D.P. Explain please :)

[D.P.]

I'm not sure what you are trying to show with those images, especially since the exif data isn't viewable (I think imgur stripped it). The phenomena is well known and widely reported by multiple independent experts.

Looks like you kept the shutter speed constant and reduced aperture and thus reduced exposure as expected, that isn't being debated. What happens with ultra fast lenses on digital sensors is that your progressively loose more and more light relative to what is expected from the physical aperture. Therefore, f/1.2 will capture more light than f/1.4, and f/1.4 more than 1.8, but much less than expected from theory.

To compensate cameras secretly increase the ISO without reporting this, even in RAW files, so you wont immediately see this decrease. The real penalty here is that at HIGH ISO you might be pushing the sensor gain an extra 0.5 stops which would add highly visible noise, but at base ISO you may not care too much which is kind of irrelevant as for the purpose of fast glass to capture more light if you are shooting at ISO 100 then you might as well just bump to ISO 200/400 etc.

By just eyeballing your photos it is very hard to spot the difference between in exposure between your f/1.2 or f/1.4 photos, supporting the notion that at f/1.2 you are not capturing significantly more light than at f/1.4 and even if there is a significant difference you have not accounted for the camera increasing the ISO secretly to compensate. Your photos are too small to see if the noise has been increased by the camera raising the sensor gain.


You may think that raising the ISO secretly if not visible at base ISO is entirely harmless, I tend to agree, but if you consider the fact that in this line f though all camera could just secretly use ISO 400 instead of 100 as base ISO giving an instant 2 stop gain to all lenses used and making slow f/4.0 zooms behave like a fast f/2.0 zoom. Clearly we can all raise the ISO to 400 with little concern these days but there is a change in the sensor performance in doing so, most observable in the shadows. You wouldn't want to buy a lens advertised as 105mm f/2.0 when it was actually only f/4.0 and tricked the camera into increasing the ISO 2 stops.


What does remain faithful to the designated aperture is the depth of focus at f/1.2 vs f/1.4 even if you don't see all the theoretical gains in light gathering.

 

[Raymond Lin]

I am trying to show...no, I shown that 1.2 captured more light than 2.2.

You claimed that

So much so that an f1.2 lens doesn't really provide any more light than a f1.4 lens and an f1.8 lens is not really noticible lower.

That is not true. If it is not noticeable then they would look the same.

They are not.

If that statement is true then they would all have the same exposure.

They do not.

I don't know about you, I can tell the difference between 1.2 to 1.4, on my mac and the back of the LCD on the camera. I made a gif so you can see all the steps. 1.2 to 1.4 is not significant, it isn't as it is not a stop of light, we knew that, but it is noticeable nevertheless. It's why I still have the 50/1.4. The 1.2 is a lot of money for that tiny stop of aperture. Where as the Canon 85mm it is between the 1.8 and 1.2, which is a bigger difference.

As the for science and theory...like I said, I wanted to see for myself.

1.2 gives more light than 1.4.

That is a FACT is it not?

As for the OP, is the 50/1.2 worth it over the 50/1.4 for the aperture?

The honest answer is no. It is a lot of money for that tiny bit of difference. You would however be paying for a newer lens design, better bokeh, better build quality. A lot of which is where the money will be going, its why we buy the L, else we'll all be shooting the 35/2.0 or the 35/1.8 and the 35/1.4 would never been made and sold like hot cakes.

 

[King Damager]

I've read and now understand this, but ultimately, D.P.

If you went out tomorrow to buy a 50... Would you settle for the 1.8 over the 1.4 (I think you shoot Nikon so couldn't go for a 1.2(?))

I understand your point behind the theory, but bottom line in my mind really is, which would you purchase if you had to? Because by your reasoning it seems like we should save the money, get the 1.8 and not bother with the extra bit. (Unless I've missed something)

kd

 

[An Exception]

I am trying to show...no, I shown that 1.2 captured more light than 2.2.

You claimed that

That is not true. If it is not noticeable then they would look the same.

They are not.

If that statement is true then they would all have the same exposure.

They do not.

As for the ISO bumping, sorry, that is irrelevant. If they are bumped up then they would be the same exposure? As I could make it 2.0 and ISO 200. It is CLEAR that 1.8 is darker than 1.2. Who cares if there is more grain, it is darker. You said it wouldn't be. Which is not what is shown here.

Dude, you are seriously making me want to bang my head against a wall here.
Read the damn article or another article (if DXO makes you puke) so you are not completely ignorant of the theory you are seemingly trying to test.

