Topics

Bigger is better and VistaVision

Thomas Gleeson
 

Art,

Thankyou for your considered response to the discussion of larger than S35 sensors and the “...increased "dimensionality” or "roundness" or "perspective" due to the larger frame size itself.” 

I totally agree with your observation in regards to the Panavision Videos https://vimeo.com/167045645 and Dan Sasaki's references to the benefits of magnification without explaining what he means with this term. I have initiated a conversation with Panavision and hopefully we can get a clear definition of the terminology.

I am surprised that only Art has engaged in this discussion. For a over a century cinematographers have been limited to the 35mm format and its many variations and for many shooters smaller than 35mm has been their fate. Since the 1920s people have pushed into larger formats but they have always been so expensive that only the largest Studio “Tent Pole” films have had access. With both Sony and Red delivering VistaVision capable cameras Large Sensor cinematography is becoming available to huge number of cinematographers. Defining the aesthetic qualities of Large Sensor cinematography seems an important task and surprisingly allusive.  

Do people think I am deluded and large sensor shooting will go the way of the 3D rig?

Stuart Brereton
 


On Dec 1, 2017, at 9:29 PM, lensboy235@... wrote
Thankyou for your considered response to the discussion of larger than S35 sensors and the “...increased "dimensionality” or "roundness" or "perspective" due to the larger frame size itself.” 
I am surprised that only Art has engaged in this discussion. 

I think one issue here is that these are completely subjective terms, referring to qualities that some people will not see as important, and other people will not see at all. If you choose focal lengths that match the FOV between s35mm and FF, and then shoot them both at apertures that match the DoF, at that point are ‘dimensionality’, or ‘roundness’ sufficient reasons to choose FF over s35? Given that you’ll almost certainly be sacrificing your favorite choice of cinema lens, and probably your favorite choice of camera to do so, is shooting ‘large’ format worth it?

Stuart Brereton
DP, LA

Michael Brennan
 

I have not found any of the explanations, tests or opinions on this subject, definitive.

Is roundness effect being discussed, partly a function/effect of the diameter of the front element as much as anything else? 

The “ants eye” perspective and the un-roundness ;) offered by a small lens diameter of 10mm is the opposite effect to the roundness effect under discussion?


Mike Brennan
DP
Melbourne 



The contents of this email are confidential

On 2 Dec 2017, at 4:29 pm, lensboy235@... wrote:

Art,

Thankyou for your considered response to the discussion of larger than S35 sensors and the “...increased "dimensionality” or "roundness" or "perspective" due to the larger frame size itself.” 

I totally agree with your observation in regards to the Panavision Videos https://vimeo.com/167045645 and Dan Sasaki's references to the benefits of magnification without explaining what he means with this term. I have initiated a conversation with Panavision and hopefully we can get a clear definition of the terminology.

I am surprised that only Art has engaged in this discussion. For a over a century cinematographers have been limited to the 35mm format and its many variations and for many shooters smaller than 35mm has been their fate. Since the 1920s people have pushed into larger formats but they have always been so expensive that only the largest Studio “Tent Pole” films have had access. With both Sony and Red delivering VistaVision capable cameras Large Sensor cinematography is becoming available to huge number of cinematographers. Defining the aesthetic qualities of Large Sensor cinematography seems an important task and surprisingly allusive.  

Do people think I am deluded and large sensor shooting will go the way of the 3D rig?

Thomas Gleeson
 

Stuart Brereton quote " think one issue here is that these are completely subjective terms, referring to qualities that some people will not see as important, and other people will not see at all. If you choose focal lengths that match the FOV between s35mm and FF, and then shoot them both at apertures that match the DoF, at that point are ‘dimensionality’, or ‘roundness’ sufficient reasons to choose FF over s35?”


Stuart,

These are good questions and I am attempting to qualify the changing aesthetic as the format size changes. This has proved strangely allusive. I am certainly not suggesting that one format is better and saying VistaVision is better than S35 as this is like saying Rasberry jam is better than Strawberry jam*. Although what is certain is that the ability to choose formatts to suit your project or even the shot is more power to cinematographers. With Arri offering the Arri65, Sony the Venice, Red the Monstro, Panavision the DXL, Cooke the S7's and Leica the Thalia’s etc we will not need be forced to far from our familiar tools.

I would still love to see if someone can clearly enunciate just how larger formatts see and represent the world differently? I have started a conversation with Red and Panavision who have at least attempted this on their Panalab site.


