A good read in the mirrorless vs DSLR debates

tko wrote:

It's a myth according to one article that has never been independently verified. As usual, DXO never tell you how they obtained the results. Funny how people disbelieve their sensor numbers, but believe this. I guess it's whether they are on your side on not.

Actually, even that article admits, by it's own number, that 1.2 glass works just fine.

Going from 1.4 to 1.2 should give you 36% more light. According to the DXO figures, you only get 24% more light. OK, not as good as expected, but still a clear benefit.

You want me to do the math for going from F2.8 to F1.2 and see who wins?
sportyaccordy wrote:

The glass is a legitimate reason to stick with DSLRs for sure, but 1.2 glass on a DSLR is a myth...

http://www.luminous-landscape.com/essays/an_open_letter_to_the_major_camera_manufacturers.shtml
Where are you getting this data? Once you get below 2.0, the real transmission threshold is about T1.8. F/1.4 gets a ~0.3-0.4 boost; F/1.2 gets a ~0.5-0.6 boost. So speed wise you are not gaining an advantage on a digital sensor, and there are sound scientific explanations as to why that is the case.

I'm not saying the lenses are pointless. There is way more to a lens than how much light it transmits wide open. The EF 85 1.2 has phenomenal IQ at any setting under any condition. But again, the idea that you are getting T1.2 from it wide open onto the sensor is just not correct, and should definitely be factored into the thought process behind purchasing one.
 
57LowRider wrote:
TrapperJohn wrote:

Gawd, that 85 1.2 is a sweet lens. It shows the same razor sharp, leap out at you definition that one finds in the best Leica M or Oly SHG ZD lenses.

But... you're really looking at two types of photographs, out of dozens of styles. Portrait, and picking one person or one thing out of an otherwise busy scene.

Steve's photos make me want to run out and grab a 5DII or 5DIII and an 85 1.2, while also keeping my EM5 for the scenarios where it really shines: discrete, small, light, but still very good IQ.

Arrrgh! So many options, so little money...
Know what you mean. As a committed Fuji X user, some of those 85 shots made me drool more than a bit - the butterflies are perfect. Unfortunately, unless I hire a bearer to lug the stuff around, I can't go there.
Parting with my Nikon 85f1.4 was hard when I switched to the Fuji X-Pro1. Hopefully the upcoming 56f1.2 will fit the bill.
 
JamieTux wrote:
Richard wrote:

I don't think you need to unlearn. I agree with your observations. I believe this shows this as well. Notice the lines behind the mannequins head as the get narrower.

http://www.wlcastleman.com/equip/reviews/85mm/mann_port/mann_port.htm

The full article

http://www.wlcastleman.com/equip/reviews/85mm/index.htm
Yeah - that's what I was expecting to see but I'm open minded and if Ross proves his point with images I'll pay attention - until then I'll keep thinking the same :)
I guess my take is a little different -- personally, I cannot see a difference in DOF between the f/1.2 and f/1.4 shots in the links above. (Though it is only 1/3 stop.)

My own f/1.2 test shots look similar. DOF difference f/1.2 to f/1.6 is imperceptible. (I would post, but the subject is a friend's portrait.) I guess people wouldn't believe me anyway based on a few jpegs -- folks here seem riled up by this (maybe because their large aperture lenses were pretty expensive?). Instead I would like to show you to an article about it, with a very extensive set of measurements made by DxO on the effect:


The pixel vignetting for large apertures differs by camera and generally by pixel pitch (more pixels on the sensor tends to make the effect worse). Modern cameras are closer to the smaller pixel pitch comparing to the cameras tested (Canon 7D has 18MP APS-C sensor). They did not measure DOF but physically speaking, losing larger aperture light will surely affect DOF.

The falloff in sensor efficiency vs. angle is not going to be a sharp cliff, but rather a gradual reduction vs. angle -- so possibly you'll find f/0.95 really has smaller DOF than f/1.4, but with existing DSLR/ILC CMOS sensor designs, you are not getting the full amount of light and DOF equivalent to the lens aperture. In the future if the sensor designs change (if they start using back-illuminated CMOS for larger sensor cameras, for example), the story will be different.
 
True, you don't need to buy bags, but emptying them can get SO messy! You hit my hot button.
 
