• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

Sony "compact" with 33 mm × 44 mm format?

Doug Kerr

Well-known member
The Canon Rumors site reports that Sony may enter the "medium format" camera field with a camera using the same 50 Mpx CMOS sensor that is to be used in the new Hasselblad H5D-50C (about 33 mm × 44 mm*).
*That would be the "ten thirds inch" size ("1/0.3 inch"), or, as we would call it here, "55 mm".​
It is suggested that this might be a "compact" body (as distinguished from the fairly bulky Hasselblad configuration).

Could be a powerful machine.

Best regards,

Doug
 

Jerome Marot

Well-known member
The Canon Rumors site reports that Sony may enter the "medium format" camera field with a camera using the same 50 Mpx CMOS sensor that is to be used in the new Hasselblad H5D-50C...

...or not. Maybe.

Did you notice that Sony has a very nice "full frame" ("kleinbild", if you prefer) point and shoot camera? It is a powerful machine.
 

Asher Kelman

OPF Owner/Editor-in-Chief
The Canon Rumors site reports that Sony may enter the "medium format" camera field with a camera using the same 50 Mpx CMOS sensor that is to be used in the new Hasselblad H5D-50C (about 33 mm × 44 mm*).
*That would be the "ten thirds inch" size ("1/0.3 inch"), or, as we would call it here, "55 mm".​
It is suggested that this might be a "compact" body (as distinguished from the fairly bulky Hasselblad configuration).

Could be a powerful machine.

Best regards,

Doug

If one were to place it above the category of the sony RX-1, 24 MO full frame camera, then it would be a darn sight more than $2400, more like $4500 or $6,000 as the Pentax i645D 2014 is going to be about $10,000 or so. I'd imagine apart from me buying 5 to use simultaneously, to get a a 200 MB stitched image, I'd doubt more than 1000 cameras would be sold world wide.

However, it it comes out, I'll place my order!

Asher
 

Asher Kelman

OPF Owner/Editor-in-Chief
Hi, Asher,


We look forward to your bracket design.

Doug,

It depends on my approach. I might do primary capture of the subject and stitch with multiple viewpoints or else photograph the image inside a camera obscura so there are no parallax issues. At the moment I'm testing my 24MP output of my 6D to see whether or not one can be "satisfied" with the image enlarged to 80" tall. With human figures, pose, expression and gesture do much to make up the meaning of a work. So perhaps, I am already in good shape. Still, with an inexpensive 50mm sensor or the quality needed for the Phase One, Hasselblad and Pentax MF, it would be a wonderful opportunity to do even better that just "good enough".

Asher
 

Doug Kerr

Well-known member
Hi, Asher,

It depends on my approach. [Or]I might . . .photograph the image inside a camera obscura so there are no parallax issues.
That's clever - a two-stage camera!

While this approach eliminates parallax issues, it does not completely eliminate "perspective" issues in the second stage. But these should be easily dealt with in the "stitching" operation, since the "object" of the second stage is planar.

At the moment I'm testing my 24MP output of my 6D to see whether or not one can be "satisfied" with the image enlarged to 80" tall. With human figures, pose, expression and gesture do much to make up the meaning of a work. So perhaps, I am already in good shape. Still, with an inexpensive 50mm sensor or the quality needed for the Phase One, Hasselblad and Pentax MF, it would be a wonderful opportunity to do even better that just "good enough".

I look forward to the results of your investigations. We suffer from a lot of incomplete thinking as to what are the actual "demands" on resolution of different contexts in which our images are to be used.

A recent case on the forum brought this to mind. We have Fahim's splendid "blue" image of a woman wading just off a beach. (The beach seems to have a tiled apron!)

p840020099.jpg

Fahim Mohammed: Untitled

I wanted to see her "close up". But (in the image resolution as presented on the forum, of course) she is gravely pixellated. Here we see a small crop at 10X magnification from pixel-for-pixel:

p840020099-C1-10X.jpg

On the other hand, it is amazing how the original presentation conveys to my eye her obviously-stunning figure!

No wonder it is hard for us to figure out "what we need" in an image.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Look Ma, I'm at f16!

The next thing, Doug, as Bart Van Der Wolf has pointed out on numerous occasions, folk think they have a 20 MP c4/3 Olympus camera but use f16 in sheer bliss. Actually, they only have about 2MP. That's the limit even if they have a superb lens. with a Hasselblad, Phase One or Pentax MF camera, at f16, the camera, even with a .55 micron, 50MP or 80MP sensor, only works as if it had much, 15MP!! see this interesting article in LULA!

What's amazing is that very good photographers feel that their instinct is good enough and they ignore physics. Diffraction effects all camera systems.

