Pages: 1 ... 4 5 [6]   Go Down

Author Topic: FF versus MF  (Read 24174 times)

Michael Erlewine

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1027
    • MacroStop.com
Re: FF versus MF
« Reply #100 on: April 29, 2017, 03:23:04 pm »

Not to be contrary, but for the sake of diversity, as mentioned before I don’t print images and have not a single photo on my walls, not one. Still, I want and require large Mpx, if only to appreciate the kind and level of detail, the micro-contrast, etc. of the photos I take. My point is that, obviously, it is not all about larger sensors and printing. There is such a thing as “Seeing” our photos in the kind of detail that satisfies something, and that is somehow liberating.
Logged
MichaelErlewine.smugmug.com. Founder MacroStop.com, MichaelErlewine.com, YouTube.com/user/merlewine

Brad P

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 258
Re: FF versus MF
« Reply #101 on: April 29, 2017, 03:34:58 pm »

Not to be contrary, but for the sake of diversity, as mentioned before I don’t print images and have not a single photo on my walls, not one. Still, I want and require large Mpx, if only to appreciate the kind and level of detail, the micro-contrast, etc. of the photos I take. My point is that, obviously, it is not all about larger sensors and printing. There is such a thing as “Seeing” our photos in the kind of detail that satisfies something, and that is somehow liberating.

An inarguable observation, even if no one else sees it (pun intended).
Logged

landscapephoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 623
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #102 on: April 29, 2017, 05:31:15 pm »

There is no theoretical limit on the resolution of lenses, except diffraction. Diffraction, however, should not be under-estimated, as we also need some depth of field unless shooting perfectly flat subjects. In normal photographic practice, the compromises between diffraction and depth of field are real, as anybody who tried to use a 8"x10" view camera would find out. This compromise also becomes apparent when stitching is used to increase resolution: only landscapes at far distance are routinely imaged.

There is no theoretical limit, but there are practical ones. Optics have made progresses, but not as fast as electronics. In practice, lenses are compromises on aberrations, price and size/weight. The optical engineer can build very good lenses for high-definition MF sensors, but they will be huge and heavy. When they are not, other tricks are used like software corrections of distortion or chromatic aberrations to relax the constraints a bit.

So what gives? At present, the maximum resolution available in a single lens frame is 200 mpix (by moving the sensor around). Presently available MF lenses cope, when stopped down a bit. An educated guess is that this would also be around the practical limit, give or take some.

Another important fact is that seemingly large increases in number of pixels correspond to relatively moderate increase in practical detail level. A good rule of thumb is that, to get a noticeable effect, one needs to double the resolution. This is very noticeable for 24x36 cameras where the manufacturers present resolution increase between, say, 36 and 42 mpix as significant while they are not unless peeping pixels. Today, the maximum resolution for 24x36 cameras is 50 mpix. An increase to, say, 70 mpix is not likely to break the lenses.

Hi,

Increasing sensor resolution will improve any lens. The MTF of the system is MTF of the lens times the MTF of the sensor. So as long as the MTF of the lens is above zero it will gain image quality when paired with a better sensor.

Best regards
Erik

Hi,

Why do you cite my message to post something completely unrelated to it? I included my message in the above quote for your convenience.

Best regards
Landscapephoto
Logged

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #103 on: April 29, 2017, 06:52:20 pm »

At present, the maximum resolution available in a single lens frame is 200 mpix (by moving the sensor around).

I'm not sure what you mean by this. Are you talking about the Hassy piezo-motion camera?

Jim

Brad P

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 258
Re: FF versus MF
« Reply #104 on: April 30, 2017, 12:53:35 am »

Unsure, but as I read through it, I thought the poster refered to a recent prosumer camera maker this last year that vibrates the sensor in favor of bigger resolution in raw files.  Interesting idea that doesn't seem yet to be catching on.  The Hassy piezo thing probably more on point, which I don't get yet.  I, for one, appreciate Erik's thoughts and do think his short reply was quite responsive and thought provoking in the equally or more interesting subjects landscapephoto touched on.