Tom Gleeson
Sydney DOP

*Clearly Strawberry jam is the best

Mark Kenfield
 

I think this is a really interesting area at the moment, and one that certainly deserves some more rigorous discussion than it’s had to date. 

 

Everyone seems to accept that ‘bigger is better’, but I agree, I don’t think the actual aesthetic qualities that people are referring to, have been well enough defined to actually mean anything.

 

There is obviously a clear distinction between the magnification properties of anamorphic glass and spherical glass (as Dan explained pretty clearly I think, in the video that lenboy235 linked to). However, I think the issue becomes much murkier when we start to compare different sizes of spherical formats. 

 

Obviously there are very noticeable jumps in resolving power, the appearance of grain/noise, and the optical separation provided by a shallower depth of field, as we step up from 8mm, to S16mm, to S35mm formats. But once we reach S35mm (and especially with the incredibly noise-free digital sensors we enjoy nowadays), the jump upwards to Vistavision and even 5-perf 65mm/Alexa65, really does become harder to discern at conventional viewing distances.

 

Add to this the fact that circle of confusion and screen viewing angles are constantly left out of these discussions (when they are so utterly fundamental to the perceptual differences rendered by these different-sized formats) and it becomes pretty clear why the whole issue has become as vague as the terms used to describe it (“roundness”, “natural perspective”, “dimensionality” etc.)

 

I think Steve Yedlin did a fantastic job of breaking a lot of this down in his “Display Prep” and “Resolution” Demos (found here: www.yedlin.net/BoringFilmTechStuff_Index.html ), and anyone who hasn’t sat through them yet, I’d strongly urge to take a look.

 

Using a depth of field and angle of view calculator (www.pointsinfocus.com has the best one going), we can see precisely the focal length and aperture setting required to match lenses between the different formats.

 

So if we take a 25mm f/1.4 lens on a digital S35mm sensor as a reference point, we would need a 38mm f/2.1 lens on a Vistavision sensor, or a 53mm f/3.0 lens on the Alexa 65 to match both the angle of view and depth of field between them.

 

Now you’d need an Alexa 65 to test this, as it can do all three formats on the same sensor (So Brett and Sean at Arri Australia, hit me up if you want to get this done!), but I suspect that if we did, and then projected the results all on the same screen, we’d find that (assuming we have nice, neutral, distortion free lenses) there would be no overly apparent optical differences in perspective or compression. And instead, the differences would lie in resolved image detail, and the relative size of the grain/noise in the image (which would get progressively smaller and less noticeable, with each step up in sensor size).

 

That’s the theory at least (from what little I understand of spherical optics).

 

Now the part that interests me most, is if you rated the Alexa 65 you were using for this test at 200 or 400 ISO (ISO levels at which the regular S35mm Alexa sensor, of which the Alexa 65 is made up of three, becomes practically noiseless); and you were sitting in the middle of a regular movie theatre, seeing the images projected at 4k, but, because you’re sitting in the middle of the theatre, where circle of confusion and your viewing angle to the screen, puts you right on the edge of being able to discern the difference in detail between a 2k and 4k image. Are you going to be able to tell the difference between a noiseless S35mm 25mm @ f/1.4 image, an even more noiseless 38mm f/2.1 Vistavision image, and a 53mm f/3.0 Alexa 65 image (which would probably appear noiseless even if you didn’t rate it at the lower ISO)?

 

By sitting in the middle of the theatre (a good reference point for the ‘average’ viewer) your ability to discern the higher resolution of the Vistavision and Alexa 65 images is limited by the resolving power of your eyes, and once the smaller S35mm format is rendered perceptually noiseless by the lower ISO, your ability to perceive that difference between the formats is limited too. 

 

So at that point, with all of the obvious perceptual differences minimised, is anyone going to be able to pick the differences between the images? And how significantly would things change if you then viewed the same images again from the middle of an IMAX theatre (where the circle of confusion/angle of view equation is changed completely)? I’d love to find out!

Cheers,

 

Mark Kenfield

Cinematographer

Melbourne

On Fri, Dec 1, 2017 at 09:34 pm, Thomas Gleeson wrote:
Thankyou for your considered response to the discussion of larger than S35 sensors and the “...increased "dimensionality” or "roundness" or "perspective" due to the larger frame size itself.” 