Perhaps use the vignette profile in Lightroom for both aperture settings. Then compare the exposure. I'm assuming they will be very close if not the same.

With the above quote, I assumed you would be doing something like 1/250 F1.4 & 1/125 F2.0 to see if the exposure is the same.

Also why test 0.3 stops, why not full stops to make any difference stand out.

Of course, doing the test this way, there shouldn't be much difference according to what DXO is reporting anyway. But you seem to have the wrong end of the stick.

 

[An Exception]

I've read and now understand this, but ultimately, D.P.

If you went out tomorrow to buy a 50... Would you settle for the 1.8 over the 1.4 (I think you shoot Nikon so couldn't go for a 1.2(?))

I understand your point behind the theory, but bottom line in my mind really is, which would you purchase if you had to? Because by your reasoning it seems like we should save the money, get the 1.8 and not bother with the extra bit. (Unless I've missed something)

kd

There is more to fast apertures than letting in more light.
DOF scales as it should, so the extra aperture speed would be worthwhile for portraits.
However not everyone buy's 1.2 or 1.4 for shallower DOF (like DP).
For example take a 1.8 lens Vs a 1.2 lens and let's say your primary concern is low light performance. 1.8 Vs 1.2 is a full stop in aperture. Do you get a full stop of extra light? No. Depending on the body, you may only get 0.3-0.6 stops more light. So now you can weigh up if 0.3 stops is worth paying an extra £XXXX for.

Another consideration would be deciding to go for a fast stabilised lens or a faster Prime. The slower stabilised lens may actually outperform the faster prime.

Ultimately knowledge empowers. It enables people to make better choices, yet some posters see this negatively it seems.

 

[Raymond Lin]

I rather eat my own puke than click on a link by DXO, no thanks to you

I understand it, 1.2 let in more light than 1.4. Because I can see it. I just saw it.

Yes, you say the camera bumps it up without telling me...who cares. It may be, it may not. I am not going to waste any more time on testing it, I wasted enough already. I am not going to shoot 128000 ISO in the field anyway and to be honest...if i were shooting ISO 12800, I rather have 1.2 than 1.4 seeing I am after that every little bit of light as I can get.

You are welcome to prove me wrong though without linking to a single article, video, in particular, by that DXO site. /puke. You can do that right?

Show me some pictures, taken yourself. That 1.4 is the same as 1.8 and that 1.4 is more grainy.

 

[D.P.]

I've read and now understand this, but ultimately, D.P.

If you went out tomorrow to buy a 50... Would you settle for the 1.8 over the 1.4 (I think you shoot Nikon so couldn't go for a 1.2(?))

I understand your point behind the theory, but bottom line in my mind really is, which would you purchase if you had to? Because by your reasoning it seems like we should save the money, get the 1.8 and not bother with the extra bit. (Unless I've missed something)

kd

I settled for the 35mm, 50mm, and 85mm f1.8 because f/1.8 is plenty shallow enough for my uses and these lenses are as sharp or sharper than the f/1.4 lenses.

I don't really get why this topic is ending up so heated with Raymond Lin so upset again, especially since it has nothing to do with Canon. This topic has arisen because there was tak about Nikon releasing f/1.2 lenses to which I said it was unlikely and the benefits are not really apparent from a light gathering perspective relative to the size, weight, cost and complexity of an f/1.2 lens vs an f/1.4 lens given the high ISO capabilities of modern sensors. Fast apertures come at a cost optically as well, more elements, often higher LoCa, and it gets hard and hardwr to make a lens sharp, and to focus accurtaely with a tiny DoF. Even on the canon mount the the Bokeh can suffer mechanical vignetting so one has to be careful when shooting it wide open.

The choice of f/1.2 vs f/1.4 vs f/1.8 is really a choice about DoF, not light gathering. Raymond likes his 85L because he like shallow DoF which the lens provides.


Now to turn your question around, would you rather have the Canon 50mm f/1.2 L or the sigma 50mm f/1.4? I also expect that if Canon had an 85mm f1.4 that had blazing fast autofocus and creamier Bokeh than the 85L then it would be far more popular than the 1.2

 

[D.P.]

I am trying to show...no, I shown that 1.2 captured more light than 2.2.

You claimed that

That is not true. If it is not noticeable then they would look the same.

They are not.

If that statement is true then they would all have the same exposure.

They do not.

I don't know about you, I can tell the difference between 1.2 to 1.4, on my mac and the back of the LCD on the camera. I made a gif so you can see all the steps. 1.2 to 1.4 is not significant, it isn't as it is not a stop of light, we knew that, but it is noticeable nevertheless. It's why I still have the 50/1.4. The 1.2 is a lot of money for that tiny stop of aperture. Where as the Canon 85mm it is between the 1.8 and 1.2, which is a bigger difference.

