Sony A7R, intersting comaprision ...

[AreBee]

Erik,

If you put a very good lens on a high resolution back with larger sensor size it will always be "better" than a smaller sensor with an equally good lens. So if you shoot flat targets, like a distant landscape, the larger format having more pixels will always win.

Not sure I understand your meaning. Is the larger sensor "better" only because it contains a greater number of pixels, or for some other reason?

Bart,

That's what you get when using lenses specifically designed for film on any digital sensor...

I am aware that lenses designed to be placed in front of film can be (always are?) problematic when placed in front of a digital sensor, but what about the reverse? Can lenses designed to be placed in front of a digital sensor be placed in front of film with impunity?

[Doug Peterson]

I am aware that lenses designed to be placed in front of film can be (always are?) problematic when placed in front of a digital sensor, but what about the reverse? Can lenses designed to be placed in front of a digital sensor be placed in front of film with impunity?

Most digital-designed lenses have too small an image circle to cover 4x5 film.

Smaller format film (e.g. 6x9 or 645) with a digital-designed lens with sufficient image circle will be fine.

List of image circles for digital-designed lenses.

[AreBee]

Thank you Doug.

[Phil Indeblanc]

Schneider has HR lenes as well. Are these "equal" to Rod HR?

Is Schneider HR for Digital or is this a mainly a film lens?

I also remember the NEWER sesores with smaller photons to benifit more from digital lenes.
Older sesonrs with 9 or larger microns are not going to gain from the Digital lenses. Correct or adjust me if I'm off a bit.

[Bart_van_der_Wolf]

I am aware that lenses designed to be placed in front of film can be (always are?) problematic when placed in front of a digital sensor, but what about the reverse? Can lenses designed to be placed in front of a digital sensor be placed in front of film with impunity?

Hi Rob,

Depends on the design. A lens designed for film only has to project the image on a flat plane (assuming flat film, some (pano) film used a cylindrical exposure area).

With a digital sensor, the image forming rays are preferably designed to strike the sensor array as perpendicular as possible. With a digital sensor, there are also several plan-parallel semi-transparent layers in the optical path, such as an IR-filter, perhaps a stack of Optical Low-Pass filters, and a sensor coverglass. When a ray of light hits these layers at an angle, there will be some refraction that will result in a small amount of shift (depends on refractive index of the material, modified by coating layers).

If a lens is specifically designed for the presence of those layers in the optical path, then that may lead to slightly different optical design for an optimal performance. If that were to be done to the extreme, then that lens would be better corrected for a digital sensor than for film. However, since there can be some variation in those layers from camera to camera, I think most current designs are optimized mostly with a larger exit pupil distance than they would have been when designed for film. If that is the only change, then that lens would be equally well suited for film.

Some traditional designs happened to already have large exit pupil distances, so they happen to also work quite well on digital sensors, but lenses specifically designed for digital sensors could possibly outperform them a bit.

There is one other specific design feature that 'digital' designs need to pay attention to, and that is the potential reflection from the sensor (potentially more reflective than film) back to the rear lens element. Therefore rear element shape and coating may get a bit more attention in a 'digital' design. Obviously, IR filter and sensor cover-glass AR-coatings also help to reduce the amount of relection. Some lenses used for IR-photography exhibit a central hotspot in the image, while others don't. That may be related to the shape and coating of that rear lens element.

Cheers,
Bart

[ErikKaffehr]

Hi,

Rodenstock and Schneider are different designs.

The Rodenstock HR lenses are very high resolution designs that cover MFD sensor format, smallish image circle extreme performance. The Rodenstock wide angles are also telecentric and have therefore less issues with color cast and vignetting.

Best regards
Erik

Schneider has HR lenes as well. Are these "equal" to Rod HR?

Is Schneider HR for Digital or is this a mainly a film lens?

I also remember the NEWER sesores with smaller photons to benifit more from digital lenes.
Older sesonrs with 9 or larger microns are not going to gain from the Digital lenses. Correct or adjust me if I'm off a bit.

[EricWHiss]

Bart,
Great explanation! Thanks!

Can this be summarized as follows?

Smaller image circle = less incident angle going out to edge = good
Larger sensor to exit pupil = less incident angle = good
Less incident angle means more problems with reflection off sensor, requiring more attention to lens shape and coatings?

I had thought that another issue with analog lenses on digital was that the color chemistry in film was at different depths, whereas they are the same depth on a sensor causing really well corrected lenses for analog to be off, resulting in axial chromatic aberrations with digital capture.

[ErikKaffehr]

Hi,

I don't think I would agree.