Lenses obviously are important to answers in the questions posted here.  In that light, I see an older article here, https://luminous-landscape.com/do-sensors-out-resolve-lenses.  Doesn't seem immediately like Michael's voice, but maybe.  Regardless, no doubt he was the publisher and many people's mentor, at least mine of sorts.

That doesn't resolve, but hopefully feeds in.
« Last Edit: April 30, 2017, 02:06:20 am by Brad P »
Logged

landscapephoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 623
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #105 on: April 30, 2017, 03:17:01 am »

I'm not sure what you mean by this. Are you talking about the Hassy piezo-motion camera?

Yes. These cameras will sample a single frame as projected by the lens with the equivalent of 200 mpix. They allow us to know how the lens would behave on a 200 mpix sensor.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #106 on: April 30, 2017, 04:29:19 am »

Hi,

The point I try to make is that most lenses will perform better if paired with a higher resolution sensor.

So, if you combine a decent quality lens with say a 50 MP sensor it will transfer better detail than it would do say with 21 MP sensor. Just as an example, Canon used to have a very good 24-70/2.8LII lens, but paired with the Canon 5DIII it could not match the weaker Nikon 24-70/2.8 lens on the D800E: https://www.lensrentals.com/blog/2013/01/a-24-70mm-system-comparison/

Another way to see it, any lens that can produce moiré on a sensor would be better served with a higher resolution sensor.

Tim Parkin has looked into this a bit back in 2012 and he found that the Nikon D800E (36MP)  delivered better detail at f/22 than the 24MP Sony Alpha 900 at f/5.6. That article is here: https://www.onlandscape.co.uk/2012/08/the-diffraction-limit-how-small-is-too-small/

My take is that pixels can shrink quiet a bit, but I don't think we will see a very rapid increase in megapixels. There is probably some ideal size of pixels, taking all parameters into account. You want to keep wiring area to photodiode area low, and I guess that there is also some concerns about crosstalk.

Lenses can probably improve a lot. There is an unfortunate trend that many high quality lenses are large aperture designs that drive both cost and weights upwards. There are some exceptions, like the Zeiss Batis line that combines relatively moderate apertures with high performance and moderate size.

Best regards
Erik

Hi,

Increasing sensor resolution will improve any lens. The MTF of the system is MTF of the lens times the MTF of the sensor. So as long as the MTF of the lens is above zero it will gain image quality when paired with a better sensor.

Best regards
Erik


Hi,

Why do you cite my message to post something completely unrelated to it? I included my message in the above quote for your convenience.

Best regards
Landscapephoto
Logged
Erik Kaffehr
 

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #107 on: April 30, 2017, 10:09:15 am »

Yes. These cameras will sample a single frame as projected by the lens with the equivalent of 200 mpix. They allow us to know how the lens would behave on a 200 mpix sensor.

Well, kinda. In H5D-200C 4-shot mode, there is no finer sampling than a 50 MP Bayer array (but better color). In 6 shot mode, there's a bit of intermediate sampling. There is no 16-shot mode any more. In addition, the fill factors do mot change from the 51 MP version of the Sony, so the effective aperture is about four times the area of that of a real 200 MP sensor. Also, Hasselblad offers no H-mount lenses optimized for 33x44mm. Put that all together and we're not seeing in 200C files what a real 200 MP camera cold do.

The Betterlight Super 10K would come closer.

Jim

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #108 on: April 30, 2017, 11:01:27 am »

Hi Jim,

This calculation of yours indicates that 2.0 micron pitch would be beneficial with the Otus at medium apertures.


An additional advantage of small pixels is that with small enough pixels we can drop the OLP filter, that also increases MTF.

But, going to very small pixels reduces DR for a few reasons. So, there is balance between DR and pixel size.

So, using 2 micron sensels we would end up with 216 MP on 24x36. But, we could use the same pixel size on 44x33 mm, yielding 363 MP.  The Otus 55 and 85 seems not to be far from covering 44x33. It may be that available lenses would not be very sharp in the corners , but having excellent reproduction over a large sweet spot may be an attractive option.

