Is full frame really necessary?

[Stab]

3 hours ago, SteveV4D said:

Thats odd as S35 cinema cameras cost a lot more than many fullframe hybrids..   and I'd rather have the money to buy a C300, a Red Komodo or an URSA 12k over a fullframe A7C,  or any other fullframe hybrids for that matter... 😆😆

You must have a very large penis. 

[fuzzynormal]

3 hours ago, Stab said:

Finally, the full image to work with. 

Sad day .  To find out I’ve only been using parts of my images for years?  Bummer. 

[SteveV4D]

23 minutes ago, fuzzynormal said:

Sad day .  To find out I’ve only been using parts of my images for years?  Bummer. 

All those movies I've seen that were shot on S35 are now rubbish as they were using a cropped sensor... 🤣🤣

[Stab]

You guys know that I was being sarcastic right...

I'm just going into other people's heads and explain why many folks have a desire for full frame. And I think a part of that isn't the technical blah blah but simply the fact that it is FULL frame. And everything smaller than that is simply an inferior product that was only viable because of full frame camera's often didn't come with the same features. And now that they do, most people are 'upgrading'. It's the 'final destination' for most people I guess.

 

[tupp]

On 9/15/2020 at 11:13 AM, noone said:

The differences are quite minor to the point of it proves theory and practice match to me

If you think that an image with a building exhibiting sharp edges matches an image of the same building showing soft edges, so be it.  Everyone has a right to an opinion.

 

To me the differences between the images are so blatant that those discrepancies must be due to some uncontrolled variable(s) -- not to the difference in formats.  The differing looks/DOF between formats is usually more subtle than what we see in your comparison.  My guess is that the apertures were not at their equivalent settings.  Plus, the 1-inch camera used a zoom lens, and it might have had sharpening enabled.

 

 

On 9/15/2020 at 11:48 AM, Jay60p said:

To test this you could use:

standard 16mm, Bolex (Kern Paillard 10mm at f/1.8

4/3rds, 17mm at f/3.2

APS-C (Nikon DX), 22mm at f/4

Full Frame, 34mm at f/6.3

8x10, 256mm at f/45

The 16mm would have to be wide open and the 8x10 would have to be completely stopped down (my 8x10 270mm is f/4.4 - 45)

This is according to this calculator:

https://www.pointsinfocus.com/tools/depth-of-field-and-equivalent-lens-calculator/#{"c":[{"f":13,"av":"8","fl":50,"d":3048,"cm":"0"}],"m":0}

Thank you for posting this!

 

The comparison would need to employ a narrower focal length -- normal or tighter -- to show a more perceptible delineation of the limits of the DOF range.  Also, the DOF should be shallower to possibly show a more dramatic difference between formats.

 

 A Zeiss Superspeed 16mm set to f2 would work as the normal lens for the Super 16 format.  A standard, normal 360mm lens for 8"x10" is a close equivalent to that 16mm Ziess Superspeed, and the equivalent aperture on the 360mm lens would be somewhere in the range of f40-f51 (this aperture would need to be dialed-in for a visual match).

 

Maybe somebody with the resources and with the gumption will eventually make such a comparison.  If so, hopefully they will conduct their test in a setting conducive to revealing DOF/focus limits, perhaps with a receding fence as shown in the photo above.

 

Thank you for the informative and helpful post!

[tupp]

20 hours ago, SteveV4D said:

Thats odd as S35 cinema cameras cost a lot more than many fullframe hybrids..   and I'd rather have the money to buy a C300, a Red Komodo or an URSA 12k over a fullframe A7C,  or any other fullframe hybrids for that matter...

Have you priced an Alexa 65 lately?

 

If I had the money to buy an C300/Komodo/Ursa-12K, I'd buy an A7s II with an Irix 15mm lens and a shift adapter and build a copy of Zev Hoover's 8"x10" rig.  I'd deposit rest of the money.

[hyalinejim]

On 9/15/2020 at 10:28 AM, tupp said:

Do you not see in the area outlined in red how the distant white building exhibits sharp edges in the 1-inch image, while it is much softer in the full frame image?

I certainly can see what you're talking about in the areas you've highlighted. It's very clear.

But if you plug equivalent settings into a DOF calculator you get the same amount of DOF for both, according to the calculator. This is the theory, and it depends on certain assumptions regarding circle of confusion that I won't pretend I understand in depth.

So is it fair to say that your position is something like this:

"Even though DOF calculators show that the theoretical DOF of equivalent shots is the same, in practice there is an observable difference in how DOF is rendered between equivalent shots"?

