Crop Factor, Low Light and Aperture with m4/3 lenses?

Started Jun 15, 2014 | Discussions
Anders W Forum Pro • Posts: 21,468
Re: Nice summary but the supplement needs improvement

tokumeino wrote:

You forgot the Handevision Ibelux 40/0.85 ? Shame on you !

Sorry not to know extensively the M43 lineup. In fact I was mainly thinking of native AF lenses. Because if you come to manual ones, then you'll find much faster than f/1.8 FF lenses as well.

Yes, but the manual lenses you can put on FE mount are not designed for the system. These manual lenses are.

Personally though, as I said, I am happy to make do with lenses in the range from f/1.7 to f/2.0. That's as much equivalence as I need.

I was actually thinking of lenses like the Pana 25/1.4 which is 1 stop behind a Sony FE 55/1.8.

But don't misunderstand me : I totally agree with you and I thnink that the advantages of large sensors are often overestimated. Many people compare at equal ISO, which does not make sense from a composition perspective.

Right.

Personally, I don't need "extreme" aperture and what M43 has to offer is more than enough to my needs. I can understand that people may have more specific needs and I don't blame them going to expensive and bulky FF, but I'm afraid that many could actually live with smaller sensors.

I don't blame them either. As I have said many times, if you want more shallow DoF than MFT can provide, it is a good idea to go FF.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
Just another Canon shooter
Just another Canon shooter Veteran Member • Posts: 4,691
A picture is worth...
1

Sergey_Green wrote:

Anders W wrote:

On the contrary, the dark tones cannot as a rule be "cleaned" without ill effects. Consider for example the hair of the girl in the example to which I link below.

http://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=oly_em1&attr13_1=sony_a7&attr13_2=sony_a7r&attr13_3=nikon_d800&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=3200&attr16_1=12800&attr16_2=12800&attr16_3=12800&normalization=print&widget=1&x=0.8342357734262059&y=0.2753099540100277

This comparison is not done with equivalent settings. If you equate the total light (assuming the same scene lighting), things go back to their places. The OM-D is slightly worse, detail vs. noise, like here, here and here, and maybe has some deep shadow advantage here to the A7 only but the A7 is more detailed. The 6D beats it every single time like here (can you read the text?), even in the deep shadows. In the greens (what DXO measures), it seems to be doing OK. This is in artificial light, and of course reflects the color filters, etc.

So the real world results seem to show that the OM-D shows more noise vs. less detail, even in the deep shadows, with normal processing, at least (and no sane person will try to pull the shadows in those cases anyway).

Now, show me your calculations to prove me wrong.

 Just another Canon shooter's gear list:Just another Canon shooter's gear list
Canon EOS 5D Mark II Canon EF 15mm f/2.8 Fisheye Canon EF 35mm F1.4L USM Canon EF 50mm f/1.2L USM Canon EF 135mm F2L USM +4 more
Just another Canon shooter
Just another Canon shooter Veteran Member • Posts: 4,691
Re: Nice summary but the supplement needs improvement
1

Anders W wrote:

Personally though, as I said, I am happy to make do with lenses in the range from f/1.7 to f/2.0. That's as much equivalence as I need.

What is your answer to pedestrian lenses like the 16/17-40, 24-xx, 70-200 on FF, all f/4? Not to mention f/2.8 zooms? Do not tell me you own one of those f/2 zoom monsters.

 Just another Canon shooter's gear list:Just another Canon shooter's gear list
Canon EOS 5D Mark II Canon EF 15mm f/2.8 Fisheye Canon EF 35mm F1.4L USM Canon EF 50mm f/1.2L USM Canon EF 135mm F2L USM +4 more
Anders W Forum Pro • Posts: 21,468
Re: A picture is worth...
3

Just another Canon shooter wrote:

Sergey_Green wrote:

Anders W wrote:

On the contrary, the dark tones cannot as a rule be "cleaned" without ill effects. Consider for example the hair of the girl in the example to which I link below.

http://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=oly_em1&attr13_1=sony_a7&attr13_2=sony_a7r&attr13_3=nikon_d800&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=3200&attr16_1=12800&attr16_2=12800&attr16_3=12800&normalization=print&widget=1&x=0.8342357734262059&y=0.2753099540100277

This comparison is not done with equivalent settings.