On the other hand, I would say the stuff I have now (Sony A7rII and a bunch of decent zoom lenses) cover my needs pretty well. I see a bit more aliasing on the Sony A7rII than I would like. Smaller pixels could obviously help with that.


Best regards
Erik


Well, kinda. In H5D-200C 4-shot mode, there is no finer sampling than a 50 MP Bayer array (but better color). In 6 shot mode, there's a bit of intermediate sampling. There is no 16-shot mode any more. In addition, the fill factors do mot change from the 51 MP version of the Sony, so the effective aperture is about four times the area of that of a real 200 MP sensor. Also, Hasselblad offers no H-mount lenses optimized for 33x44mm. Put that all together and we're not seeing in 200C files what a real 200 MP camera cold do.

The Betterlight Super 10K would come closer.

Jim
Logged
Erik Kaffehr
 

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #109 on: April 30, 2017, 11:09:50 am »

Hi Jim,

This calculation of yours indicates that 2.0 micron pitch would be beneficial with the Otus at medium apertures.


Yes, although MTF10 is rather a specialized metric. I have since discovered that my earlier testing (that formed the basic ofr the model used for that graph) shortchanged the Otus at f/2 and f/2.8, though.

Jim

landscapephoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 623
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #110 on: April 30, 2017, 11:29:18 am »

The point I try to make is that most lenses will perform better if paired with a higher resolution sensor.

OK. Now I understand what you mean.

The idea that a lens will perform better if paired with a higher resolution sensor is indeed true, up to a point. The problem here is "up to a point", which means that your theory will not hold for vastly different resolutions or for very small apertures, but that discussion would lead us too far for a simple photography forum.

Anyway, what I was saying is different; namely that we are approaching practical limits of what can be done with a camera on standard subjects. Your argument, if I have correctly understood, is that for any given resolution and camera, we will increase final sharpness if we add a few pixels to resolution. Under some assumptions this is true, but the increase of final sharpness is less and less visible. It is like an asymptotic limit, if you want to see it that way.
Logged

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #111 on: April 30, 2017, 11:44:59 am »

OK. Now I understand what you mean.

The idea that a lens will perform better if paired with a higher resolution sensor is indeed true, up to a point. The problem here is "up to a point", which means that your theory will not hold for vastly different resolutions or for very small apertures, but that discussion would lead us too far for a simple photography forum.

Anyway, what I was saying is different; namely that we are approaching practical limits of what can be done with a camera on standard subjects. Your argument, if I have correctly understood, is that for any given resolution and camera, we will increase final sharpness if we add a few pixels to resolution. Under some assumptions this is true, but the increase of final sharpness is less and less visible. It is like an asymptotic limit, if you want to see it that way.

Yes. People working with orbiting cameras have developed a metric to help them balance lens and sensor resolution:

http://blog.kasson.com/the-last-word/whats-your-q/

However, they normally work with diffraction-limited lenses and monochromatic sensors. It is possible to generalize the metric for Bayer GFAs, with some guesswork:

http://blog.kasson.com/the-last-word/interpreting-q-in-the-real-world/

http://blog.kasson.com/the-last-word/example-q-calculations-for-bayer-cfas/

By the way, I'm not sure the relationship between pitch and resolution-adjusted EDR that Erik talked about is all that solid.

Jim

landscapephoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 623
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #112 on: April 30, 2017, 11:46:05 am »

Well, kinda. In H5D-200C 4-shot mode, there is no finer sampling than a 50 MP Bayer array (but better color). In 6 shot mode, there's a bit of intermediate sampling. There is no 16-shot mode any more. In addition, the fill factors do mot change from the 51 MP version of the Sony, so the effective aperture is about four times the area of that of a real 200 MP sensor. Also, Hasselblad offers no H-mount lenses optimized for 33x44mm. Put that all together and we're not seeing in 200C files what a real 200 MP camera cold do.

A few corrections.