If so, and if I understand you correctly, that would mean that a DOF calculator is showing us 2 points on the DOF continuum, the point of near focus and the point of far focus, and these are the same for both formats if the focal length and aperture is equivalent. But it's not telling us anything about the DOF characteristics elsewhere on the continuum, which is noticeably different.

That's certainly very interesting and worth investigating. I think for most people it's enough to know that they can match the field of view precisely using equivalence theory, and that the DOF is "the same" according to its assumptions. But others may notice and be very interested in differences in DOF behaviour that's not described by equivalence theory... if what you say is true! So we would be talking about a kind of DOF rolloff, which is shorter on larger formats and longer on small formats, according to equivalence sceptics.

Is this understanding correct @tupp?

 

[SteveV4D]

2 hours ago, tupp said:

Have you priced an Alexa 65 lately?

 

If I had the money to buy an C300/Komodo/Ursa-12K, I'd buy an A7s II with an Irix 15mm lens and a shift adapter and build a copy of Zev Hoover's 8"x10" rig.  I'd deposit rest of the money.

Oddly enough no.  My main point is that smaller sensors can cost more than fullframe.  Having more money doesn't necessary equate choosing fullframe.  Its as much down to personal choice of what you need or want to use - hybrids, smaller cameras, larger cinema cameras.  

I wouldn't buy a Sony as I've never liked using or editing their footage.  And to be honest, I don't shoot enough photos to need a camera that does both.  Maybe one fullframe I am thinking about to cover those small jobs where I do.  Mostly though, I want to stick to dedicated video cameras.

 

[tupp]

2 hours ago, hyalinejim said:

So is it fair to say that your position is something like this:  "Even though DOF calculators show that the theoretical DOF of equivalent shots is the same, in practice there is an observable difference in how DOF is rendered between equivalent shots"?

Yes, but the differences are not contained only within the front and back DOF limits.  Additionally, the character of the focus transitions at the DOF limits can differ between optics designed for different sized formats.  Likewise, how the focus behaves outside of the DOF limits can generally differ between formats.

 

 

2 hours ago, hyalinejim said:

If so, and if I understand you correctly, that would mean that a DOF calculator is showing us 2 points on the DOF continuum, the point of near focus and the point of far focus, and these are the same for both formats if the focal length and aperture is equivalent. But it's not telling us anything about the DOF characteristics elsewhere on the continuum, which is noticeably different.

Yes.  However, the differences are very slight (and sometimes non-existent) between optics designed for formats of similar size, for instance, M4/3 and APS-C, or APS-C and FF.

 

 

2 hours ago, hyalinejim said:

But others may notice and be very interested in differences in DOF behaviour that's not described by equivalence theory...if what you say is true!

The basic idea is that what's in focus is not limited the factors addressed in the DOF calculations (aperture, focal length, subject distance) -- the refractive optical elements have a huge influence over what is in focus, and how things resolve in a camera image.

 

An obvious example of how refractive optics can affect the range of focus is a split diopter.  A split diopter can near objects and distant objects into critical focus, regardless of whether the aperture is wide open or closed to its smallest setting.  By the same token, a split diopter can be used in "reverse" to make close objects sharp while making distant objects blurry, even if the aperture is stopped down considerably.  There are other examples of refractive optics affecting the focus range.

 

However, there is more involved in "format specific looks" than DOF and range of focus.  Optics for larger formats are generally less prone to aberrations and can usually resolve more lines per format frame.  Optics for smaller formats can require more aberration correction (more glass) and, although they necessarily resolve more lines per mm, optics for smaller formats have a harder time squeezing the same number of lines of resolution into the smaller format frame that is possible with lager formats.  These properties and limitations can affect focus, the flatness/shape of the focal plane, resolving and, hence, the "look."

 

 

2 hours ago, hyalinejim said:

So we would be talking about a kind of DOF rolloff, which is shorter on larger formats and longer on small formats, according to equivalence sceptics.

DOF "rolloff" is the street name for it.  Brian Caldwell expressed it in more technical terms.

 

I wouldn't say that the rolloff is "shorter" or more rapid on larger formats nor "longer" nor slower on smaller formats.  Nor would I say that the rate of the falloff is smooth/constant in larger or smaller formats.

 

[tupp]

24 minutes ago, SteveV4D said:

And to be honest, I don't shoot enough photos to need a camera that does both.

The A7s with the 8"x10" rig is not really intended for stills.