Yes it is. I compared ISO 3200 on MFT with ISO 12800 on FF, i.e., a two EV difference. This is what you need for the FF camera to reach the same DoF.

If you equate the total light (assuming the same scene lighting), things go back to their places.

I already equated total light. You didn't. You incorrectly compare ISO 3200 on MFT with ISO 6400 on FF.

The OM-D is slightly worse, detail vs. noise, like here, here and here, and maybe has some deep shadow advantage here to the A7 only but the A7 is more detailed. The 6D beats it every single time like here (can you read the text?), even in the deep shadows. In the greens (what DXO measures), it seems to be doing OK. This is in artificial light, and of course reflects the color filters, etc.

Even in the incorrect comparisons you provide, at a one- rather than two-stop ISO difference, the E-M1 beats the A7, A7r and D800 thanks to the greater efficiency of the smaller sensor. I didn't say that the efficiency advantage of the smaller sensor is more than one stop in comparison with any FF camera although for the above three it is.

So the real world results seem to show that the OM-D shows more noise vs. less detail, even in the deep shadows, with normal processing, at least (and no sane person will try to pull the shadows in those cases anyway).

See above.

Now, show me your calculations to prove me wrong.

I have already showed me your calculations and they, as well as what I said above, prove you wrong.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
Just another Canon shooter
Just another Canon shooter Veteran Member • Posts: 4,691
Re: A picture is worth...
1

Anders W wrote:

Just another Canon shooter wrote:

Sergey_Green wrote:

Anders W wrote:

On the contrary, the dark tones cannot as a rule be "cleaned" without ill effects. Consider for example the hair of the girl in the example to which I link below.

http://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=oly_em1&attr13_1=sony_a7&attr13_2=sony_a7r&attr13_3=nikon_d800&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=3200&attr16_1=12800&attr16_2=12800&attr16_3=12800&normalization=print&widget=1&x=0.8342357734262059&y=0.2753099540100277

This comparison is not done with equivalent settings.

Yes it is. I compared ISO 3200 on MFT with ISO 12800 on FF, i.e., a two EV difference.

Wrong. ISO has nothing to do with EV.

 Just another Canon shooter's gear list:Just another Canon shooter's gear list
Canon EOS 5D Mark II Canon EF 15mm f/2.8 Fisheye Canon EF 35mm F1.4L USM Canon EF 50mm f/1.2L USM Canon EF 135mm F2L USM +4 more
texinwien Veteran Member • Posts: 3,326
Not what he says

Sergey_Green wrote:

texinwien wrote:

What does "the lens is stressed less" mean? Is that accepted optics terminology? Do you have some links to authoritative sources that talk about this in more detail?

I think he means that to record the same amount of detail with the same number of pixels the lens that is in front of the smaller pixels must be resolving better, than the lens that is in front of the larger pixels.

I think I'll wait to hear from Truthiness. He was talking about sensor size and enlargement factor. You're talking about pixel size and pixel count. These factors are not synonymous, and they shouldn't be confused as such.

You've obviously confused them here, but there's no reason for me to believe that Truthiness has made the same elementary mistake in reasoning of which you're guilty.

That is the smaller pixels on a smaller area with "stress" the lens more, than the larger pixels on the larger area.

Do you think that the sizes of the image circles that the respective lenses and sensors are designed for play any role in this "stress" equation? If so, how would you describe the role they play?

thus the lens can be of lesser quality to achieve the same image quality.

Lesser quality in what respect?

As stated above.

Nope.

-- hide signature --

- sergey

 texinwien's gear list:texinwien's gear list
Panasonic Lumix DMC-GM5 Olympus E-M5 II Olympus 12-40mm F2.8 Pro OnePlus One Canon EOS 300D +20 more
Anders W Forum Pro • Posts: 21,468
Re: Nice summary but the supplement needs improvement

Just another Canon shooter wrote:

Anders W wrote:

Personally though, as I said, I am happy to make do with lenses in the range from f/1.7 to f/2.0. That's as much equivalence as I need.

What is your answer to pedestrian lenses like the 16/17-40, 24-xx, 70-200 on FF, all f/4? Not to mention f/2.8 zooms? Do not tell me you own one of those f/2 zoom monsters.