First, there have been different cameras with 6-shots mode. The present one uses the Sony CMOS sized 33x44mm. Older ones used the Kodak CCD sized 39x50mm.
Second, in 6-shot mode, there is exactly the same bit of intermediate sampling as there is in the Bayer array (which does not sample all colors at all positions either).
Third, the aperture effect, which is indeed real, correspond to a relatively mild, non-destructive, low-pass filter. The effect can be removed in post.

I think that the 200 piezo backs indeed give us a good idea of the lens response on a native 200 mpix Bayer sensor.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #113 on: April 30, 2017, 11:51:29 am »

Hi,

It is often called diminishing returns.

Personally, I am often shooting landscape and landscapes are often shot at infinity or we can often utilise Scheimpflug (that is tilts). So, I feel that the region I shoot within now, around 40 MP, is quite beneficial. On my P45+, moiré is a frequent problem, with the Sony A7rII less so. I am mostly shooting zooms and I still see some moiré on the A7rII, but I wouldn't really call it a problem.

I don't think that decreasing pixel size on the A7rII would make for a big improvement with my print sizes as I normally don't print larger than 28"x39". The major benefit of smaller pixels in MFD would be the reduction in aliasing.

My take is really that increasing sensor resolution may make it possible to have smaller, lighter and more affordable gear. Just as an example, we have really good 4/3" and APS-C gear now and such gear is quite adequate for A2/C-size prints.

Best regards
Erik


OK. Now I understand what you mean.

The idea that a lens will perform better if paired with a higher resolution sensor is indeed true, up to a point. The problem here is "up to a point", which means that your theory will not hold for vastly different resolutions or for very small apertures, but that discussion would lead us too far for a simple photography forum.

Anyway, what I was saying is different; namely that we are approaching practical limits of what can be done with a camera on standard subjects. Your argument, if I have correctly understood, is that for any given resolution and camera, we will increase final sharpness if we add a few pixels to resolution. Under some assumptions this is true, but the increase of final sharpness is less and less visible. It is like an asymptotic limit, if you want to see it that way.
Logged
Erik Kaffehr
 

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #114 on: April 30, 2017, 11:52:14 am »

A few corrections.

First, there have been different cameras with 6-shots mode. The present one uses the Sony CMOS sized 33x44mm. Older ones used the Kodak CCD sized 39x50mm.
Second, in 6-shot mode, there is exactly the same bit of intermediate sampling as there is in the Bayer array (which does not sample all colors at all positions either).
Third, the aperture effect, which is indeed real, correspond to a relatively mild, non-destructive, low-pass filter. The effect can be removed in post.

I think that the 200 piezo backs indeed give us a good idea of the lens response on a native 200 mpix Bayer sensor.

OK. Let's agree to disagree on that one. In a few years, we'll find out.

Jim

landscapephoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 623
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #115 on: April 30, 2017, 12:09:01 pm »

Yes. People working with orbiting cameras have developed a metric to help them balance lens and sensor resolution:

http://blog.kasson.com/the-last-word/whats-your-q/

However, they normally work with diffraction-limited lenses and monochromatic sensors. It is possible to generalize the metric for Bayer GFAs, with some guesswork:

http://blog.kasson.com/the-last-word/interpreting-q-in-the-real-world/

http://blog.kasson.com/the-last-word/example-q-calculations-for-bayer-cfas/

It is a good post, and a good reminder that noticing the first effects of diffraction does not mean that one should strop increasing resolution. But it is not what I was talking about.

In your study, you make an untold assumption: that the subject is perfectly flat. You are only interested in what happen at the plane of sharpness. That implies a perfectly flat subject as in a test chart or a landscape far away (or an orbiting camera). It is useful, but there is a little more to photography than perfectly flat subjects.

If the subject is not flat, we as photographers need to balance sharpness and depth of field.

Interestingly, depth of field depends on resolution. To effectively use the extra resolution, we need bigger prints (and we also need to let the public look at them close). But then, on these huge prints, depth of field becomes a bigger problem as everything out of the plane of focus will appear more noticeably unsharp. If we want the tridimensional object that is photographed to appear reasonably sharp all over on  a high resolution large print, we will need a smaller aperture than on a low resolution small print and then the effects of diffraction becomes more noticeable. The two requirements collide.