[hyalinejim]

23 minutes ago, tupp said:

I wouldn't say that the rolloff is "shorter" or more rapid on larger formats nor "longer" nor slower on smaller formats.  Nor would I say that the rate of the falloff is smooth/constant in larger or smaller formats.

 

 

26 minutes ago, tupp said:

However, there is more involved in "format specific looks"

This implies that you maintain that there is a look inherent to a format, independent of variations between lenses. If so, it should be consistent as format size changes and it should be describable. How does the look of small format compare to the look of a larger format, at equivalent focal lengths and apertures?

I'm just interested here, as I use DOF calculators to help my understanding when moving between FF, micro 4/3 and speedboosted micro 4/3. But I also see a huge difference in the images posted, which I would not have expected.

 

[SteveV4D]

1 hour ago, tupp said:

The A7s with the 8"x10" rig is not really intended for stills.

Probably not.. but it is a camera designed to do both regardless of how you rig it. Like my GH5 rigged to a gimbal is still a hybrid designed for a Photographer to use.  But no one would shoot photos that way.  At least, I hope not. 😆

[tupp]

6 hours ago, tupp said:

However, there is more involved in "format specific looks"

5 hours ago, hyalinejim said:

This implies that you maintain that there is a look inherent to a format, independent of variations between lenses.

 

The quotation marks that I employed imply something else.  I'm saying that refractive optical elements can affect focus and the focus range, and, additionally, that there are general tendencies, advantages and problems inherent in refractive optics designed for larger formats and likewise with refractive optics designed for smaller formats.  Of course, there are exceptions and some lenses for smaller formats possess some of the qualities generally found in larger format lenses, and vice versa.

 

 

6 hours ago, hyalinejim said:

If so, it should be consistent as format size changes and it should be describable.  How does the look of small format compare to the look of a larger format, at equivalent focal lengths and apertures?

There is consistency, but there is also seems to be more than one variable at play, so there is some complexity.

 

I would describe the look of larger formats as generally having a flatter and more "solid" focus plane with a faster "rolloff" at the DOF limits, but with a smoother and better resolved "macro-contrast" outside of the DOF limits.

 

 

6 hours ago, hyalinejim said:

I'm just interested here, as I use DOF calculators to help my understanding when moving between FF, micro 4/3 and speedboosted micro 4/3.

There is not a huge difference between FF and M4/3.

 

Using a speedbooster or focal reducer can allow the qualities of the larger format optics to be captured on a smaller format.

 

 

6 hours ago, hyalinejim said:

But I also see a huge difference in the images posted, which I would not have expected.

The dramatic discrepancies between the two images shown above are not due to any tendencies inherent in different sized formats.  My guess is that the DOF was not equivalently matched, plus the 1-inch camera likely had a built-in zoom lens (which can look/behave different than a prime) and excessive in-camera sharpening could have been enabled.

 

There are a lot of variables that need to be controlled in such comparisons, otherwise the tests are invalid.

[hyalinejim]

Well, you've convinced me that there's something worth considering there for sure. I had always conceived of the equivalence debates as being along the lines of flat-earthers etc 🤣

But now I see that for some it's just about a level of complexity and possibly even bokeh connosseurship.

For my purposes, these differences are so rarefied as to be irrelevant. However, the idea of testing for them is interesting. How about a test where you simulate a small sensor by using a center crop from a full frame stills camera? Like this:

Full frame
200mm, f7.1

Simulated 4x crop sensor
50mm, f1.8

These would be different lenses. I could do this (when I have time) with a Canon 50 1.8 and Sigma 70-200 f2.8 and use ACR to correct for lens aberrations. Or instead with an OM Zuiko 50 1.8 and OM Zuiko 200mm f4 (possibly similar primes? But can't correct for aberrations)

Would 4x be enough to show a difference?

I realise you're dealing with a lower megapixel image from the central portion of a wide open lens for the simulated 4x so the image will be softer.  Would these variables make the test invalid?

[seanzzxx]

There is an interesting post from cinematographer Steve Yedlin talking about this exact issue. As far as he's concerned, there is no unique 'look' to larger formats unless using a different format somehow forces you to set up your camera differently. https://www.yedlin.net/NerdyFilmTechStuff/MatchLensBlur.html

 

He has some pretty rigorous testing to go with it.