No I don't. If I want more subject isolation by means of background blur than my MFT zooms can provide, I carry a prime or two for that particular purpose (usually 45/1.8 and/or 75/1.8). I prefer more "FL control" (more lenses in my bag rather than in a drawer back home) to more "DoF control".

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
Truthiness Regular Member • Posts: 122
Re: Handevision Ibelux ?
3

texinwien wrote:

Truthiness wrote:

Sergey_Green wrote:

tokumeino wrote:

I was actually thinking of lenses like the Pana 25/1.4 which is 1 stop behind a Sony FE 55/1.8.

Yes, it is 50/2.8 equivalent. An Ok to have, but nothing to give you an edge when it comes to DoF and light. And stopped down to f/2.8 fast FF lenses do usually produce better results, than a wide open on a smaller formats equivalents.

Additionally there is one inherit advantage larger formats have over smaller ones: enlargement factor. A large format image needs to be enlarged less than a medium format image, which needs to be enlarged less than FF image which needs to be enlarged less than APS-C image which needs to be enlarged less than m4/3 image which needs to be enlarged less than cell phone image

This inherent factor is taken into account by switching from "Screen" to "Print" when viewing DXO's measurement graphs.

I wasn't talking about SNR of the signal of such, but about the image as rendered by the lens, before anything sensor does.

and when less enlargement is needed, the lens is stressed less,

What does "the lens is stressed less" mean? Is that accepted optics terminology? Do you have some links to authoritative sources that talk about this in more detail?

I have something better than that - logic:

If you want to produce a 20 cm by 30 cm picture (600 cm^2) from a full frame system where the lens draws you an image of 36 mm by 24 mm, you need to enlarge the image the lens draws to 6940% of the original size, 69 times larger the size the lens draws).

If you system has for example 17.3 mm by 13 mm as in m43 and you want to print the same size (600cm^2), you need to enlarge the image by 26700%, 267 times larger  the size the lens draws.

The more you need to enlarge the image, the better quality the image lens draws needs to be for one to get the result one wishes.

If you use the very same lens on both these systems, the aberrations of the lens will be more visible due to the larger enlargement on the system with the smaller image sensor.

thus the lens can be of lesser quality to achieve the same image quality.

Lesser quality in what respect?

It can have stronger aberrations as they (among with everthing else) are enlarged less.

Because of this the very best lenses are in the smallest formats and using lenses meant for larger formats often produce sub-optimal outcome.

Regardless, there is a limit how good a lens can be, thus a larger format in principle has a higher ceiling for image quality in this regard as well.

Anders W Forum Pro • Posts: 21,468
Re: A picture is worth...
2

Just another Canon shooter wrote:

Anders W wrote:

Just another Canon shooter wrote:

Sergey_Green wrote:

Anders W wrote:

On the contrary, the dark tones cannot as a rule be "cleaned" without ill effects. Consider for example the hair of the girl in the example to which I link below.

http://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=oly_em1&attr13_1=sony_a7&attr13_2=sony_a7r&attr13_3=nikon_d800&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=3200&attr16_1=12800&attr16_2=12800&attr16_3=12800&normalization=print&widget=1&x=0.8342357734262059&y=0.2753099540100277

This comparison is not done with equivalent settings.

Yes it is. I compared ISO 3200 on MFT with ISO 12800 on FF, i.e., a two EV difference.

Wrong. ISO has nothing to do with EV.

Yes it does. Let me explain:

Suppose the MFT user can take a certain photo at 25 mm, f/2.8, 1/100 s, and ISO 3200. To capture the same photo, the FF user needs 50 mm, f/5.6, 1/100 s, and ISO 12800. Do you get it?

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
eyeswideshut Regular Member • Posts: 333
Re: Handevision Ibelux ?

I get the sense you are inferring from enlargements of film using an enlarger lens?

Truthiness wrote:

texinwien wrote:

Truthiness wrote:

Sergey_Green wrote:

tokumeino wrote:

I was actually thinking of lenses like the Pana 25/1.4 which is 1 stop behind a Sony FE 55/1.8.