I guess deconvolution can recover some losses but I'd guess only a stop or so. Even at f11 an 80 MP sensor only can give MTF's as if there were just 31MP. Things get better with larger sensors but then chromatic aberration, vignetting, distortions and many other optical issues come to bear as the lenses get larger.

Asher
 

Doug Kerr

Well-known member
Hi, Asher,

see this interesting article in LULA!


Thank you for that nice cite. That is an interesting article, which I think I have not seen before.

Especially valuable is its discussion of the numerous "metrics" of "resolution", what they mean, and how they are typically related. There is much misunderstanding about this area.

Of course a major "takeaway" is the matter of diffraction and how its impact, "other factors being equal", depends on format size.

A nice intuitive outlook is that, remembering that the exposure implication of a lens is (nominally) determined by the f-number, then if we want to retain the same exposure implication, but scale up the whole system (by adopting a larger format size), we have in effect reduced the (relative) wavelength of the light, thus "shrinking" the (relative) diffraction.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Another take away is that one should consider both f stops and pixel density in doing work for some particular end result.

Getting increase in DOF with smaller aperture costs loss of resolving power at a contrast level that is increasingly degrading to the image quality.

I personally hardly ever have an aperture beyond 5.6 with a APS-C sized or full sizes "35" mm sensor.

With LF, it jumps to f16.

Asher
 

Doug Kerr

Well-known member
Hi, Asher,
Another take away is that one should consider both f stops and pixel density in doing work for some particular end result.

Getting increase in DOF with smaller aperture costs loss of resolving power at a contrast level that is increasingly degrading to the image quality.

I personally hardly ever have an aperture beyond 5.6 with a APS-C sized or full sizes "35" mm sensor.

With LF, it jumps to f16.
Sure.

Of course, there is no single metric that expresses the "resolution" of a system, but we can certainly adopt some consistent arbitrary metric that will allow us to compare the "resolution" performance of two postulated systems. We might for example use the spatial frequency at the focal plane, in cycles per picture height where the MTF drops to 20% of its "low frequency" value.

Now let's only consider the impact of diffraction on resolution.

If we do that, then if for a format with a diagonal size of about 43 mm (ff35), we get a certain resultion (on the arbitrary basis adopted above) with an aperture of f/5.6, then with a format with a diagonal size of 55 mm, we would get that same resolution with an aperture of f/7.2.

To get that same resolution performance at an aperture of f/16, we would need to contemplate a format with a diagonal dimension of 123 mm (e.g., 74 mm × 98 mm). (I am ignoring the slight complication of differences in aspect ratio.)

Your discussion was quite reasonable couched in terms of the desire to use a "small enough" aperture to attain certain depth of field performance. Of course, here again we must be careful as to the conditions of comparison.

A pivotal matter here is our criterion for "negligible" blurring (as codified by our choice of a circle of confusion diameter limit). Two common outlooks are often used there.

A. We adopt a criterion of based on human visual acuity, assuming viewing of the image under certain highly-arbitrary conditions. That is, we consider as negligible that amount of blurring from imperfect focus that could not be noticed by the viewer (assuming that this blurring is not "swamped" by the blurring from diffraction and lens aberrations).

B. We adopt a criterion based on the resultion of the optical system, considering as negligible that amount of blurring by imperfect focus that would not significantly degrade the system resolution at the focal plane.

If we choose to work with criterion B, we have no hope of finding any general rule to help us see how diffraction can influence our ability to get a certain DoF performance with different format sizes.

So I will adopt criterion A as the premise for adopting a COCDL for DoF reckoning.

So the overall set of conditions I will adopt for comparison of DoF performance will be:

1. Focal length to give a consistent field of view.

2. Same focus distance

3. Same COCDL expressed as a fraction of the image dimensions.

4. Same f-number (except that in this exercise we will change that as needed for our purposes).

Now we will consider:

• camera X (with a format size of 30 mm × 40 mm, not a recognized format, but handy numbers)

• camera Y (with a format size of 60 mm × 80 mm, twice as large).

We will adopt a COCDL of 1/1400 the frame diagonal dimension in both cases.

We will assume focus at a distance of 100 m in both cases.

We will start by using, on camera X, a focal length of 100 mm and an aperture of f/2.0. The total depth of field is reckoned as about 300 m.

Now we will switch to camera Y, using a focal length of 200 mm, and for the moment still use an aperture of f/5.6. The total depth of field is now reckoned as about 81 m.

To regain the total depth of field of the previous case, we would need to adopt an aperture of about f/4.

This, simplistically, under the conditions of comparison I had adopted, we find that, to maintain the same total depth of field as we increase the format size, we must increase the f-number roughly proportionately to the format size (and of course this is a well known relationship).

But we saw above that, in order to have a consistent impact of diffraction on image resolution (normalized to image size; that is, as would be perceived by a viewer observing a consistent size print at a consistent distance), we may use an f-number that is proportional to format size.