Recently, I moved to DSLr, from compact using it for a decade. From the moment I recieved DSLR, it feels luxury, not just build quality is excellent but DSLR feels like more traditional camera style with mirror in it. Not just sensor size big / nor same as mirrorless, but if you want the real photography style I would suggest DSLR is still winner. Somehow, mirrorless cameras feels like more smartphone style/nor toy gadgets, although it's being small and compact size with light, still you cannot put on your pocket that goes same as DSLR. So my suggestion is that if you want more higher pic quality, go for DSLR, period and obvious answer.
 
ragmanjin wrote:
...speaking of expensive commercials, who was it that hired Ashton Kutcher to promote their inferior sub-par cameras all over TV again?
Probably doesn't matter.

--
Probably the the same guy that hired him to replace Charlie Sheen in 'Two and a Half Men' . .

. . . now that did matter!
 
Here's a sample:

"Being full frame lenses with HUGE front elements and a light sucking 1.2 aperture, these lenses are HUGE, FAT, HEAVY and loaded with abilities that can make almost anyone with an ounce of skill into an abstract artist. NOTHING renders like an 85L 1.2 lens and it is more like a big fat paintbrush than a camera lens. Many have nicknamed it “THE KEG” because it looks like a mini keg of beer. It is large but it sure can pull off a special and one of a kind look."

This guy's writing style is just execrable in every way. The typography resembles something produced by a secretary newly designated company graphic designer - using Pagemaker in about 1989.

Roy
 
rossdoyle wrote:
JamieTux wrote:
Richard wrote:

I don't think you need to unlearn. I agree with your observations. I believe this shows this as well. Notice the lines behind the mannequins head as the get narrower.
Yeah - that's what I was expecting to see but I'm open minded and if Ross proves his point with images I'll pay attention - until then I'll keep thinking the same :)
I guess my take is a little different -- personally, I cannot see a difference in DOF between the f/1.2 and f/1.4 shots in the links above. (Though it is only 1/3 stop.)

My own f/1.2 test shots look similar. DOF difference f/1.2 to f/1.6 is imperceptible. (I would post, but the subject is a friend's portrait.) I guess people wouldn't believe me anyway based on a few jpegs -- folks here seem riled up by this (maybe because their large aperture lenses were pretty expensive?). Instead I would like to show you to an article about it, with a very extensive set of measurements made by DxO on the effect:

http://www.dxomark.com/index.php/en/Reviews/F-stop-blues
The DXO article talks about amount of light. I never debated this though I did give an example. Anyway I was talking about DOF and people have been saying you cannot see a difference between F1.2 and 1.6 and I have to disagree.

Here is my example. I think even a 1/3 stop is noticeable. In fact I would say it is quite linear the size of the bokeh balls getting larger and softer progressively.

I think it is even more noticeable when you have a subject further away and a closer background. Either way, the lens at F1.2 makes creamy dreamy backgrounds and bokeh balls progressively. You may disagree but my example proves to me in my mind what I have seen and the difference is not imperceptible but quite noticeable, But again, your perception of what is noticeable and imperceptible may be different than mine. I will let others make their own decision. The below animated gif does not work on my browser. Click the link below to see the difference. The pics that made up this animation are below with exif data included.

lensDOF.gif


animated gif if it does not work above http://www.mxphotos.net/dof/lensDOF.gif

IMG_9500.jpg


IMG_9501.jpg


IMG_9503.jpg


IMG_9504.jpg
 
Last edited:
rossdoyle wrote:
JamieTux wrote:
Richard wrote:

I don't think you need to unlearn. I agree with your observations. I believe this shows this as well. Notice the lines behind the mannequins head as the get narrower.

http://www.wlcastleman.com/equip/reviews/85mm/mann_port/mann_port.htm

The full article

http://www.wlcastleman.com/equip/reviews/85mm/index.htm
Yeah - that's what I was expecting to see but I'm open minded and if Ross proves his point with images I'll pay attention - until then I'll keep thinking the same :)
I guess my take is a little different -- personally, I cannot see a difference in DOF between the f/1.2 and f/1.4 shots in the links above. (Though it is only 1/3 stop.)

