F-stop blues

[JDittlen]

http://www.dxomark.com/index.php/Publications/DxOMark-Insights/F-stop-blues is indicating that going wider than about f/2 is of no benefit for lowering ISO or speeding up the shutter on APS-C systems. Full frame is less affected. Does any of you know if something similar applies to MFT, other mirrorless, or enthusiast compacts?

Best regards,
Johan

 

[Detail Man]

JDittlen wrote:

http://www.dxomark.com/index.php/Publications/DxOMark-Insights/F-stop-blues is indicating that going wider than about f/2 is of no benefit for lowering ISO or speeding up the shutter on APS-C systems. Full frame is less affected. Does any of you know if something similar applies to MFT, other mirrorless, or enthusiast compacts?

“We can suspect,” Guichard continued, “that sensors collect the incoming light all the more improperly, in that this light comes from a more oblique angle. Since faster lens have, by definition, a wider opening, they raise the proportion of oblique light, hence the proportion of lost energy which never lands on the pixels.”

Good evidence that life, in all of it's myriad enchantments towards distractions, is indeed more a "state of mind" than a "state of matter/energy". The triumph of form over function is complete

 

[hello world]

Now this is something that DPR should incorporate in their Lens reviews... a T-stop value for the lens instead of an F-stop. That would actually make a whole lot of sense. Should be measurable...
B

 

[Hen3ry]

...can be done with infrared.

Seriously, the pursuit of the the biggest aperture goes beyond reason. As for whether a bigger aperture has more effect on one size sensor than on another, that's just busy work, like digging holes then filling them in again.

Cheers, geoff
--
Geoffrey Heard
http://pngtimetraveller.blogspot.com/2011/10/return-to-karai-komana_31.html

 

[Detail Man]

Hen3ry wrote:
...can be done with infrared.

Seriously, the pursuit of the the biggest aperture goes beyond reason. As for whether a bigger aperture has more effect on one size sensor than on another, that's just busy work, like digging holes then filling them in again.

Hi Geoff ,

Seems pretty relevant to me (whether one is virtually wasting their valuable resources on little but an illusion) - unless the acquisition is all about external appearances (the other kind of "optics").

DM

 

[hello world]

I for one would be interested if investing in a F1.4 lens or so would actually make sense considdering that article.

You pay a heck of a lot more for a fast prime than a 1 stop slower lens. If both are effectively able to bring the same amount of light to the sensor but with the faster lens the software in the camera 'cheats' you, it wouldn't really make sense to invest in that, would it?

I realize it also comes down to pixel density, and more MP on the same surface is always a point of discussion - I guess everyone who understands it would pick a 12 MP sensor over a 16 MP sensor if surface area and technology are equal..., but explain that to the marketing guys... I guess they damn well know it, but the generic consumer does not, simply because "more" equals "better" in the human mind.

And now we also have to take the loss of light of the lens in consideration as apperantly - if that article may be trusted - it is a software cheating game.

Anyhow it can also be that our slow lenses, say F3.5 do not let in that amount of light either. Who knows.

Would make sense to have lens reviews that test the effective amount of light a lens can let through to the sensor area.

 

[Tim F 101]

Are you suggesting that people who use smaller sensors suffer from some collective psychosis that leads them to think falsely that their pics get brighter from f/2.8 to f/1.7? No, it really happens. Honest. You will not get as much benefit from a wide aperture in terms of narrow DOF but the incident light that just makes it through the PL25mm at f/1.4 really does strike the sensor and register a signal. If anyone really thinks it is necessary I can set up my GH2 with a PL20 in a controlled environment (my living room) and take a series of pics with everything kept constant but the aperture.

 

[Frank B]

I would be interested.
--
Frank B
Nikon P7100 Gallery
http://www.pbase.com/frank_b/nikon_p7100
All Galleries
http://www.pbase.com/frank_b

 

[bobn2]

http://www.dxomark.com/index.php/Publications/DxOMark-Insights/F-stop-blues is indicating that going wider than about f/2 is of no benefit for lowering ISO or speeding up the shutter on APS-C systems. Full frame is less affected. Does any of you know if something similar applies to MFT, other mirrorless, or enthusiast compacts?

This effect relates to the speed of the microlenses. In short, the microlenses on the sensor must have an f-number less than the f-number of the taking lens times by the linear proportion of the sensor area that is actually sensitive to light. That's a bit convoluted, so perhaps numerical example will help. Imagine we want to get full benefit from an f/1.4 lens, and the photodiodes on the sensor occupy 70% (linear not area) of the pixel, then the microlens must be faster than f/1.4 times 0.7, = f/1.