Yes, it is 50/2.8 equivalent. An Ok to have, but nothing to give you an edge when it comes to DoF and light. And stopped down to f/2.8 fast FF lenses do usually produce better results, than a wide open on a smaller formats equivalents.

Additionally there is one inherit advantage larger formats have over smaller ones: enlargement factor. A large format image needs to be enlarged less than a medium format image, which needs to be enlarged less than FF image which needs to be enlarged less than APS-C image which needs to be enlarged less than m4/3 image which needs to be enlarged less than cell phone image

This inherent factor is taken into account by switching from "Screen" to "Print" when viewing DXO's measurement graphs.

I wasn't talking about SNR of the signal of such, but about the image as rendered by the lens, before anything sensor does.

and when less enlargement is needed, the lens is stressed less,

What does "the lens is stressed less" mean? Is that accepted optics terminology? Do you have some links to authoritative sources that talk about this in more detail?

I have something better than that - logic:

If you want to produce a 20 cm by 30 cm picture (600 cm^2) from a full frame system where the lens draws you an image of 36 mm by 24 mm, you need to enlarge the image the lens draws to 6940% of the original size, 69 times larger the size the lens draws).

If you system has for example 17.3 mm by 13 mm as in m43 and you want to print the same size (600cm^2), you need to enlarge the image by 26700%, 267 times larger the size the lens draws.

The more you need to enlarge the image, the better quality the image lens draws needs to be for one to get the result one wishes.

If you use the very same lens on both these systems, the aberrations of the lens will be more visible due to the larger enlargement on the system with the smaller image sensor.

thus the lens can be of lesser quality to achieve the same image quality.

Lesser quality in what respect?

It can have stronger aberrations as they (among with everthing else) are enlarged less.

Because of this the very best lenses are in the smallest formats and using lenses meant for larger formats often produce sub-optimal outcome.

Regardless, there is a limit how good a lens can be, thus a larger format in principle has a higher ceiling for image quality in this regard as well.

-- hide signature --

Cheers
eyeswideshut

Anders W Forum Pro • Posts: 21,468
Re: Very simple:

tokumeino wrote:

It become less and less "Very simple" Hard for me to follow.

Intuitively, I'm considering that, if you can produce sensors with a particular efficiency (/mm2) for M43, then you just have to make it bigger and you'll have the same efficiency for FF. Aren't FF manufacturers aware of the technology used for smaller sensors ? Of course, this is only intuition and I don't know of quantums, and can obviously go wrong.

This is in fact a interresting debate and if you could avoid to answer to each other point by point, it could be more readable and interresting. As it, I'm afraid it come to an argue, instead.

We don't know the precise mechanisms behind the regularities shown by my analyses. But one likely reason why smaller sensors tend to be more efficient than larger is that new technology gets used for smaller sensors before they migrate to larger. Smaller sensors are produced in greater number and have shorter production cycles. Yield problems with new technologies are less pressing for smaller sensors. New production facilities with finer geometries are first used for smaller sensors. And so on.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
eyeswideshut Regular Member • Posts: 333
Re: A picture is worth...

Anders W wrote:

Just another Canon shooter wrote:

Anders W wrote:

Just another Canon shooter wrote:

Sergey_Green wrote:

Anders W wrote:

On the contrary, the dark tones cannot as a rule be "cleaned" without ill effects. Consider for example the hair of the girl in the example to which I link below.

http://www.dpreview.com/reviews/image-comparison?attr18=lowlight&attr13_0=oly_em1&attr13_1=sony_a7&attr13_2=sony_a7r&attr13_3=nikon_d800&attr15_0=raw&attr15_1=raw&attr15_2=raw&attr15_3=raw&attr16_0=3200&attr16_1=12800&attr16_2=12800&attr16_3=12800&normalization=print&widget=1&x=0.8342357734262059&y=0.2753099540100277

This comparison is not done with equivalent settings.

Yes it is. I compared ISO 3200 on MFT with ISO 12800 on FF, i.e., a two EV difference.

Wrong. ISO has nothing to do with EV.

Yes it does. Let me explain:

Suppose the MFT user can take a certain photo at 25 mm, f/2.8, 1/100 s, and ISO 3200. To capture the same photo, the FF user needs 50 mm, f/5.6, 1/100 s, and ISO 12800. Do you get it?