Image circle is not part of the equation. You can put a Hasselblad lens on a Sony sensor and it works just fine. What may matter is baffling. Some adapters like the Mirex T/S have internal buffling that reduces the risk of internal reflections in mount and adapter internals. This can be a big problem with badly designed adapters!

The size of the pupil, I don't think it matters, it is the distance between sensor and the pupil that matters most. A larger pupil is worse than a small pupil, you want the rays coming from lens to sensor as perpendicular to sensor as possible.

The reflections from the sensor may matter. Lenses for digital have better coating on element pointing to the sensor.

The major problem with film was curvature but also the stochastic nature of silver-halogenid capture. It takes many grains to describe a density, while one pixel does it. Film also has a problem with curvature. The sensors are flat.

Best regards
Erik

Bart,
Great explanation! Thanks!

Can this be summarized as follows?

Smaller image circle = less incident angle going out to edge = good
Larger sensor to exit pupil = less incident angle = good
Less incident angle means more problems with reflection off sensor, requiring more attention to lens shape and coatings?

I had thought that another issue with analog lenses on digital was that the color chemistry in film was at different depths, whereas they are the same depth on a sensor causing really well corrected lenses for analog to be off, resulting in axial chromatic aberrations with digital capture.

[Paul2660]

Actually the modern digital lenses from both Schneider and Rodenstock tend to have pretty large image circles.  

The Rodenstock HR (now called S) lenses which are the 23, 28, both have smallish 70mm image circles and thus are very limited to shifting.  

The Rodenstock HR-W lenses, 32mm, 40mm, 70mm (their newest besides the HR-SW 90mm) have 90mm image circles which allow for a greater amount of shifting.  up to 18mm on the 40mm on a 60mp back.  The 70mm has a 100mm IC, and the 90mm in the HR-W (no longer made) has a 125mm image circle

The Schneiders, in the APO Digitar series, 28mm, 35mm, both have a 90mm image circle however due lens design, you will not be able to shift the 35mm much past 8mm on a 60mp back as much past that you get too much magenta shift and detail smearing.   The Schneider 43mm has a IC of 110mm and can shift to around 12mm without any loss of detail or saturation and can be pushed to 16mm depending on the subject on a 60mp back.  

The Schneider Digitar 60mm has a 125mm image circle and is an amazing lens, which will allow shift of up to 25mm before you start to see any loss of detail or harsh magenta problems.  

You can read on the Rodenstock's here: which is the link to their latest lens brochure.  http://www.rodenstock-photo.com/mediabase/original/Objektive_digitale_Fotografie_e_10468.pdf

Schneider's website is very limited to lens info, but if you go to the main Cambo website, http://www.cambo.com  you can read all about the Schneiders and Rodenstocks.  Cambo's site is wealth of knowledge on lenses.

The Rodenstock's are all retrofocus designs, thus tend to be heavier and larger than the Schneiders.  The Schneider wides up to the 43mm all have issues on the 80MP backs when shifted (harsh magenta color shifts, possible center folding issues that won't resolve with LCC, and detail smearing)  The Rodenstocks will shift very well on both the 60MP and 80MP backs, however Rodenstock by design places a disc inside the lens (apparently to let the user know where the limit of the image circle is) and this creates a hard vignette that ruins the corners.  Nothing can fix this.  It's a tragedy to me since I know that the 28mm would easily shift to 10mm maybe 12mm without any magenta shifting or detail smearing, but the disc shows up at around anything past 7mm of horizontal shift.  You can get a bit more rise.  You will see much less magenta color shifting with the Rodenstocks than the Schneiders.  This is very easy to see in the LCC, both lenses will show considerable light fall off, but the LCC will fix this to an amazing degree.  The magenta shift is not as easy to fix.  

Paul Caldwell

[ErikKaffehr]

Thanks for good info!

Erik

Actually the modern digital lenses from both Schneider and Rodenstock tend to have pretty large image circles.  

The Rodenstock HR (now called S) lenses which are the 23, 28, both have smallish 70mm image circles and thus are very limited to shifting.  

The Rodenstock HR-W lenses, 32mm, 40mm, 70mm (their newest besides the HR-SW 90mm) have 90mm image circles which allow for a greater amount of shifting.  up to 18mm on the 40mm on a 60mp back.  The 70mm has a 100mm IC, and the 90mm in the HR-W (no longer made) has a 125mm image circle

The Schneiders, in the APO Digitar series, 28mm, 35mm, both have a 90mm image circle however due lens design, you will not be able to shift the 35mm much past 8mm on a 60mp back as much past that you get too much magenta shift and detail smearing.   The Schneider 43mm has a IC of 110mm and can shift to around 12mm without any loss of detail or saturation and can be pushed to 16mm depending on the subject on a 60mp back.  