My own f/1.2 test shots look similar. DOF difference f/1.2 to f/1.6 is imperceptible. (I would post, but the subject is a friend's portrait.) I guess people wouldn't believe me anyway based on a few jpegs -- folks here seem riled up by this (maybe because their large aperture lenses were pretty expensive?). Instead I would like to show you to an article about it, with a very extensive set of measurements made by DxO on the effect:

http://www.dxomark.com/index.php/en/Reviews/F-stop-blues

The pixel vignetting for large apertures differs by camera and generally by pixel pitch (more pixels on the sensor tends to make the effect worse). Modern cameras are closer to the smaller pixel pitch comparing to the cameras tested (Canon 7D has 18MP APS-C sensor). They did not measure DOF but physically speaking, losing larger aperture light will surely affect DOF.

The falloff in sensor efficiency vs. angle is not going to be a sharp cliff, but rather a gradual reduction vs. angle -- so possibly you'll find f/0.95 really has smaller DOF than f/1.4, but with existing DSLR/ILC CMOS sensor designs, you are not getting the full amount of light and DOF equivalent to the lens aperture. In the future if the sensor designs change (if they start using back-illuminated CMOS for larger sensor cameras, for example), the story will be different.
Hi Ross,

To me less of the hat is in focus at f1.2 but it could be because focus is shifted slightly?

That article and measurement you linked to is all based on the amount of light hitting the sensor, not the DOF, unless I'm missing something they don't even mention the DOF, only the brightness as recorded by the sensor and I've not argued with that at all, we were only talking about the DOF.

I would also never expect any lens to have the same T and f-stop values as that would make it a perfect lens and that's just not practical :)

So out of curiosity do you not use your lens beyond f1.6 any more? I'm guessing that it has a nicer way of drawing than the cheaper alternatives.

Thanks Ross
 
Faster lenses still have other optical advantages. They are more likely sharper stopped down and have less vignetting & other IQ issues. So if you want the best IQ possible def go for the fastest lens you can get. But know you are not getting its true t-stop wide open.
 
sportyaccordy wrote:

Faster lenses still have other optical advantages. They are more likely sharper stopped down and have less vignetting & other IQ issues. So if you want the best IQ possible def go for the fastest lens you can get. But know you are not getting its true t-stop wide open.
That's just nonsense! If a designer is not worrying about as big a lens aperture as possible they can create much better and smaller lenses, ask Leica what their best 50mm is in terms of all aspects of IQ other than maximum aperture and I would be amazed if anyone there said the Noctilux...

If what you're trying to say is that if you spend more you get something better I'd agree most of the time.


Going back to the Canon 85L - I woudl never have owned one as the AF is too slow whereas the 85 f1.8 was pretty much perfect for me optically and had great focus speed too
 
JamieTux wrote:
sportyaccordy wrote:

Faster lenses still have other optical advantages. They are more likely sharper stopped down and have less vignetting & other IQ issues. So if you want the best IQ possible def go for the fastest lens you can get. But know you are not getting its true t-stop wide open.
That's just nonsense! If a designer is not worrying about as big a lens aperture as possible they can create much better and smaller lenses, ask Leica what their best 50mm is in terms of all aspects of IQ other than maximum aperture and I would be amazed if anyone there said the Noctilux...

If what you're trying to say is that if you spend more you get something better I'd agree most of the time.

Going back to the Canon 85L - I woudl never have owned one as the AF is too slow whereas the 85 f1.8 was pretty much perfect for me optically and had great focus speed too
You are correct. The fastest glass is often designed for different aspects than pure sharpness stopped down. that appears to be a misconception often held by the inexperienced.
 
JamieTux wrote:
sportyaccordy wrote:

Faster lenses still have other optical advantages. They are more likely sharper stopped down and have less vignetting & other IQ issues. So if you want the best IQ possible def go for the fastest lens you can get. But know you are not getting its true t-stop wide open.
That's just nonsense! If a designer is not worrying about as big a lens aperture as possible they can create much better and smaller lenses, ask Leica what their best 50mm is in terms of all aspects of IQ other than maximum aperture and I would be amazed if anyone there said the Noctilux...

If what you're trying to say is that if you spend more you get something better I'd agree most of the time.