As the for science and theory...like I said, I wanted to see for myself.

1.2 gives more light than 1.4.

That is a FACT is it not?

As for the OP, is the 50/1.2 worth it over the 50/1.4 for the aperture?

The honest answer is no. It is a lot of money for that tiny bit of difference. You would however be paying for a newer lens design, better bokeh, better build quality. A lot of which is where the money will be going, its why we buy the L, else we'll all be shooting the 35/2.0 or the 35/1.8 and the 35/1.4 would never been made and sold like hot cakes.

None of your photos show what you are trying to prove because you haven't taken into account what the camera is doing to the sensor gain. Secondly, I didn't say the f/1.2 doesn't let in more light than the f/1.4, but much less than it should do, and so does the f/1.4 lens let in less light than the f/1.8 lenses. The end result is there is diminishing returns as the aperture increases on digital sensors which will make the difference between say an f/1.2 lens compared to an f/1.4 lens much smaller than it should be in terms of aperture.

That is an undeniable fact that has been shown numerous times, no photos you can produce can disprove that. You can ignore DxOmark as much as you like but they clearly present the raws facts in this case, and they are not biasing any manufacturer at all, simply presenting an observation. Nikon, Sony and Canon, and likely all other camera exhibit this behavior as it is a basic result of the physics of light hitting reflective surfaces at oblique angles. Last week I watched every evening the sun setting behind the Pacific Ocean and as the angle of the sun lowered the special highlights increases, something completely absent in the middle of the day

 

[D.P.]

There is more to fast apertures than letting in more light.
DOF scales as it should, so the extra aperture speed would be worthwhile for portraits.
However not everyone buy's 1.2 or 1.4 for shallower DOF (like DP).
For example take a 1.8 lens Vs a 1.2 lens and let's say your primary concern is low light performance. 1.8 Vs 1.2 is a full stop in aperture. Do you get a full stop of extra light? No. Depending on the body, you may only get 0.3-0.6 stops more light. So now you can weigh up if 0.3 stops is worth paying an extra £XXXX for.

Another consideration would be deciding to go for a fast stabilised lens or a faster Prime. The slower stabilised lens may actually outperform the faster prime.

Ultimately knowledge empowers. It enables people to make better choices, yet some posters see this negatively it seems.

This really.

If you primary concern is low light performance then the difference in 1.8 vs 1.2 is much smaller than one would expect given the stated apertures.

however if your goal is shallower DoF then the f1.2 may well be a better choice.

The only caveat I personally add is that a shallow DoF doesn't necessarily lead to a smoother bokeh as so many other things matter such as lens design and complexity, subject background, distance to subject and from subject to background, focal length and just aspects of the lens such as contrast and ability to control highlights etc.

 

[Raymond Lin]

You say knowledge empowers...and allow people to make better choices. I haven't read any of the DXO stuff, yet:

I still have the Canon 85/1.8 (yes I do)
I still have the Canon 50/1.4

I don't need to read some Lab test to explain to me why I have them, why I haven't got the 50/1.2. Real world experiences tells me that:

The 85/1.2 is a bokeh machine
The 85/1.2 is a hell of a piece of engineering
The 85/1.2 gives more light that is actually useful in a real world situation over the 1.8 most of the time.
The 85/1.2 is not a perfect lens, it's slow to focus.
The 85/1.2 is £1,300 more than the 1.8

The 85/1.8 is a bargain of a lens
The 85/1.8 creates bokeh almost the same as the 1.2 when it's at 1.8
The 85/1.8 is a very sharp lens
The 85/1.8 is lightning fast to focus
The 85/1.8 would make a nice travel portrait lens due to its size
The 85/1.8 is more plastic than the 1.2

I haven't got the 50/1.2 because

I don't think that tiny bit of light is worth £800 difference.

That's it.

I would buy it because

It's a better quality lens in conatruction
It is possibly sharper, I haven't got it yet so wouldn't know now
It is physically a bigger lens so better to hold compare to the stump that is the 1.4
It has nicer bokeh

No DXO links required and those are my personal experience with them.

That's said, I want the 50/1.2 still.

Just because.

I normally just go to PoTN and look at their lens archive and look at hundreds of photos taken with a lens taken by people around the world and see what images they get. I find that is more useful than reading numbers. It certainly is less boring too.