The Schneider Digitar 60mm has a 125mm image circle and is an amazing lens, which will allow shift of up to 25mm before you start to see any loss of detail or harsh magenta problems.  

You can read on the Rodenstock's here: which is the link to their latest lens brochure.  http://www.rodenstock-photo.com/mediabase/original/Objektive_digitale_Fotografie_e_10468.pdf

Schneider's website is very limited to lens info, but if you go to the main Cambo website, http://www.cambo.com  you can read all about the Schneiders and Rodenstocks.  Cambo's site is wealth of knowledge on lenses.

The Rodenstock's are all retrofocus designs, thus tend to be heavier and larger than the Schneiders.  The Schneider wides up to the 43mm all have issues on the 80MP backs when shifted (harsh magenta color shifts, possible center folding issues that won't resolve with LCC, and detail smearing)  The Rodenstocks will shift very well on both the 60MP and 80MP backs, however Rodenstock by design places a disc inside the lens (apparently to let the user know where the limit of the image circle is) and this creates a hard vignette that ruins the corners.  Nothing can fix this.  It's a tragedy to me since I know that the 28mm would easily shift to 10mm maybe 12mm without any magenta shifting or detail smearing, but the disc shows up at around anything past 7mm of horizontal shift.  You can get a bit more rise.  You will see much less magenta color shifting with the Rodenstocks than the Schneiders.  This is very easy to see in the LCC, both lenses will show considerable light fall off, but the LCC will fix this to an amazing degree.  The magenta shift is not as easy to fix.  

Paul Caldwell

[AreBee]

Bart,
Great explanation! Thanks!

+1

Thank you Bart.

[bcooter]

Everytime a new high megapixel camera comes out we always compare it to a medium format back, which is a good yardstick.

What we don't get to see is how we work in the real world, which means nothing, or can mean everything.

If we're shooting with strobe and a lot of light medium format just shines.  Continuous light, if there is enough volume it's great, but at a certain point the smaller cameras catch up.

I'm very curious about the A7R though I've told myself I'd never go Sony again, but it does look interesting and I guess the only way to really know is to shoot it in the conditions I work.

I do think the the camera and lens makers play too strong to the pixel examiners on the web, with reports of ca or slight softness coming out like it's the end of the world, though some of the prettiest lenses I use aren't that sharp and always need some post work to correct some CA, but I shoot people so usually pure pixel sharpness isn't my only goal.

IMO

BC

[BernardLanguillier]

I do think the the camera and lens makers play too strong to the pixel examiners on the web, with reports of ca or slight softness coming out like it's the end of the world, though some of the prettiest lenses I use aren't that sharp and always need some post work to correct some CA, but I shoot people so usually pure pixel sharpness isn't my only goal.

At least Nikon cannot be accused of that.

I attended a Nikon Df/58mm f1.4 AF-S event on Saturday in Tokyo, and 98% of the marketing message around the 58mm f1.4 is about the look of the bokeh and its behavior as a function of the distance behind the plane of sharpness. The remaining 2% were about sharpness, but only as a contrasting aspect to bokeh.

I have to admit that the samples were extremely nice and supportive of the message.

Cheers,
Bernard

[ErikKaffehr]

Hi,

Yes, I think so. Good enough is good enough. I have upgraded time to time and now I have quite a few cameras from Hasselblad V/P45+ to Sony RX-100. The best pictures I have taken were probably with a Sony Alpha APS-C at 12 MP that is now inherited by my parents. I sort of feel that those 12 MP were good enough.

On the other hand, I really enjoy all my present cameras.

Best regards
Erik

Everytime a new high megapixel camera comes out we always compare it to a medium format back, which is a good yardstick.

What we don't get to see is how we work in the real world, which means nothing, or can mean everything.

If we're shooting with strobe and a lot of light medium format just shines.  Continuous light, if there is enough volume it's great, but at a certain point the smaller cameras catch up.

I'm very curious about the A7R though I've told myself I'd never go Sony again, but it does look interesting and I guess the only way to really know is to shoot it in the conditions I work.

I do think the the camera and lens makers play too strong to the pixel examiners on the web, with reports of ca or slight softness coming out like it's the end of the world, though some of the prettiest lenses I use aren't that sharp and always need some post work to correct some CA, but I shoot people so usually pure pixel sharpness isn't my only goal.

