We investigate the case where we want the same shutter speed.
We somewhat agree that for sensors with the same efficiency per area:
45 mm f/1.4 is equivalent to 90 mm f/2.8 on FF.
If a u43 user can tolerate more noise than the FF user, he may want to use a slower lens on u43, and say use a 45 mm f/2 instead.
If a FF users can tolerate more noise than the u43 user, he may want to use a slower lens on FF, and say use a 90 mm f/4.0. This is suddenly absurd? ...Show more →
I get what you mean. Here's another scenario, in terms of aps-c and 135. If someone made an awesome new 23mm f.9 lens for my NEX-5N, it wouldn't give me any advantage over a 35mm f1.4 on my D800, because if I shot the new 23/.9 at ISO 200 on the 5N, I'd get similar DOF and noise results with the D800 and the 35/1.4 at ISO 400. So, the 23/.9 isn't all THAT exotic, in terms of performance available on other formats.
When calculating exposure in the field, it obviously makes senses to call an f2 lens f2, but when comparing lenses from different systems, the equivalents make more sense. So, everyone is right!
douglasf13 wrote:
I get what you mean. Here's another scenario, in terms of aps-c and 135. If someone made an awesome new 23mm f.9 lens for my NEX-5N, it wouldn't give me any advantage over a 35mm f1.4 on my D800, because if I shot the new 23/.9 at ISO 200 on the 5N, I'd get similar DOF and noise results with the D800 and the 35/1.4 at ISO 400. So, the 23/.9 isn't all THAT exotic, in terms of performance available on other formats.
When calculating exposure in the field, it obviously makes senses to call an f2 lens f2, but when comparing lenses from different systems, the equivalents make more sense. So, everyone is right! ...Show more →
I agree....and where did you already get your D800?
douglasf13 wrote:
Ha! I wish. Just used that as a convenient example, since the sensors of the D800 and 5N are supposedly similar at the pixel level.
Yeah, I know..but there is always that outside chance.
I'm thinking 5N sensor performance on FF should be outstanding.
On another note, Olympus has apparently gone out of their way to state that the E-M5 is not a professional Olympus camera, even calling into question it's reliability for that use. I guess attaching the "OM" name to it caused some confusion. Reasons from the recent Olympus interview here
No. It is a camera for advanced amateurs. Logically, certainly also some professionals use because their quality and performance are really good. But I still do not think it enough for us as we raise professional camera range.
What is missing?
Most important for the professional photographer is the reliability of the machine. And you also need more configuration options, such as in the E-5 , our professional SLR range."
and about the EVF in the E-M5 Olympus says:
"For example, the electronic viewfinder. This we have used in the E-M5 is really good, but we are convinced that there is still room for improvement in resolution or response."
douglasf13 wrote:
When calculating exposure in the field, it obviously makes senses to call an f2 lens f2, but when comparing lenses from different systems, the equivalents make more sense. So, everyone is right!
Agree!
The term that needs a little clarification is "light gathering". We should be a little more specific on when we mean "light gathering per sensor area" or "light gathering per image".
Light gathering per sensor area is exposure. Light gathering per image is exposure*sensor area. So exposure alone is only an operational parameter. If the sensor area is 4 times as large, the light gathered per image is 4 times as much for the same exposure.
This was seen from the sensor perspective.
When seen from the lens perspective, f2 is still f2. But the f-number is a derivative number. The physical properties of the lens are the focal length and the diameter of the entrance pupil.
A 50 mm f/2 lens is a 50 mm Focal Length 25 mm Entrance Pupil lens.
A 100 mm f/2 lens is a 100 mm Focal Length 50 mm Entrance Pupil lens
From the lens perspective we see exactly the same as from the sensor perspective. In terms of exposure they are equal, but in terms of light gathering per image, the 50 mm entrance pupil lens gathers 4 times as much light as the 25 mm entrance pupil lens.
douglasf13 wrote:
Here's another scenario, in terms of aps-c and 135. If someone made an awesome new 23mm f.9 lens for my NEX-5N, it wouldn't give me any advantage over a 35mm f1.4 on my D800, because if I shot the new 23/.9 at ISO 200 on the 5N, I'd get similar DOF and noise results with the D800 and the 35/1.4 at ISO 400. So, the 23/.9 isn't all THAT exotic, in terms of performance available on other formats.
Whe we take the problem with angle of incidence into account, the f/0.9 lens will not effectively be 1 stop faster than the f/1.4 lens when captured by the the same pixels, maybe just 0.5 stops faster or even less:
The term that needs a little clarification is "light gathering". We should be a little more specific on when we mean "light gathering per sensor area" or "light gathering per image".
Light gathering per sensor area is exposure. Light gathering per image is exposure*sensor area. So exposure alone is only an operational parameter. If the sensor area is 4 times as large, the light gathered per image is 4 times as much for the same exposure.
This was seen from the sensor perspective.
When seen from the lens perspective, f2 is still f2. But the f-number is a derivative number. The physical properties of the lens are the focal length and the diameter of the entrance pupil.
A 50 mm f/2 lens is a 50 mm Focal Length 25 mm Entrance Pupil lens.
