Interesting analysis and conclusions by DxO regarding loss of light gathering ability at wide apertures (1.2 and 1.4) as a function of sensor design and how camera makers get around this to give correct exposures.
the effect of a sensor on light gathering power to light rays that aren't perpendicular to the sensor plane isn't exactly a secret. the effective aperture of the sensor's wells is a well known design-time specification and that is why the big deal over back-illuminated sensor fabrication techniques. same with the falloff from the lens axis that is steeper than the lens's falloff on film. it's the overall compensation for it based on aperture that isn't well known.
cogitech wrote:
The conclusion of the DxO chief scientist: "I am therefore inclined to question the real benefit of faster lenses.” - Spoken like a true scientist.
Hmm. I wonder what he considers to be the "true benefit of faster lenses" to be.
well if the idea is to shoot a lower ISO with a wider aperture to preserve noise then what is the point? it's not the only or even necessarily the main use of fast lenses but it is in fact one justification for having one.
HerbChong wrote:
well if the idea is to shoot a lower ISO with a wider aperture to preserve noise then what is the point? it's not the only or even necessarily the main use of fast lenses but it is in fact one justification for having one.
Herb...
So, if you remove that one benefit, would you be inclined to "question the real benefit of faster lenses"?
I don't think much people on the alt forum are looking for extremely high speed, most are looking for short depth of field.
And I don't get their conclusions either. Yes a T stop is not a F stop, and the T stop is different on each camera. But on a given camera a f/1.4 lens will still be faster and have a shorter depth of field than an f/2 lens, and that's all that matters.
Finally I don't think there is any influence on the depth of field, and if there is one its easy to prove: shoot the same lens on a film camera, and one - or several - digital camera. I'm still waiting for evidences rather than cryptic maths =p
Indeed. In this age of ISO 12,000,000 (or whatever it is at now) I really do not think anyone can imagine that speed is the only (or even primary) benefit of large aperture lenses (except scientists, I guess).
edwardkaraa wrote:
Having shot both film and digital until recently (2 years ago) I too have noticed shallower DOF and more vignetting with digital.
This is a result of the 'depth' or thickness of the film's emulsion which helps record rays at slightly different focal distances points as being focused.
IIRC this affects the presence of CA too.
To the comentators on horseback - good on you. I never said this revolutionary info.
I'm surprised that no one has brought up the point that most DSLR's increase the sensor gain to compensate for the loss of light at wide apertures (anything below f/4). This only works with electronically coupled lenses of course.
It (the thread title) is also one of the few remaining "very good" reasons for going recommending FF/FX cameras, since the actual, TRUE gain you get is even bigger than what just the area theory says when you shoot at large apertures.
Modern high-density APS cameras loose so much light at large apertures that they're almost a full stop down in T-stop measurements.
This means that a 5D2 with a 135F/2L @ F/2.4 (!) actually gathers more light than a 7D with an 85L @ F/1.2.... So if you REALLY need the low-light capability at medium-short reach, then FF is a much better and also CHEAPER alternative!
The 135F/2 w. 5D2 gathers almost 2/3Ev more light than the 7D w. 85F/1.2L.
The 85F/1.8 w. 5D2 gathers almost 2/3Ev more light than the 7D w. 50F/1.2L.
The 50F/1.8 w. 5D2 gathers almost a full Ev more light than the 7D w. 35F/1.4L.
Not to mention that the FF/FX combos also will be sharper, cheaper and sometimes also lighter in weight...
When you get to F/2.8+ this advantage is gone, though. So if the main bulk of shooting is at F/4.0 or higher, then reach and lens availability is of more importance than sensor size when choosing your main system.
AhamB wrote:
I'm surprised that no one has brought up the point that most DSLR's increase the sensor gain to compensate for the loss of light at wide apertures (anything below f/4). This only works with electronically coupled lenses of course.
I am baffled and confused by this.
Consider two faster-than-f/4 lenses with identical T-stop, one manual focus lens, and one lens with aperture coupling (electronic or otherwise). With constant camera ISO and shutter speed settings, photos made with each lens will have different exposures? If camera 'knows' a fast lens is attached and aperture setting is less than f/4 the ISO (sensor sensitivity, gain) is boosted so my ISO 100 photo is actually ISO 160 (or whatever boost is)? Is there is a noise trade-off in low light, the manual focus lens requires longer exposure time (more noise), the auto lens less exposure time but more noise due to increased sensor gain? Is sensor gain increased or is sensor signal multiplied by some factor by hardware/software?