The effect is well known on large format film cameras. On these, the sensor (a sheet of film) has effectively a lot more resolution than necessary and everything is limited by diffraction because of the small apertures normally used to have some depth of field. Still: in practice depth of field and diffraction are a problem and movements are necessary to nudge the plane of focus to where it is most needed.
Logged

landscapephoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 623
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #116 on: April 30, 2017, 12:10:46 pm »

OK. Let's agree to disagree on that one. In a few years, we'll find out.

Or we could compare yesteryears multishot backs with today's higher resolution single shot backs.
Logged

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #118 on: April 30, 2017, 12:54:44 pm »

Hi,

Yes of course! But, many times medium apertures just work fine. Selective focus is one of the abstractions of photography. I always carry a HCam Master TS, a pocket size device that gives me tilt capability with my lenses from 16 to 135 mm on the A7rII.

I would agree that need of resolution has it's limits. But, I would also suggest that resolution is nice to have and never makes damage to pictures.

Best regards
Erik

It is a good post, and a good reminder that noticing the first effects of diffraction does not mean that one should strop increasing resolution. But it is not what I was talking about.

In your study, you make an untold assumption: that the subject is perfectly flat. You are only interested in what happen at the plane of sharpness. That implies a perfectly flat subject as in a test chart or a landscape far away (or an orbiting camera). It is useful, but there is a little more to photography than perfectly flat subjects.

If the subject is not flat, we as photographers need to balance sharpness and depth of field.

Interestingly, depth of field depends on resolution. To effectively use the extra resolution, we need bigger prints (and we also need to let the public look at them close). But then, on these huge prints, depth of field becomes a bigger problem as everything out of the plane of focus will appear more noticeably unsharp. If we want the tridimensional object that is photographed to appear reasonably sharp all over on  a high resolution large print, we will need a smaller aperture than on a low resolution small print and then the effects of diffraction becomes more noticeable. The two requirements collide.

The effect is well known on large format film cameras. On these, the sensor (a sheet of film) has effectively a lot more resolution than necessary and everything is limited by diffraction because of the small apertures normally used to have some depth of field. Still: in practice depth of field and diffraction are a problem and movements are necessary to nudge the plane of focus to where it is most needed.
Logged
Erik Kaffehr
 

landscapephoto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 623
Re: FF versus MF: will lenses be the ultimate limit on sensor downsizing?
« Reply #119 on: May 01, 2017, 05:02:17 am »

http://blog.kasson.com/the-last-word/do-sharper-lenses-have-more-or-less-dof/

The article are interesting, and I would like to add some historical context. Around the beginning of the 20th century, opticians produced lenses to increase depth of field (which was a problem with slow emulsions in large format view cameras). One of the most famous makers were Pinkham & Smith, but there were others.

These lenses were all soft focus lenses, which made sense as limited depth of field were a problem mainly for portraits. They achieved both soft focus and increased depth of field by adding aberrations, for example spherical aberration. Pinkham & Smith hand retouched the lenses to give the same effect. Others used chromatic aberration to the same effect (on B&W film, of course).

The idea is that a lens with spherical aberration or a hand retouched lens will have different focus distances according to the path the rays take in the lens. So the image looks like a composite of several images taken with slightly varied focus. We could do the same on static subjects by varying the focus a bit while taking multiple exposures on the same piece of film. This makes highlights look as if they glow with light but also increases apparent depth of field.

In the 70s some add-on filters were developed to the same effect. Basically, they had little low-power lenses on their surfaces. Some, I think from Minolta, made the power of these extra lenses depend on wavelength so that the effect would only appear on some colors, while preserving apparent sharpness on others. This way skin detail could be preserved with softening and increased depth of field on different colors.

These effects are impossible to emulate in post, unless using a light field camera. That is: the softening can be emulated, but the increased depth of field cannot, unless combining pictures taken with varied focus (which a light field camera does by principle). But I am not aware of software doing that.
Logged
Pages: 1 ... 4 5 [6]   Go Up