A 100 mm f/2 lens is a 100 mm Focal Length 50 mm Entrance Pupil lens
From the lens perspective we see exactly the same as from the sensor perspective. In terms of exposure they are equal, but in terms of light gathering per image, the 50 mm entrance pupil lens gathers 4 times as much light as the 25 mm entrance pupil lens. ...Show more →
You're just describing why full frame needs larger lenses.
alundeb wrote:
I wonder how the new 16 MP u43 sensors behave with respect to this for lenses faster than f/2.8.
No problems at all at least down to 1.4. Tested with multiple lenses, many with no electronic communication to the body. I'm not sure if anyone tested a 0.95 lens, but I recall a 1.2 being done. All showed no reduction in exposure at the center of the frame to within measurement error (about 0.1 EV).
hauxon wrote:
You're just describing why full frame needs larger lenses.
I am describing why you need larger lenses to gather more light per image. To gather the same light per image as a 100 mm Focal length 50 mm Entance pupil, you need a 50 mm Focal Length 50 mm entrance pupil, or 50 mm f/1.
Lens size and weight comparison between formats for equivalent lenses give different results depending on the spcific case. The 75 mm f/1.8 for u43 will likely be smaller and lighter than a 150 mm f/3.5.
kwalsh wrote:
No problems at all at least down to 1.4. Tested with multiple lenses, many with no electronic communication to the body. I'm not sure if anyone tested a 0.95 lens, but I recall a 1.2 being done. All showed no reduction in exposure at the center of the frame to within measurement error (about 0.1 EV).
alundeb wrote:
To gather the same light per image as a 100 mm Focal length 50 mm Entance pupil, you need a 50 mm Focal Length 50 mm entrance pupil, or 50 mm f/1.
Is that true? No right? Yeah, I'm pretty sure it's not.
The reason people a FF/100/2 is equal to a µ43/50/1 if because of the DOF and not exposure values. Exposure values remain the same. A 100/1 at ISO200 properly exposing a white piece of paper will be 1/1000 (for example) on both a FF, APS-C, and µ4/3 - theoretically speaking. In practice there may be a tiny bit of difference here or there depending on the sensor tech and camera electronics but not counting that - it'll be the same.
Bifurcator wrote:
Is that true? No right? Yeah, I'm pretty sure it's not.
The reason people a FF/100/2 is equal to a µ43/50/1 if because of the DOF and not exposure values. Exposure values remain the same. A 100/1 at ISO200 properly exposing a white piece of paper will be 1/1000 (for example) on both a FF, APS-C, and µ4/3 - theoretically speaking. In practice there may be a tiny bit of difference here or there depending on the sensor tech and camera electronics but not counting that - it'll be the same.
Or did I misunderstand your example?
No, you did not misunderstand. It is just that I refuse to use the same exposure on different formats, becuase it is exposure*sensor area that matters for image quality, not exposure alone.
If we have the same number of pixels on the small and the large sensor, for the same exposure, the numbers of photons collected per pixel is four times as much with the larger sensor. The dominating noise source is photon shot noise, and that is the square root of the number of photons collected.
If I reduce the exposure for the large sensor, by using the same entrance pupil diameter (hence a different f-stop) and a 4 times higher ISO setting (to get the same shutter speed), I end up with collecting just as many photons in the image as the smaller sensor with 4 times better exposure.
That doesn't make any sense to me. You make it sound as if the photons have an awareness level and decide to spread themselves out when they see a larger sensor and/or lock arms when they see a small one.
I could be mistaken I guess but I'm pretty sure your thinking on this is incorrect. The results of two photographs compared seem to say so too. Got any (unedited!) examples of this phenomena?
alundeb wrote:
If I reduce the exposure for the large sensor, by using the same entrance pupil diameter (hence a different f-stop) and a 4 times higher ISO setting (to get the same shutter speed), I end up with collecting just as many photons in the image as the smaller sensor with 4 times better exposure.
so what you're saying is that if I'm looking at 2 printed photos one from m4:3 and one from a old 5D that have been shot with the same f stop same iso same shutter speed, the smaller format photo will be 4 times worse in terms of image quality?
Ok I'm not the most observant viewer, but in real life I just dont see this.
Spyro P. wrote:
so what you're saying is that if I'm looking at 2 printed photos one from m4:3 and one from a old 5D that have been shot with the same f stop same iso same shutter speed, the smaller format photo will be 4 times worse in terms of image quality?
Ok I'm not the most observant viewer, but in real life I just dont see this.
unless I misunderstood?
Interesting question.
1) Signal to noise ratio does not depend linearly with captured light, but with the square root. So technically, it will only be 2 times worse.
2) When the noise is low, the reduction from little noise to very little noise is barely noticeable. This is the point many u43 users are making.
Also, the ratios above are only really valid for the same sensor tech. Newer sensors are more efficient at gathering light.
Other comes a point where the technology is good enough for pretty much all your needs. Those needs will be different for every photographer. For me, m43 fulfills it for pretty much everything I shoot.
Just used that as a convenient example, since the sensors of the D800 and 5N are supposedly similar at the pixel level.