The f-number is the focal length divided by the aperture diameter. The aperture diameter is set by the pixel width and the focal length by the depth of the wiring on top of the chip. So, to work well with fast lenses a sensor must have big pixels or shallow wiring or a combination of both (the advantage of bak side illumination, by the way, is that it reduces the wiring depth to 0, so allows very fast microlenses.).

So far as FT and MFT is concerned, DxO did not give any figures, but one of the advantages of the maicovicon (aka 'Live MOS') pixel design is that it has a low wiring depth, which might compensate for the small pixels that go with a smaller sensor. The GH series (and, I think the G3) use a classical CMOS structure, not maicovicon, and I don't know whether Panasonic has managed to keep down the wiring depth with those, too.
--
Bob

 

[hello world]

To quote;

1.
Loss of light at wider aperture

“We have been very surprised,” explained Frédéric Guichard, chief scientist at DxO Labs, “to find out that some of the gain from wider lens openings seems to be offset by the present state of sensor technology. Our measurements all point in the same direction: as you go further than f/ 4 – to f 2 and wider, the accrued quantity of light falls marginally onto the sensor. A stronger and stronger part of this additional light is blocked or lost. I am therefore inclined to question the real benefit of faster lenses.”

2.
Raising ISO to compensate light loss

A further assessment is still more troubling. It seems that many camera makers are raising the ISO sensitivity when the lens opening is at its widest, probably to cover up the lack of efficiency. “We have ascertained,” explained Guichard, “that on many cameras, the ISO is pushed up on the order of 1/3 to 1/2 stop when the lens opening comes close to its upper limit.

Since the camera will most likely correct the loss of light by raising it's ISO in secret it makes no sense to test with a camera attached

It would take a hell of a lot equipment and a labratory to accurately calculate the amount of light which SHOULD hit the sensor area based on the size of aperture and the amount of light, and the amount of light which actually HITS the sensor area.

 

[Detail Man]

bobn2 wrote:

So far as FT and MFT is concerned, DxO did not give any figures, but one of the advantages of the maicovicon (aka 'Live MOS') pixel design is that it has a low wiring depth, which might compensate for the small pixels that go with a smaller sensor. The GH series (and, I think the G3) use a classical CMOS structure, not maicovicon, and I don't know whether Panasonic has managed to keep down the wiring depth with those, too.

The newly developed 16.05-megapixel multi-aspect Live MOS sensor adopts the high speed digital νMaicovicon. In addition to the high speed, which is an advantage of the CMOS sensor, Panasonic adds its original pixel mix technology. A dedicated process exclusively for νMaicovicon of the DMC-GH2 is newly developed by applying the top-class high performance transistor and its interconnection process needed for the high speed operation of digital circuit to make low-noise photodiode and transistor with reduced wiring to achieve both high resolution and high aperture-ratio as the DMC-GH1 .

From the GH2 "16.05-megapixel Live MOS Sensor" section at:

http://panasonic.net/avc/lumix/systemcamera/gms/gh2/high_speed.html
.

The newly developed 16.0-megapixel Live MOS sensor for the DMC-G3 plays a big role in high quality image rendering .

From the G3 "16.0-megapixel Live MOS Sensor" section at:

http://panasonic.net/avc/lumix/systemcamera/gms/g3/high_image.html

 

[bobn2]

bobn2 wrote:

So far as FT and MFT is concerned, DxO did not give any figures, but one of the advantages of the maicovicon (aka 'Live MOS') pixel design is that it has a low wiring depth, which might compensate for the small pixels that go with a smaller sensor. The GH series (and, I think the G3) use a classical CMOS structure, not maicovicon, and I don't know whether Panasonic has managed to keep down the wiring depth with those, too.

The newly developed 16.05-megapixel multi-aspect Live MOS sensor adopts the high speed digital νMaicovicon. In addition to the high speed, which is an advantage of the CMOS sensor, Panasonic adds its original pixel mix technology. A dedicated process exclusively for νMaicovicon of the DMC-GH2 is newly developed by applying the top-class high performance transistor and its interconnection process needed for the high speed operation of digital circuit to make low-noise photodiode and transistor with reduced wiring to achieve both high resolution and high aperture-ratio as the DMC-GH1 .

From the GH2 "16.05-megapixel Live MOS Sensor" section at:

http://panasonic.net/avc/lumix/systemcamera/gms/gh2/high_speed.html

Yet Chpworks found it to be conventional CMOS. Perhaps the trade name 'Maicovicon' has been redefined to mean 'made by Panasonic'.