Have you tried varying shutter speed and leaving iso untouched?

I think this equivalence back and forth is beginning to cloud all our minds

-- hide signature --

Cheers
eyeswideshut

Sergey_Green
Sergey_Green Forum Pro • Posts: 11,996
I am confuced again ;) ?

texinwien wrote:

You've obviously confused them here, but there's no reason for me to believe that Truthiness has made the same elementary mistake in reasoning of which you're guilty.

Which reasoning was that?

That is the smaller pixels on a smaller area with "stress" the lens more, than the larger pixels on the larger area.

Do you think that the sizes of the image circles that the respective lenses and sensors are designed for play any role in this "stress" equation? If so, how would you describe the role they play?

The size of the pixels (not the image circle). Smaller the pixels the better the lens must be. I can see it on D800, some lenses are considerably better on it (when compared against 12MP cameras), whereas others are only marginally better (as few old wide open fast primes for example). That's what I thought he meant in his post.

thus the lens can be of lesser quality to achieve the same image quality.

Lesser quality in what respect?

As stated above.

Nope.

Generally yes.

-- hide signature --

- sergey

bobn2
bobn2 Forum Pro • Posts: 62,457
Re: Very simple:
1

Anders W wrote:

bobn2 wrote:

Anders W wrote:

Just another Canon shooter wrote:

Anders W wrote:

Just another Canon shooter wrote:

Anders W wrote:

FF has the same noise for equivalent images only if the sensors are equally efficient. But as a rule they are not. Smaller sensors tend to be more efficient than larger. See here:

http://www.dpreview.com/forums/post/53189207

Not true. QE of newer sensors is about 53%, see sensorgen, from m43 to APS-C to FF.

Sensorgen doesn't have data for many of the sensors in my sample and are based on fitting curves to numbers that results from previous curve fitting.

It does, for the most relevant bodies.

Why should we use partial data if we can use more complete data? And why should we use fitted data points if we can use real data points?

Anders, if you have good data for a full range of cameras, why not publish it so we can all use it?

I don't have a website for things like these and don't intend to start one. Furthermore, my data are collected for the purpose of a set of specific analyses (those presented in the post I linked to) and are sampled and designed specifically to that end. I don't think they are of much interest for any other purpose.

But if you or anyone else want a copy of my Excel spreadsheet for the purpose of replicating or verifying what I have done, I am of course willing to share it. The raw data from which I start out are already available via DxO and, in line with scientific practice, I have already made it clear exactly which data I use and what I do with them. Hence, everyone has had the opportunity to replicate my analysis from the very beginning.

That would be interesting. I've read some of your posts on this, but am still far from understanding how your analysis works, just that it produces results that you like. So, a look at the spreadsheet would be revealing.

Then the sensorgen data with its acknowledged faults would be redundant.

I don't consider the sensorgen data redundant. I just don't consider them as well suited to answering the questions about sensor efficiency that I asked as those I was in a position to assemble myself.

I think the issue is that you cite your own data as the justification for the comments, but it's not clear how the data is derived. If it provides useful information, then I think a wider audience might like to see it.

As it is, your 'real data points' seem to be derived from exactly the same data and your method involves 'interpolation' - so the meaning or the word 'real' here is not maybe what most of us would think of as 'real'.

I have made it perfectly clear in which ways my data differ from those of Sensorgen.

Perhaps, but I for one have found it hard to follow.

-- hide signature --

Bob

Anders W Forum Pro • Posts: 21,468
Re: Very simple:

Just another Canon shooter wrote:

Anders W wrote:

So you are using the DXO data to compute some "efficiency factor" that you define yourself to draw conclusions about noise, which was your starting point to begin with?

As I already pointed out, I use DxO data to compute the efficiency of the sensors to be compared. I do so on the basis of four different measures of signal-noise performance. What's wrong with that?

Do you understand how ridiculous is that?

On what specific ground would that be ridiculous?

And do you understand how ridiculous you make yourself by asking that question.

Sensorgen do very interesting analysis because they compute factors which are not in the DXO data, like well capacity and read noise in e-. Those two parameters have a clear meaning unlike your "efficiency" one.