[tupp]

30 minutes ago, hyalinejim said:

How about a test where you simulate a small sensor by using a center crop from a full frame stills camera? Like this:

Full frame
200mm, f7.1

Simulated 4x crop sensor
50mm, f1.8

These would be different lenses. I could do this (when I have time) with a Canon 50 1.8 and Sigma 70-200 f2.8 and use ACR to correct for lens aberrations. Or instead with an OM Zuiko 50 1.8 and OM Zuiko 200mm f4 (possibly similar primes? But can't correct for aberrations)

Keep in mind that, to do a proper comparison, one must use the optics made for the formats being tested.

 

In addition, testing wide angle focal lengths is going to make it more difficult to discern any differences, so stick to lenses that are a normal focal length or tighter.

 

200mm is in the normal focal length range for the  4"x5" format.  So you need to use a 150mm-210mm lens made for 4"x5" and it has to be focused to a 4"x5" sensor/film sheet/DOF adapter.

 

50mm is the normal focal length for FF, so you need to use a 50mm FF lens on a FF camera.

 

By the way, here's another yet another overlooked variable -- view camera lenses for large format (2'x3", 4"x5", 8"X10", 11"X14", etc.) are designed produce image circles that are much larger than their format, because they have to allow for tilts, swings and shifts.

[tupp]

14 minutes ago, seanzzxx said:

There is an interesting post from cinematographer Steve Yedlin talking about this exact issue.

Thanks for the link!

 

I doubt that he is talking about this exact issue.  Like most other folks who do equivalency tests, he likely limits his attention to mathematical DOF, and his tests use wider angle lenses and there is no delineation of the front and rear DOF limits with a lot of other detail thrown away or ignored.

 

 

13 minutes ago, seanzzxx said:

He has some pretty rigorous testing to go with it.

I don't have time right now to read the linked page, but if the images shown are the extent of his comparison, his tests are invalid.  He does not show how the limits of DOF are delineated.  He seems to be using wide angles focal lengths, and I can see a difference in one of the images with just a glance.

[Mark Romero 2]

Late to the conversation here and haven't read all the replies.

I moved from my crop-sensor a6500 to a full frame Panasonic S1 for the better low light ability / increased dynamic range, and for the 10-bit 4:2:2 codec.

But that is because I shoot real estate and have to deal with ambient lighting. When shooting real estate videos, there often isn't time (or a budget) to set up lighting. So I often have a COUPLE of shots per video which are at the ISO 1600 to 3200 range.

Other times I am dealing with extreme levels of dynamic range.

The S1 does pretty well in these situations. Far better than the a6500.

[seanzzxx]

29 minutes ago, tupp said:

Thanks for the link!

 

I doubt that he is talking about this exact issue.  Like most other folks who do equivalency tests, he likely limits his attention to mathematical DOF, and his tests use wider angle lenses and there is no delineation of the front and rear DOF limits with a lot of other detail thrown away or ignored.

 

 

I don't have time right now to read the linked page, but if the images shown are the extent of his comparison, his tests are invalid.  He does not show how the limits of DOF are delineated.  He seems to be using wide angles focal lengths, and I can see a difference in one of the images with just a glance.

I feel like you are kind of moving the goalposts. This guy has a use case where he has shot hundreds of shots on cameras with all kinds of film backs (for camera comparisons), and somehow this does not count because his lenses are too wide? This is based on real-world experience with everything from an IMAX down to a super 35 camera. He even admits in the article that his matching is not perfect due to practical limitations (t-stops and f-stops not aligning, lenses not matching exactly to their equivalent counterparts, etc.), but his argument is that the likeness between shots is so convincing and consistent that the sensor size obviously does not play a role in the actual image, and that any perceived difference is due to bias or particular (non image circle-related) lens characteristics, not due to the size of the film back. In fact, where you have been previously arguing about recognizing a larger format due to increased lens blur (in your examples where you are circling a number of shots), the Alexa 65 actually seems to have slightly LESS lens blur in the examples provided by Steve Yedlin, likely due to my aforementioned reasons. This, again, seems to provide an argument that any perceived differences are more likely to be due to individual lens characteristiscs or other uncontrolled variables which are not related to the film back size.

EDIT: I hope this does not come off as argumentative, as I do appreciate -and enjoy- the discussion!

[noone]

11 hours ago, tupp said:

Have you priced an Alexa 65 lately?

 

If I had the money to buy an C300/Komodo/Ursa-12K, I'd buy an A7s II with an Irix 15mm lens and a shift adapter and build a copy of Zev Hoover's 8"x10" rig.  I'd deposit rest of the money.

Buy me a Arri master prime 150 1.3 and an adapter to use it on M43 and I will happily compare it to my ancient lowly Tamron 300 2.8 FF.

As for the rest, in your reply to me above, I disagree