Going back to the Canon 85L - I woudl never have owned one as the AF is too slow whereas the 85 f1.8 was pretty much perfect for me optically and had great focus speed too
 
JamieTux wrote:
sportyaccordy wrote:

Faster lenses still have other optical advantages.
Going back to the Canon 85L - I woudl never have owned one as the AF is too slow whereas the 85 f1.8 was pretty much perfect for me optically and had great focus speed too
The focus speed was increased between the 85L and the 85L II which is what I have. The focus works pretty well although you are not going to be able to use it for sports (slower moving sports, it would probably work). It was not meant for that, it is more specialized than a general lens, most fast glass in the F0.95 to F1.4 usually is.

I have never had regrets, and there are other lenses I own that cover the fast AF piece. I looked at your sites, most of the pictures there would be covered by the 85L but I admit, if I were going to shoot fast moving action, the 85L would not be the lens.
 
Richard wrote:
JamieTux wrote:
sportyaccordy wrote:

Faster lenses still have other optical advantages.
Going back to the Canon 85L - I woudl never have owned one as the AF is too slow whereas the 85 f1.8 was pretty much perfect for me optically and had great focus speed too
The focus speed was increased between the 85L and the 85L II which is what I have. The focus works pretty well although you are not going to be able to use it for sports (slower moving sports, it would probably work). It was not meant for that, it is more specialized than a general lens, most fast glass in the F0.95 to F1.4 usually is.
Thanks Richard, I tried the version 2, but I had had the Sigma f1.4 and that thing is so fast that the 85L II just felt (this is the keyword, felt :) ) too slow to me. I loved the Noctilux when I played with it - I could focus very quickly with that on an M9.
I have never had regrets, and there are other lenses I own that cover the fast AF piece. I looked at your sites, most of the pictures there would be covered by the 85L but I admit, if I were going to shoot fast moving action, the 85L would not be the lens.
Yeah, the really static stuff ends up being shot with anything (most lenses are pretty good at f8 or f11) and that's what pays most of the bills at the moment. I do have a few f1.4 lenses but I don't shoot Canon anymore. I'm sure that I could have made use of the 85L but I really only wanted low light for shots of my kids at the time (so AF was more important to me).
 
Hi Ross,
To me less of the hat is in focus at f1.2 but it could be because focus is shifted slightly?
That article and measurement you linked to is all based on the amount of light hitting the sensor, not the DOF, unless I'm missing something they don't even mention the DOF, only the brightness as recorded by the sensor and I've not argued with that at all, we were only talking about the DOF.
I would also never expect any lens to have the same T and f-stop values as that would make it a perfect lens and that's just not practical :)
So out of curiosity do you not use your lens beyond f1.6 any more? I'm guessing that it has a nicer way of drawing than the cheaper alternatives.
Thanks Ross
Jamie,

Yes, the article didn't mention DOF at all. However, if the light past a certain aperture is not recorded on the sensor, it doesn't matter whether it is due the lens iris or a mini iris on each pixel (which is approximately what is happening at each pixel on modern camera sensors -- all FF, u4/3, and APS-C sensors are front-illuminated CMOS with microlens arrays). Take a look at my images on the other thread started by Richard. There is a noticeable DOF change f/2.8-f/2.2, whereas there is subtle or no DOF difference from f/1.6-f/1.2, both of which are 2/3 stop difference. People saying only brightness is affected but DOF is not -- are incorrect.

I usually do travel photography and landscapes, so shallow DOF hasn't really been my thing. I only recently acquired this nice Pentax-A 50mm f/1.2 manual focus lens -- it is a beautiful beast, but I haven't really tried it enough to get used to its performance at various apertures. I will definitely keep this lens, though -- perhaps shallow DOF may be the next frontier of sensor improvement?
 
Thanks Ross,

Its weird that the 2 points that proved it to me in my head were the Canon f1.2 and the Leica Noctilux as Leica are known to have an angled microlens array and the Canon seems to work differently to the Pentax and Nikon. Maybe the sensor has more of an impact than we suspected.

As for advances, I'm pretty sure that there will be advances in this area as Leica and Kodak proved that there needs to be to avoid colour shifts in some scenarios.

I've heard great things about Pentax primes (weirdly I've never even handled a Pentax camera despite having used so many cameras from pocket cams all the way up to 5x4 kit) so I would hold onto it too.

Thanks again for explaining and for sharing your examples.
--
James
 

Keyboard shortcuts

Back
Top