Oh, I'm not upset about this and I don't think it's a heated discussion (you've not seen me in a heated discussion lol), I am just nauseated from the mention of DXO. Yes I know what they do and I know it is valid but the mere mention of them makes me want to sling mud at the screen.

Thanks Rhys

 

[ElDude]

On an aside don't you worry that the 50mm 1.2 has softness issues on the edges? I keep seeing people commenting that only the centre is sharp. If true (as with everything Internet they could be parroting an incorrect article they read!) isn't it a bit of a waste in that extra money? That's a nice holiday taking photos with your 1.4

 

[Raymond Lin]

I think I've said it before, I want it (rather than need) mainly because so I can complete the set lol (24,35,50,85,100,135).

The reason why I haven't is what you said, there are better things to get than to do that. Like a holiday, set up some personal shoots, a NAS for my media, save up for a house deposit or taxes even.

It doesn't stop me still wanting it though.

 

[Johnny|lucidcomposure]

This thread hasn't one bit let me fall out of love with my 85 F1.2.

To be honest I think its really interesting some people are so passionate about the technical limitations of equipment to the feather detail and I respect that so much.

For me though I (suppose) ignorantly try not to go on these sites and digest their analysis.
I would rather put that time into enjoying my passion of using the equipment and getting the shots I want.

 

[darelmiler]

I generally use sigma f/1.4 EXDGHCM lens. It give a large maximum aperture and fixed focal length prime lens. The fast maximum aperture offers several key advantages over zooms, including the ability to isolate a subject by selectively blurring the background.

 

[D.P.]

This thread hasn't one bit let me fall out of love with my 85 F1.2.

To be honest I think its really interesting some people are so passionate about the technical limitations of equipment to the feather detail and I respect that so much.

For me though I (suppose) ignorantly try not to go on these sites and digest their analysis.
I would rather put that time into enjoying my passion of using the equipment and getting the shots I want.

"This thread hasn't one bit let me fall out of love with my 85 F1.2. "

Nor should it, it is one of the best lenses Canon makes. Nothing anyone has said has been to criticize the 85L, I am very surprised that it ended up so heated and Raymond taking it very seriously. The thread is just about an explanation of a physical phenomena that results in diminishing returns for light gathering capability at ultra fast apertures that effect all manufacturers equally.



I am interested in the technical side of photography no more than I am interested in the technical side of anything I do. i am a scientists, i have a PhD, I like to know how things work and why, I spend how on Wikipedia just learning anything and everything from biology, chemistry, physics, history, economics, etc., etc. I far prefer to be out in the mountains hiking and taking photos but monday-Friday I am in the office all day so that isn't possible. What is possible is reading articles that interest me in a 5 min coffee break.

 

[CGrieves]

Nor should it, it is one of the best lenses Canon makes. Nothing anyone has said has been to criticize the 85L, I am very surprised that it ended up so heated and Raymond taking it very seriously. The thread is just about an explanation of a physical phenomena that results in diminishing returns for light gathering capability at ultra fast apertures that effect all manufacturers equally.

I am interested in the technical side of photography no more than I am interested in the technical side of anything I do. i am a scientists, i have a PhD, I like to know how things work and why, I spend how on Wikipedia just learning anything and everything from biology, chemistry, physics, history, economics, etc., etc. I far prefer to be out in the mountains hiking and taking photos but monday-Friday I am in the office all day so that isn't possible. What is possible is reading articles that interest me in a 5 min coffee break.

I'm also not sure why this got so heated. Exactly as asked, you've clarified interesting information I didn't know, and given me another dimension to consider when choosing equipment. Personally, I'd much rather know than not....

I'm guessing that as sensor technology and efficiency advances, the off-axis sensitivity might also increase? If so, from that standpoint, going for the largest aperture possible could still be viewed as a good long-term investment. Unlike computing technology, and one could argue to a certain degree camera bodies, lenses aren't prone to becoming obsolete over the short term.

Even knowing this limitation, I'd still consider the 50 f1.2. Performance and build quality aside, while I hate to admit it, the consumerist in me still gets satisfaction from pulling an L out of the bag

 

[Johnny|lucidcomposure]

"This thread hasn't one bit let me fall out of love with my 85 F1.2. "

Nor should it, it is one of the best lenses Canon makes. Nothing anyone has said has been to criticize the 85L, I am very surprised that it ended up so heated and Raymond taking it very seriously. The thread is just about an explanation of a physical phenomena that results in diminishing returns for light gathering capability at ultra fast apertures that effect all manufacturers equally.