IMO

BC

[EricWHiss]

The best pictures I have taken were probably with a Sony Alpha APS-C at 12 MP that is now inherited by my parents. I sort of feel that those 12 MP were good enough.

Erik,
I couldn't resist myself and am going to tease you a little bit so don't take this too seriously…..  But maybe the reason why you took the best pictures with your 12mp camera was because you were interested in picture taking then and not spending your free time making lots of comparisons?  In other words,   can you analyze image quality and make good/interesting images at the same time?

[ErikKaffehr]

Hi Eric,

Yes I think so. I have been always interested in optics and science. One of the main reasons I made a lot of good images with the 12 MP 700 that I was visiting new places, like Yellowstone and Grand Teton. I went back to those places last year but I had some bad luck with smog because of wildfires. Still I got some new images.

The point may be that the Alpha 700 was good enough. The cameras I had before it were perhaps less than good enough.

I do testing, because I am interested in both photography and the theory behind. I take pictures because it gives me peace of mind. But, I consider myself to be an engineer, not an artist.

Some of my work is here: http://echophoto.smugmug.com

Best regards
Erik

Ps. Teasing is OK! And I enjoy your postings, just so you know!

Erik,
I couldn't resist myself and am going to tease you a little bit so don't take this too seriously…..  But maybe the reason why you took the best pictures with your 12mp camera was because you were interested in picture taking then and not spending your free time making lots of comparisons?  In other words,   can you analyze image quality and make good/interesting images at the same time?

[Doug Peterson]

At least Nikon cannot be accused of that.

I attended a Nikon Df/58mm f1.4 AF-S event on Saturday in Tokyo, and 98% of the marketing message around the 58mm f1.4 is about the look of the bokeh and its behavior as a function of the distance behind the plane of sharpness. The remaining 2% were about sharpness, but only as a contrasting aspect to bokeh.

Always glad to hear.

Though the messaging at that one event is likely to have been catered to the audience they expected at that event. Nikon (like nearly every camera company I can think of) puts resolution high on it's marketing checklist.

When we host an event for fine art photographers our messaging is not the same as when we host an event for studio still life photographers even though the equipment is the same and both messages are honest.

[ErikKaffehr]

Hi,

Let me put this way, resolution is a good thing unless it affects noise and DR. Reducing pixel size affects DR but not really noise.

Smaller pixels have advantages, less aliasing, sharpen better. The only real disadvantage is processing time and in case of MFD high price.

Production cost is almost independent of pixel size, but MFD vendors charge a much higher price on 80 MP than say 60 MP, so it is probably very good business.

Best regards
Erik

Always glad to hear.

Though the messaging at that one event is likely to have been catered to the audience they expected at that event. Nikon (like nearly every camera company I can think of) puts resolution high on it's marketing checklist.

When we host an event for fine art photographers our messaging is not the same as when we host an event for studio still life photographers even though the equipment is the same and both messages are honest.

[bcooter]

Always glad to hear.

Though the messaging at that one event is likely to have been catered to the audience they expected at that event. Nikon (like nearly every camera company I can think of) puts resolution high on it's marketing checklist.

When we host an event for fine art photographers our messaging is not the same as when we host an event for studio still life photographers even though the equipment is the same and both messages are honest.

In the advertising world (whatever that is today) for life, illustrative and fashion photography, detail is a given.   No client really knows or cares is we use 18 or 31 mpx and a lot of our work goes to large in-store format.

What they do care about is the "look" and imagery that covers a lot of territory. 

What I care about is the look, the onset workflow and post workflow.  Today everything has to be faster and as unique as possible and faster doesn't mean less quality.

In other words whatever is shot on stills my clients want in motion imagery and whatever is shot in either medium, they want shot well so time is at a high premium.

Then once you have it they want a unique look.

It's nice Nikon is talking about the look of a lens rather than the graphs and charts, more makers should look at this.

I'm currently buying some duclos converted m43 f. 095 lenses.  They're not the sharpest lenses in the world but for motion imagery and some stills they do have a unique look, they glow in highlights and the flare like we expect film to flare when tracking across the scene into the sun.

Consequently for stills that oversharp, layered look is distinctive and is much easier to produce with a medium format camera than a dslr. 

IMO

BC

[ondebanks]

Is the larger sensor "better" only because it contains a greater number of pixels, or for some other reason?

It's not the greater number of pixels per se - it's the greater quantity of information. You can subdivide a fixed sensor size into more and more, smaller and smaller pixels, but at some point when the image is oversampled, there is no further gain in information. Maintaining the pixel size and lens performance while increasing the sensor area is still the best route to increased information (detail).

Ray