Now! You see why many EF 50mm f1.2L users complained about the lens is not much better than the EF 50mm f1.4 or even the cheapy EF 50mm f1.8?
Only those who have the extra cash and not know how to spend it will buy into the marketing gimmick.
Even at the local Canon Regional Head Office, they use the very reliable and cheapy EF 50mm f1.8 II for testing. Or maybe they can't afford to use the EF 50mm f1.2L , for those who think that if someone don't use it is because they can't afford it.
For those who use the AIS 50mm f1.2 on film cameras will be a different story.
Now you see why Nikon or Sigma never bother to come up with a 50mm f1.2 for DSLR? Maybe the difference is just too minute.
Time to go horseback riding for more information and shoot moving targets.
melanopsin wrote:
I am baffled and confused by this.
Why? It is all explained in the article. There is nothing baffling or confusing about it. It is simple physics.
Consider two faster-than-f/4 lenses with identical T-stop, one manual focus lens, and one lens with aperture coupling (electronic or otherwise). With constant camera ISO and shutter speed settings, photos made with each lens will have different exposures?
Yes, exactly, because the camera increases the gain to compensate for light loss.
If camera 'knows' a fast lens is attached and aperture setting is less than f/4 the ISO (sensor sensitivity, gain) is boosted so my ISO 100 photo is actually ISO 160 (or whatever boost is)?
Yes.
Is there is a noise trade-off in low light, the manual focus lens requires longer exposure time (more noise), the auto lens less exposure time but more noise due to increased sensor gain?
6 of one and half a dozen of the other.
Is sensor gain increased or is sensor signal multiplied by some factor by hardware/software?
hans98ko wrote:
Now! You see why many EF 50mm f1.2L users complained about the lens is not much better than the EF 50mm f1.4 or even the cheapy EF 50mm f1.8?
No, I don't see this at all. I don't think this has anything to do with someone concluding that the 1.2L is not better than the other Canon 50s.
Only those who have the extra cash and not know how to spend it will buy into the marketing gimmick.
I completely disagree. You seem to think that more light is the only benefit of faster lenses. It is not a gimmick at all. Are you a scientist, perchance?
Even at the local Canon Regional Head Office, they use the very reliable and cheapy EF 50mm f1.8 II for testing.
For testing what, and how is that relevant at all to the discussion at hand?
Or maybe they can't afford to use the EF 50mm f1.2L , for those who think that if someone don't use it is because they can't afford it.
I can't afford the EF 50/1.2 L, but this does not mean I somehow try to explain away the benefits of f1.2 lenses. Instead, I buy and use f1.2 lenses that I can afford.
For those who use the AIS 50mm f1.2 on film cameras will be a different story.
Exactly 0.25 Ev different than my 5D + Rokkor 58/1.2. I can correct for this with EC or ISO bump and still have a much cleaner, more detailed result.
Now you see why Nikon or Sigma never bother to come up with a 50mm f1.2 for DSLR?
cogitech wrote:
No, I don't see this at all. I don't think this has anything to do with someone concluding that the 1.2L is not better than the other Canon 50s.
I thought that you have read and understood the link well regarding about wider aperture opening does not mean more light falling onto the sensor (depends on micro lens convex angle and the dept of the light well). And also about the gain boost adjustment by each individual manufacturers (the slight aperture advantage can always be taken care by gain boost). Also there is this so call CA error kicking into effect.
Regarding the other disagreements, I can always agree to disagree with each other. No one are born equal even for twins.
But one thing we have in common and that is we still kept our antique EOS 5D, almost sold it if not for the low price it can fetch. So decided to keep it as a backup and snap shooter.
hans98ko wrote:
I thought that you have read and understood the link well regarding about wider aperture opening does not mean more light falling onto the sensor (depends on micro lens convex angle and the dept of the light well). And also about the gain boost adjustment by each individual manufacturers (the slight aperture advantage can always be taken care by gain boost). Also there is this so call CA error kicking into effect.
Apparently you, like the scientist at DxO, assume that increased light is the only benefit of large aperture lenses. I know this is not the case, therefore I disagree with any conclusions based on that assumption.