The newly developed 16.0-megapixel Live MOS sensor for the DMC-G3 plays a big role in high quality image rendering .

From the G3 "16.0-megapixel Live MOS Sensor" section at:

http://panasonic.net/avc/lumix/systemcamera/gms/g3/high_image.html

Live MOS is actually an Olympus trade mark. I think it means 'made for a Four Thirds camera'.
--
Bob

 

[Detail Man]

bobn2 wrote:

So far as FT and MFT is concerned, DxO did not give any figures, but one of the advantages of the maicovicon (aka 'Live MOS') pixel design is that it has a low wiring depth, which might compensate for the small pixels that go with a smaller sensor. The GH series (and, I think the G3) use a classical CMOS structure, not maicovicon, and I don't know whether Panasonic has managed to keep down the wiring depth with those, too.

The newly developed 16.05-megapixel multi-aspect Live MOS sensor adopts the high speed digital νMaicovicon. In addition to the high speed, which is an advantage of the CMOS sensor, Panasonic adds its original pixel mix technology. A dedicated process exclusively for νMaicovicon of the DMC-GH2 is newly developed by applying the top-class high performance transistor and its interconnection process needed for the high speed operation of digital circuit to make low-noise photodiode and transistor with reduced wiring to achieve both high resolution and high aperture-ratio as the DMC-GH1 .

From the GH2 "16.05-megapixel Live MOS Sensor" section at:

http://panasonic.net/avc/lumix/systemcamera/gms/gh2/high_speed.html

Yet Chpworks found it to be conventional CMOS. Perhaps the trade name 'Maicovicon' has been redefined to mean 'made by Panasonic'.

Interesting. I guess that I will "buy that" (since I cannot myself afford to buy their reports) at:

http://chipworks.force.com/catalog/ProductDetails?sku=PAN-DMC-GH2-K_Pri-Camera&viewState=DetailView

The newly developed 16.0-megapixel Live MOS sensor for the DMC-G3 plays a big role in high quality image rendering .

From the G3 "16.0-megapixel Live MOS Sensor" section at:

http://panasonic.net/avc/lumix/systemcamera/gms/g3/high_image.html

Live MOS is actually an Olympus trade mark. I think it means 'made for a Four Thirds camera'.

I suppose it has a better cognitive resonance to it than "Moribund MOS" might evoke ! ...

Yet another of my "bubbles" has been popped by this realization that (nearly) everybody involved in my life is in some way knowingly and willfully deceiving me. In the US we have a "Federal Trade Commission" to provide consumers the comfortable illusion that "truth in advertising" is required ...

However, now that the US Supreme Court (in the ironically named "Citizens United" decision) has ruled that "corporations are of the same status as individual persons, it appears that they, too are "under no obligation to tell the truth unless under oath in a court of law" ...

 

[micksh6]

Raising ISO to compensate light loss

A further assessment is still more troubling. It seems that many camera makers are raising the ISO sensitivity when the lens opening is at its widest, probably to cover up the lack of efficiency. “We have ascertained,” explained Guichard, “that on many cameras, the ISO is pushed up on the order of 1/3 to 1/2 stop when the lens opening comes close to its upper limit.

Since the camera will most likely correct the loss of light by raising it's ISO in secret it makes no sense to test with a camera attached

It would take a hell of a lot equipment and a labratory to accurately calculate the amount of light which SHOULD hit the sensor area based on the size of aperture and the amount of light, and the amount of light which actually HITS the sensor area.

It's easy. Just mount F1.4 legacy lens. Camera won't know the aperture and won't change ISO.
Make 3 shots of the same object with the same exposure -
a) F1.4 ISO 200
b) F2.0 ISO 200
c) F2.0 ISO 400
(Or, better, keep ISO the same and change exposure)

Measure intensity of the same area. Ideally shots (a) and (c) should have the same intensity. In practice shot (a) may be darker.

If this happens all it means that lens T-number doesn't change linearly with F-number. It doesn't necessarily depend on sensor, it can be function of the lens.

F-stops should still be correct for estimating DOF. For light intensity T-stops should be used.

"This loss seems to increase when the pixel size decreases, as shown on the figure below"

The key word is "seems". It's not necessarily true, it depends on what lenses they tested on APS-C and FF systems. Some lenses will lose more light than the others. It's possible that FF lenses are less affected. And then there is such thing as vignetting. Was it taken in consideration?