As I already pointed out, the DxO figures on SNR on DR that I use are more direct measures of signal-noise performance, and thus image quality, than things like well capacity and read noise. That's part of the reason why I use the former.

You misinterpreted my remark about the resolution of the system. Given the same noise curve, a system with better resolution has some advantage. You can apply some NR and have a lower noise and the same resolution as a lower res system.

No I didn't misinterpret anything. The per-pixel signal-noise performance of two equally efficient sensors will not be the same if one has more pixels than the other. The performance of the sensor with more pixels will be worse on a per-pixel basis.

This disadvantage can be compensated by applying more NR (e.g., by downsampling) to the images from the sensor with more pixels. As I already pointed out, this is already done in my comparisons of sensor efficiency. The figures I use are for DxO "print mode", i.e., a resolution of 8 MP for all sensors alike, no matter how many pixels they originally have.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
texinwien Veteran Member • Posts: 3,326
Re: Handevision Ibelux ?
1

Truthiness wrote:

texinwien wrote:

Truthiness wrote:

Sergey_Green wrote:

tokumeino wrote:

I was actually thinking of lenses like the Pana 25/1.4 which is 1 stop behind a Sony FE 55/1.8.

Yes, it is 50/2.8 equivalent. An Ok to have, but nothing to give you an edge when it comes to DoF and light. And stopped down to f/2.8 fast FF lenses do usually produce better results, than a wide open on a smaller formats equivalents.

Additionally there is one inherit advantage larger formats have over smaller ones: enlargement factor. A large format image needs to be enlarged less than a medium format image, which needs to be enlarged less than FF image which needs to be enlarged less than APS-C image which needs to be enlarged less than m4/3 image which needs to be enlarged less than cell phone image

This inherent factor is taken into account by switching from "Screen" to "Print" when viewing DXO's measurement graphs.

I wasn't talking about SNR of the signal of such, but about the image as rendered by the lens, before anything sensor does.

and when less enlargement is needed, the lens is stressed less,

What does "the lens is stressed less" mean? Is that accepted optics terminology? Do you have some links to authoritative sources that talk about this in more detail?

I have something better than that - logic:

If you want to produce a 20 cm by 30 cm picture (600 cm^2) from a full frame system where the lens draws you an image of 36 mm by 24 mm, you need to enlarge the image the lens draws to 6940% of the original size, 69 times larger the size the lens draws).

If you system has for example 17.3 mm by 13 mm as in m43 and you want to print the same size (600cm^2), you need to enlarge the image by 26700%, 267 times larger the size the lens draws.

The more you need to enlarge the image, the better quality the image lens draws needs to be for one to get the result one wishes.

If you use the very same lens on both these systems, the aberrations of the lens will be more visible due to the larger enlargement on the system with the smaller image sensor.

Here's a logical thought experiment. Perhaps you could tell me how it fits in with the logic you've provided above:

1. We attach an FF lens to an ideal 16MP FF camera and take a photo of a static scene.

2. We attach the same FF lens to an ideal 16MP m43 camera via an optically perfect speedbooster and take a photo of the same static scene.

3. We enlarge both photos to the same size - what will the difference between the final, enlarged images be?

Is the enlargement factor the real issue here?

thus the lens can be of lesser quality to achieve the same image quality.

Lesser quality in what respect?

It can have stronger aberrations as they (among with everthing else) are enlarged less.

Because of this the very best lenses are in the smallest formats and using lenses meant for larger formats often produce sub-optimal outcome.

How much of the sub-optimal outcome is due to the fact that, in most cases, when using a lens meant for a larger format on a smaller format camera, a significant part of the lens' resolution is thrown away by the smaller format sensor?

Regardless, there is a limit how good a lens can be, thus a larger format in principle has a higher ceiling for image quality in this regard as well.

Again, I'd be interested to hear your take on the role of image circle size in this equation.

 texinwien's gear list:texinwien's gear list
Panasonic Lumix DMC-GM5 Olympus E-M5 II Olympus 12-40mm F2.8 Pro OnePlus One Canon EOS 300D +20 more
bobn2
bobn2 Forum Pro • Posts: 62,457
Re: Very simple:

Anders W wrote:

tokumeino wrote:

It become less and less "Very simple" Hard for me to follow.

