1+2/3) ≈ 1.7818.)

[gazonk]

Ultra fast lenses aren't really needed anymore....

Except that no cameras have really good results at ISO 3200 yet and that shallow DOF has its own virtues. Here's a shot @f/1.7 with the M-50/1.7:

[Digitalis]

Takes A deep breath

Yes, I have stopped a 1.2 down and a 1.4 and 1.7 to f/2 and frankly, no, I see little difference in sharpness. Hard to compare with Pentax as most lenses pre-date MTF, but Canon has par examples discussed extensively on their forums:

the difference is in the accutance between tonal values. The spherical aberration from the 50mm f/1.2 will be much lower than the f/1.4 lens at f/2, and therefore micro tonal values will be rendered with more clarity.

As one can see, the 1.4 is measured at 1.8, not 2.0, but if you follow the curve, by f/2.0 they are essentially equal. In fact, at 1.7 to 1.8 they are essentially equal. At 2.8 the 1.2 is actually inferior on the borders by a very small amount. Both are soft at 1.4.

from my testing the pentax K 50mm f/1.2 has a different resolution characteristic from the canon 50mm f/1.2 which I own and use from time to time. In my testing with a pentax LX on T-Max 100 at f/2.8 the SMCP 50mm f/1.2 has higher overall resolution in the periphery of the imaging field than the Pentax SMCP-FA 50mm f/1.4. The Canon 50mm f/1.2L is close to the SMCP 50mm f/1.2 @f/2.8 however the image quality at the edges of the frame are a bit softer in comparison to the SMCP 50mm f/1.2

The 1.2 has 5x the CA. Equal on the vignetting. The verdict says it better than I could.

The SMCP 50mm f/1.2 has extremely low Purple fringing*,however it does possess transverse chromatic aberration which is exacerbated at closer focusing distances, it is most apparent at the periphery of the imaging field. It is my personal opinion in regards to optical aberrations the Canon 50mm f/1.2 is an inferior performer when compared to the SMCP 50mm f/1.2

The Pentax 50/1.4 seems to be a little behind the Canon, but then again, it's a much older lens. Maybe the 25 year-old A 1.2 is that much better, but we'd need an MTF comparison which is not available. I cannot see how it is much superior to the Canon 1.2; in fact, one could reasonably assume it is not, and I doubt Pentax could make a better 1.2 than Canon. The DA*55 fares a little better than the venerable FA 50/1.4. I do not see how the A 1.2 is "noticeably sharper" than a Pentax 50 1.4 by f/2.0.

The DA*55mm has higher amounts of PF and CA than the SMCP 50mm f/1.2. The FA 50mm f/1.4 is an old design, but it is basically identical to most of the 50mm f/1.4 "me too" lenses. The DA*55mm f/1.4 uses a different optical formula in comparison to the SMCP 50mm f/1.2 and FA50mm f/1.4 lenses. And since the lenses are of a different focal length it becomes increasingly difficult to quantify and compare results.

Back to the Canon where there is objective data. Is the extra stop worth $1,150? Some of that price difference is other quality aspects to the 1.2 lens, but essentially that's gravy. It's a cool grand to get a stop softer meat.

In the case of the canon lens no, it really isn't worth it. because the Canon EF 50mm f/1.2L has a T stop of T=f/1.33- the SMCP 50mm f/1.2 has a T stop T=f/1.22(as I recall) so IMO it isn't even worth comparing the two lenses.

Also, I would take issue with the "increase performance" aspect. You lose contrast wide open. That's lost colour and shadows data you never get back. Fast glass is a tradeoff. I agree you get headroom by starting lower, but the tradeoffs are still there until about f/2.0 to 2.8, even with the Leica. Look at the MTF's in this review

the colour and contrast doesn't just vanish, it's still there. what do you think PP,darkroom filters and variable contrast papers are for? Do you expect every image from your lenses to be perfect in colour and contrast? even without application of rudimentary PP? I have never seen a jeweller present a finished ring with an uncut diamond in its setting.

I like the line "hardly anything will be in focus". Again, a tradeoff. Same for the 6-stop ND filters! Read the other compromises inherent to the optics. For every increase in the aperture size, something has to give. The "too bright" comment is telling. You'd have to stop down if there's a single candle in the frame in a dark jazz club, by which point you're probably in the price range and quality of a lens worth $7,000 less, like the 1.4 Summilux where they are identical in performance. (In fact, it is highly recommended to have the Noctilux fixed with a factory 3-stop ND filter!). You don't just get a faster glass and everything else stays constant. Your wallet is certainly lighter. Don't get me wrong, fast glass is fun. But at a certain point it's like driving a car capable of going 600 MPH. Are you really going to push it that far often? No, and it makes a lousy vehicle for just going to the store for a jug of milk. The Noctilux largely about bragging rights. It's not about making someone a better photographer in a dimly lit, smoky blues bar. Sad for the guy whose Noctilux was mistaken for a beer bottle to be thrown at the chicken wire.

Well that would be a $15,000 mistake. but from my ownership of leica lenses and cameras i'm sure the 50mm f/0.95 ASPH can take more abuse than most lenses will ever be subject to. I use my noctilux 50mm f/1.0 without the ND filter. As Ansel Adams said "If your lens has bad vignetting, use it to frame your subject" those ND filters are for nitwits who obsess over achieving this ridiculous ideal of perfection. The tools we are using are imprecise, compare the shutter mechanism of a Nikon F5 to a swiss designed invar watch movement and the shutter from the F5 will look like a crude contrivance.

*I'm not entirely sure what causes PF, there are hypotheses some claiming it is UV light,some saying it is a sensor defect. I have another theory, but I need more hard data to come to any conclusions. All I know for certain lenses suffer from inordinate amounts of it; FA77LTD and DA 12-24mm f/4 are the worst culprits. Yet, some lenses appear to be remarkably free from it.

[falconeye]

Look at the MTF's in this review:
diglloyd: Leica 50mm f/0.95 Noctilux-M ASPH

Thanks for the link.
Leica normally has 5/10/20/40 lp/mm curves. 40% MTF in the center at 40 lp/mm and f/0.95 is not that bad, corresponding to resolving a 6 MP full frame sensor. But you're right, at 10k$, you may expect better.

[Digitalis]

Thanks for the link.
Leica normally has 5/10/20/40 lp/mm curves. 40% MTF in the center at 40 lp/mm and f/0.95 is not that bad, corresponding to resolving a 6 MP full frame sensor. But you're right, at 10k$, you may expect better.

hence the reason why I stayed with the "classic" 50mm f/1.0 Noctilux - and now, thanks to my new M9 I got on my birthday I can now make the most of it