I am interested in the technical side of photography no more than I am interested in the technical side of anything I do. i am a scientists, i have a PhD, I like to know how things work and why, I spend how on Wikipedia just learning anything and everything from biology, chemistry, physics, history, economics, etc., etc. I far prefer to be out in the mountains hiking and taking photos but monday-Friday I am in the office all day so that isn't possible. What is possible is reading articles that interest me in a 5 min coffee break.

I know you are. I find it fascinating you go into such depth.

But I think in fairness to Raymond he's just on the same fence as me as would just rather not delve into such details and just take photos. Probably not a thread we should be in.

Thanks though for such an in-depth look into the technicals.

 

[D.P.]

I'm also not sure why this got so heated. Exactly as asked, you've clarified interesting information I didn't know, and given me another dimension to consider when choosing equipment. Personally, I'd much rather know than not....

I'm guessing that as sensor technology and efficiency advances, the off-axis sensitivity might also increase? If so, from that standpoint, going for the largest aperture possible could still be viewed as a good long-term investment. Unlike computing technology, and one could argue to a certain degree camera bodies, lenses aren't prone to becoming obsolete over the short term.

Even knowing this limitation, I'd still consider the 50 f1.2. Performance and build quality aside, while I hate to admit it, the consumerist in me still gets satisfaction from pulling an L out of the bag

I don't think that much will change in the design of sensors to accommodate the off-axis light. We already have fairly advanced microlenses to try to help. I see it as one of those physical things that we just have to accept such as small apertures causing diffraction, or long lenses with fast apertures being big and heavy. In fact things might get worse as it appears to be strongly related to the pixel size so going forwards we might experience it more and more, although conceivably that is when they will try harder and harder to minimise the impact.

I also don't think the camera manufacturers worry much about it currently, there are more important things like simply improving the sensor to have better noise and Dynamic range characteristics. I am still waiting for digital sensors to get the DR and highlight headroom of good film, esp B&W film which pushes 20 stops. I often shoot scenes pushing 25 stops DR and wish I could do this in a single exposure without filters.

 

[Vertigo]

To try to boil it down, it appears that what D.P. is asserting is this:

If we take two fast apertures, one stop apart, say f/1.4 and f/2. If the same shot is taken at both of these apertures, then twice as much light is reaching the sensor at f/1.4 than at f/2.

However...

The claim is that the sensor itself isn't capable of "gathering" or absorbing all of this light, so in fact doesn't actually benefit from all of it. To counteract this effect, the camera raises the sensitivity or gain of the sensor above what is "normal" for the configured ISO, to produce the correct exposure.

The end result is the same exposure-wise but, because the sensor gain was raised, more noise may be evident in the resulting picture than if the sensor was actually able to gather all of the available light and thus didn't need to raise the sensor gain. The camera will still record the configured ISO speed in the image file, thus there will be no indication that this process took place.

Am I understanding correctly?

Not saying this is correct or incorrect, merely trying to clarify the assertion

 

[D.P.]

To try to boil it down, it appears that what D.P. is asserting is this:

If we take two fast apertures, one stop apart, say f/1.4 and f/2. If the same shot is taken at both of these apertures, then twice as much light is reaching the sensor at f/1.4 than at f/2.

However...

The claim is that the sensor itself isn't capable of "gathering" or absorbing all of this light, so in fact doesn't actually benefit from all of it. To counteract this effect, the camera raises the sensitivity or gain of the sensor above what is "normal" for the configured ISO, to produce the correct exposure.

The end result is the same exposure-wise but, because the sensor gain was raised, more noise may be evident in the resulting picture than if the sensor was actually able to gather all of the available light and thus didn't need to raise the sensor gain. The camera will still record the configured ISO speed in the image file, thus there will be no indication that this process took place.

Am I understanding correctly?

Not saying this is correct or incorrect, merely trying to clarify the assertion

That is spot on.

There should be 1 stop of extra light reaching the sensor with an f/1.4 vs f/2.0 lens but depending on the sensor up to 0.5 a stop of light will be lost. With f/1.2 lenses wide open up to 1 complete stop of light is lost, as per the DXOmark link. The physical aperture is real and so the DoF is appropriately witnessed, also the exposure meter will also accordingly react as if there was the full amount of light absorbed by the sensor.


Raymond claims he doesn't care because he doesn't see it, which is a fine attitude. But with the same logic I could claim my f/1.8 prime is actually an f/0.7 prime because if I raise the ISO and print small enough then i also wont see any negative downside and my lens is acting like a much faster lens in the context of gathering light. Could I sell my f/1.8 lens as a f/0.7 lens for thousands of pounds and tell people to simply up their ISO?