 

[Tim F 101]

I just put my GH2 on the tripod with a manual OM 50mm 1.8 and took some pics of the living room drapes. Manual mode at ISO 800, 1/15 sec, at 1 f-stop intervals from 4 to 1.8 . The histogram tells me that each stop gained me about 1 stop more light all the way down to 1.8. Focus got noticeably less sharp at 1.8, but that is not much of a surprise nor a big departure from how it behaves on a film camera. Conclusion: unless the camera somehow can tell that an old manual lens is stopped down (it can't), you get more light from a wide aperture. I have no way to show that it works below 1.8, but I would guess people who use the Voigtlander 25 might have useful input.

 

[tedolf]

Detail Man wrote:
[snip]

Good evidence that life, in all of it's myriad enchantments towards distractions, is indeed more a "state of mind" than a "state of matter/energy". The triumph of form over function is complete

is he talking about?

Tedolph

 

[tedolf]

Hen3ry wrote:
...can be done with infrared.

Seriously, the pursuit of the the biggest aperture goes beyond reason. As for whether a bigger aperture has more effect on one size sensor than on another, that's just busy work, like digging holes then filling them in again.

Hi Geoff ,

Seems pretty relevant to me (whether one is virtually wasting their valuable resources on little but an illusion) - unless the acquisition is all about external appearances (the other kind of "optics").

DM

what is he saying!

TEdollph

 

[tedolf]

bobn2 wrote:

So far as FT and MFT is concerned, DxO did not give any figures, but one of the advantages of the maicovicon (aka 'Live MOS') pixel design is that it has a low wiring depth, which might compensate for the small pixels that go with a smaller sensor. The GH series (and, I think the G3) use a classical CMOS structure, not maicovicon, and I don't know whether Panasonic has managed to keep down the wiring depth with those, too.

The newly developed 16.05-megapixel multi-aspect Live MOS sensor adopts the high speed digital νMaicovicon. In addition to the high speed, which is an advantage of the CMOS sensor, Panasonic adds its original pixel mix technology. A dedicated process exclusively for νMaicovicon of the DMC-GH2 is newly developed by applying the top-class high performance transistor and its interconnection process needed for the high speed operation of digital circuit to make low-noise photodiode and transistor with reduced wiring to achieve both high resolution and high aperture-ratio as the DMC-GH1 .

From the GH2 "16.05-megapixel Live MOS Sensor" section at:

http://panasonic.net/avc/lumix/systemcamera/gms/gh2/high_speed.html

Yet Chpworks found it to be conventional CMOS. Perhaps the trade name 'Maicovicon' has been redefined to mean 'made by Panasonic'.

Interesting. I guess that I will "buy that" (since I cannot myself afford to buy their reports) at:

http://chipworks.force.com/catalog/ProductDetails?sku=PAN-DMC-GH2-K_Pri-Camera&viewState=DetailView

The newly developed 16.0-megapixel Live MOS sensor for the DMC-G3 plays a big role in high quality image rendering .

From the G3 "16.0-megapixel Live MOS Sensor" section at:

http://panasonic.net/avc/lumix/systemcamera/gms/g3/high_image.html

Live MOS is actually an Olympus trade mark. I think it means 'made for a Four Thirds camera'.

I suppose it has a better cognitive resonance to it than "Moribund MOS" might evoke ! ...

Yet another of my "bubbles" has been popped by this realization that (nearly) everybody involved in my life is in some way knowingly and willfully deceiving me. In the US we have a "Federal Trade Commission" to provide consumers the comfortable illusion that "truth in advertising" is required ...

However, now that the US Supreme Court (in the ironically named "Citizens United" decision) has ruled that "corporations are of the same status as individual persons, it appears that they, too are "under no obligation to tell the truth unless under oath in a court of law" ...

Now he's quoting the Citizens United decision.

Somebody please kill me now before he posts again.

TEdolph

 

[Tim F 101]

The idea is that expensive glass is like paying extra for the Leica version of a LX3. It gives you status in a red dot without adding anything meaningful. If you dug through the prose long enough you might find a nigh pathological need to patronize people from a position of superiority justified by florid prose and protected by an attitude of amused detachment, but life is too short for crap like that.

 

[kenw]

I did test a legacy 50/1.4 on my G1. It was exactly one stop from 1.4 to 2.0 as far as the RAW files went. Again, a legacy lens so no way for the camera to auto-boost the ISO.

As someone else said, someone with a 0.95 lens might give it a try...
--
Ken W
See plan in profile for equipment list