Thus I don't see how an increase in format size helps us to attain a certain DoF objective in the situation where diffraction limits our ability to use an arbitrarily-small aperture.

This is of course all based on about 15 minutes of thought (and before breakfast as well). I may have gone badly wrong someplace.

Best regards,

Doug
 

Doug Kerr

Well-known member
Hi, Asher,

My apologies for the rambling nature of that last dissertation - It was not a "prepared paper", but rather just stream-of-consciousness palaver. (My serum glucose level was, however. 120 mg/dL.)

But to summarize its conclusions:

1. If we double the format size, then for the same diffraction impact we can use twice the f-number.

2. If we double the format size, then for the same depth of field performance we must use twice the f-number.

Best regards,

Doug
 

Jerome Marot

Well-known member
1. If we double the format size, then for the same diffraction impact we can use twice the f-number.
2. If we double the format size, then for the same depth of field performance we must use twice the f-number.

If we print at the same size.

But why would we do that? I don't know about your practice, but as far as I am concerned, when I use a camera with a bigger sensor, I normally get more pixels. And if I get more pixels, I can use them to print bigger.

The sensor we are talking about has pixels of 5.3 µm. For this large sensor, we have 50 mpix. A sensor with the same pixel pitch would have about 13 mpix in APS-C size and about 30 mpix in "full-frame"/"kleinbild" size. The large sensor is good for prints about 4 times as big as the APS-C one. If I don't want diffraction to kick in at the pixel size, I will need to use the same aperture (a maximum of about f/11).
 

Asher Kelman

OPF Owner/Editor-in-Chief
If we print at the same size.

But why would we do that? I don't know about your practice, but as far as I am concerned, when I use a camera with a bigger sensor, I normally get more pixels. And if I get more pixels, I can use them to print bigger.

The sensor we are talking about has pixels of 5.3 µm. For this large sensor, we have 50 mpix. A sensor with the same pixel pitch would have about 13 mpix in APS-C size and about 30 mpix in "full-frame"/"kleinbild" size. The large sensor is good for prints about 4 times as big as the APS-C one. If I don't want diffraction to kick in at the pixel size, I will need to use the same aperture (a maximum of about f/11).


Jerome,

Actually, as f11, with about the size of pixel you specify, the practical effects of diffraction, assuming the lens was pretty perfect, could be no more than about 7MP, not 13, no matter how many pixels are actually there. An APS C sized sensor must be used at f5.6 to get the full value of actual pixel count.

Look at the table in LULA, here. Or else, have we some condition where the larger number is in effect and I'm mistaken?

Asher
 

Doug Kerr

Well-known member
Hi, Jerome,

If we print at the same size.

But why would we do that? I don't know about your practice, but as far as I am concerned, when I use a camera with a bigger sensor, I normally get more pixels. And if I get more pixels, I can use them to print bigger.

Or not, depending on what you are trying to produce.

However, the discussion to which I was responding was as to the advantage of a larger sensor in allowing one to, within the aperture limits imposed by diffraction, attain a certain DoF objective.

We can of of course adopt many other models of comparison, such as "camera will fit in my glove compartment".

Oh, never mind....
 

Jerome Marot

Well-known member
Actually, as f11, with about the size of pixel you specify, the practical effects of diffraction, assuming the lens was pretty perfect, could be no more than about 7MP, not 13, no matter how many pixels are actually there. An APS C sized sensor must be used at f5.6 to get the full value of actual pixel count.

Look at the table in LULA, here. Or else, have we some condition where the larger number is in effect and I'm mistaken?


There are probably more than one "condition" that are not taken into account. The most obvious is the effect of the Bayer filter. The second most obvious is the choice of the criteria for determining the size of the diffraction spot. The third is the effect of lens aberrations.

In practice, when taking actual photographs of finely detailed real life subjects under ideal focus conditions, I notice the first minute lowering of contrast between f/5.6 and f/8 on the D800 (pixel pitch is 4.88 µm) and between f/8 and f/11 on the A900 (pixel pitch is 5.9 µm). f/11 and f/16 may be used without obvious ill effects even when peeping pixels but f/22 is obviously degraded.

Why don't you test your favourite cameras yourself? It is pretty simple to do, just chose a finely detailed subject (e.g. a landscape by clean air), and do a series at various apertures.
 

Jerome Marot

Well-known member
Or not, depending on what you are trying to produce.

However, the discussion to which I was responding was as to the advantage of a larger sensor in allowing one to, within the aperture limits imposed by diffraction, attain a certain DoF objective.

We can of of course adopt many other models of comparison, such as "camera will fit in my glove compartment".

Oh, never mind....