Intuitively, I'm considering that, if you can produce sensors with a particular efficiency (/mm2) for M43, then you just have to make it bigger and you'll have the same efficiency for FF. Aren't FF manufacturers aware of the technology used for smaller sensors ? Of course, this is only intuition and I don't know of quantums, and can obviously go wrong.

This is in fact a interresting debate and if you could avoid to answer to each other point by point, it could be more readable and interresting. As it, I'm afraid it come to an argue, instead.

We don't know the precise mechanisms behind the regularities shown by my analyses. But one likely reason why smaller sensors tend to be more efficient than larger is that new technology gets used for smaller sensors before they migrate to larger. Smaller sensors are produced in greater number and have shorter production cycles. Yield problems with new technologies are less pressing for smaller sensors. New production facilities with finer geometries are first used for smaller sensors. And so on.

I think the efficiency increasing techniques used on small sensors are particularly BSI and light pipes. Bothe wet designed to recover the lost efficiency due to lower fill factors, and both have over-compensated. The most surprising thing I find about your analysis is the linear correlation between sensor size and efficiency, I would have expected more of a step-wise function.

-- hide signature --

Bob

Sergey_Green
Sergey_Green Forum Pro • Posts: 11,996
Wow!

Anders W wrote:

Yes it does. Let me explain:

Suppose the MFT user can take a certain photo at 25 mm, f/2.8, 1/100 s, and ISO 3200. To capture the same photo, the FF user needs 50 mm, f/5.6, 1/100 s, and ISO 12800. Do you get it?

That would be a very, very dark place. Are you sure it is not a security camera you are talking about  ? Generally I take my camera out when I see good and interesting light, not when it is pitch dark and I can not find the dials. You do it differently?

-- hide signature --

- sergey

texinwien Veteran Member • Posts: 3,326
You're confused, not 'confuced' (sic)

Sergey_Green wrote:

texinwien wrote:

You've obviously confused them here, but there's no reason for me to believe that Truthiness has made the same elementary mistake in reasoning of which you're guilty.

Which reasoning was that?

The reasoning that led you to equate sensor size to pixel size and pixel count. I think it's obvious for all to see, and, as Truthiness has confirmed, your interpretation of his remarks was wrong.

That is the smaller pixels on a smaller area with "stress" the lens more, than the larger pixels on the larger area.

Do you think that the sizes of the image circles that the respective lenses and sensors are designed for play any role in this "stress" equation? If so, how would you describe the role they play?

The size of the pixels (not the image circle).

Let me get this right - you're saying the size of the image circle plays no role in this equation?

Smaller the pixels the better the lens must be. I can see it on D800, some lenses are considerably better on it (when compared against 12MP cameras), whereas others are only marginally better (as few old wide open fast primes for example). That's what I thought he meant in his post.

He's confirmed what he meant now, and it's not what you thought.

thus the lens can be of lesser quality to achieve the same image quality.

Lesser quality in what respect?

As stated above.

Nope.

Generally yes.

"Nope" refers to your misinterpretation of Thruthiness' words.

-- hide signature --

- sergey

 texinwien's gear list:texinwien's gear list
Panasonic Lumix DMC-GM5 Olympus E-M5 II Olympus 12-40mm F2.8 Pro OnePlus One Canon EOS 300D +20 more
Truthiness Regular Member • Posts: 122
Re: Handevision Ibelux ?
1

eyeswideshut wrote:

I get the sense you are inferring from enlargements of film using an enlarger lens?

No.

I'll try to simplify this:

  • The lens of the camera draws an image (43mm diagonal on full frame, of which 36mm by 24mm is used by the sensor of film)
  • This image is enlarged to the output size, maybe A4 or whatever by irrelevant means
  • The more you need to enlarge this image, the more the imperfections of the camera lens will show.
  • Because the smaller formats require larger enlargement, the lens used byt the camera needs to have better optical quality than if a larger format is used. (And this is often the case.) Also it means that adapter FF vintage lenses don't necessarily perform as well on m43 than they do on FF.
  • It doesn't matter if the device recording the image the lens draws is an image sensor or film
Keyboard shortcuts:
FForum MMy threads