I think you have mistaken my point. I always said that the discussion of depth of field resides on assumptions that do not correspond to actual photographic practice. The formulas underlying hypothesis is that, whatever the camera, one will produce prints of the same size (or observe larger prints from a greater distance). This is not what people do and for a good reason: it does not make sense to buy and carry around a bigger and more expensive camera to produce the same pictures as with a smaller and cheaper camera.

A model that is internally consistent but resides on false assumptions is not scientific. You, of all persons, should be able to understand that.
 

Doug Kerr

Well-known member
Hi, Jerome,

I think you have mistaken my point. I always said that the discussion of depth of field resides on assumptions that do not correspond to actual photographic practice. The formulas underlying hypothesis is that, whatever the camera, one will produce prints of the same size (or observe larger prints from a greater distance).
That is true of only one of the several bases that may be adopted in the choice of COCDL. Others are often used.

The knowledgeable investigator, considering some actual characterizable model of photographic usage (characterizable up through the viewing context), in the light of which model we are interested in "depth of field" performance, may appropriately adopt some other basis.

I showed what the relationships were for a certain, well-stated situation. You suggest that it is an unrealistic characterization of the general case. I cannot quarrel with that.

I think that hardly qualifies as a case of "false assumptions".

Your view seems to be that generally speaking, photographers using a "larger format" camera will most often have its images viewed at a larger physical size than with the use of a "smaller format" camera. That is probably a reasonable outlook. But it does not lead to a different model than the one, upon which certain rules of thumb for adopting a COCDL are based, that the image will most typically be viewed with a certain angular size.

I fact, as I have often stated, if we wish to inquire into the depth of field implications of a certain photographic task, we must have (or adopt) a rather complete (if not necessarily precise) set of parameters of the entire context, including the context in which the image will be viewed, both in terms of geometry and in other more subjective terms.

Faced with this, we should certainly eschew trying to articulate general suggestions as to the impact on depth of field considerations of such differences between candidate tools as sensor size. For example, "A value of a larger format size is that we can then, within the limits on aperture imposed by diffraction considerations, more readily attain some desired depth-of-field performance."

Which in fact was my point.

Best regards,

Doug
 

Doug Kerr

Well-known member
Hi, Jerome,
In practice, when taking actual photographs of finely detailed real life subjects under ideal focus conditions, I notice the first minute lowering of contrast between f/5.6 and f/8 on the D800 (pixel pitch is 4.88 µm) and between f/8 and f/11 on the A900 (pixel pitch is 5.9 µm).

That is a very interesting and valuable observation.

If we consider the "diameter" of the Airy disk that results from the spreading of a point source under diffraction to be the diameter of the first zero ("dark ring") (we can simplistically consider this to be the diffraction "spot diameter"), and make our calculations at the commonly used wavelength of 555 nm, then in both cases, for the "critical f-number" (where degradation of modulation is first perceptible - I used halfway between the two bracketing values you mention), the diameter of the Airy disk is very nearly twice the pixel pitch.

Thanks for that data.

Best regards,

Doug
 

Doug Kerr

Well-known member
Hi, Jerome,

We can describe the "diffraction phenomenon" in terms of an MTF. To a very close approximation over most of the range, the diffraction MTF falls linearly with increase in spatial frequency, falling to approximately zero at a spatial frequency of 1/r, where r is the "radius" of the Airy disc (at the assumed wavelength).
The curve actually turns up a bit from linear toward the high end, so the actual value of the MTF at that frequency is about 10%.​

Based on your empirical data, where it seems as if the first notable intrusion of diffraction occurs where the radius of the Airy disc is approximately the pixel pitch, this would be when the MTF representing the effect of diffraction came to about 9% at the Nyquist frequency of the sensor (the frequency at or above which the sensor cannot capture components of the image).

I'm not yet sure what to make of that intuitively, but it is an interesting relationship.

I have a suspicion that the 9% is related to the MTF implications of the Rayleigh criterion for the ability of a system under diffraction to resolve separate points of light, but I need to ponder that some more.

Best regards,

Doug
 
I don't see the reason of Sony making an MF "compact" Doug... they may very well "crop" the sensor and make a much cheaper FF "compact" out of it... MO is that in the applications in which one would choose to use a "compact" the extra image area would add very little if any at all to affect the quality of the photograph... But again, MO also is that MF in the digital age has a meaning only if used in applications that FF sensors are of disadvantage (view cameras, still photography, multishot, architectural ...things like that), I don't see a reason at all for one using an ultra large sensor (and spend a fortune with respect to the FF alternative) just to do exactly the same kind of pictures he would by using FF... Clearly, DOF has a wider range to use with an FF sensor/fast lens combination, bulk is much reduced and cost difference is huge... Further more, I don't know if Sony would find it sensible on investing to develop a new system with different image area and different image area to mount distance and perhaps different lenses... In my view, it's like one buying a Ferrari to drive 20 blocks away from home (with traffic lights in the path) to work...
 
Top