I totally agree with your observation in regards to the Panavision Videos https://vimeo.com/167045645 and Dan Sasaki's references to the benefits of magnification without explaining what he means with this term. I have initiated a conversation with Panavision and hopefully we can get a clear definition of the terminology.

Thomas Gleeson
 

Mark,

Don’t worry I have already hit up Arri and Panavision. In Australia a Arri65 is presently Unicorn poop and the DXL not much better. Sony will have a Venice at ACS HQ Sydney next Thursday and we will have a Monstro Vistavision to test with the ACS Technical Committee over the coming weeks.

There are several issues that need to be teased out over large sensors. One of the drivers of early large film formats was technical with major improvements in picture quality with better grain and resilience to the lossy analogue printing processes. I would also concur Mark that the huge improvements in first film stock grain then sensor noise means we really don’t need larger sensors for their marginal technical improvements. Its also not just about the pixel count or resolution as we have reached pixel densities that need not increase. Note that the 65mm Arri and Sony Venice are 6K and Red 8K so these cameras are really not pushing the pixel race.

The “difference” is a lens/geometry/perspective/DOF thing. Firstly the perspective is not determined by lenses and formats but by the position of the camera and geography of the subject and world photographed. Larger Formats do allow us to work closer using any given focal length due to their larger angle of view. These longer focal lengths naturally bring their DOF characteristics with them.

Say you had a Monstro camera shooting a MS of a subject in a room you shoot a shot using the S35 format with a zoom lens set to 24mm then switch the camera to VistaVision. We could then zoom into a 35mm focal length to match the AOV and have exactly the same picture except the DOF would decrease. Now stop down around two stops and the two images are identical.

Now do the same test but instead of zooming in to match the shot we stay on a 24mm but move forward to match the MS on the subject. We match the MS but now the perspective on the background has changed as we have a wider AOV and we have moved forward. The Subject and background relationship has changed as the subject has grown larger in relationship to the background. The other effect in moving forward is we are focusing closer to keep the focus on the subject allowing for greater focus separation as DOF decreases. This is where the magic may lie?

In my research looking at Big Sensor cinematography I keep coming across compelling images that jus look so different that I am compelled to dig deeper although my ignorance of optical physics does not help.

Tom Gleeson
Sydney DOP

Paul Curtis
 


On 2 Dec 2017, at 08:15, Mark Kenfield <mark@...> wrote:

Everyone seems to accept that ‘bigger is better’, but I agree, I don’t think the actual aesthetic qualities that people are referring to, have been well enough defined to actually mean anything.


In an end to end imaging chain there are numerous factors that affect the image. Some of these have a big contribution and others a small one. I think we all have a habit of looking at the minutiae due to the nature of what we do. I know i'm guilty.

For me the single biggest effect of a larger sensor is simply focal length vs DOF, that's what 99% of people will see and feel. I think that the other contributions and differences are really only discernible to folk like us, who are looking for them.

It's also a case of physical lens design too. On a large sensor the capability to make a wide lens with shallow DOF is much easier than on a smaller sensor. So we have a wide aesthetic now with clear separation. Just like medium format. So physics and practicality of manufacturing plays a part in this too.

All other aspects - noise, resolution and so on are valid, but i don't think an audience is really going to pick up on that like we do, but there will be some subconscious effects, that's true.

It's like the Red 8K image, anyone working with it can see that it shows the limitations of most lenses, if you pixel peep, but step back then there is a smoothness which yields an image i feel that's very organic - i'm sure the same is true for other similar resolution cameras. Where as with HD res, it's a bit easier to 'see' and 'feel' the pixels and aliasing.

My 2c

Cheers
Paul

Paul Curtis, VFX & Post | Canterbury, UK


George Palmer
 

Just rewatched the Vistavision  “How the West Was Won”.  The story of this truly magnificent film is largely fullfilled by legendary performers portraying the stories of ordinary people framed, dwarfed by mind boggling breathtaking awesome natural (not synthetic, CGI) American West settings.  Typical of places like the Great Australian Outback, or the mountains of Switzerland, or the Gobi desert.  Hence the ability to frame performers prominently and yet fully retain the awsome grandeur of awesome natural settings.  Set on the streets or in the buildings of NYC, or Chicago, or Sydney (vs. for instance the Great Outback) would we see any discernible (to the viewer!) benefits of Vistavision or these ultra large formats?

To me the point and the highest and best (only useful?) purpose of these formats is to portray the stories that capture the pedestrian stories and struggles of ordinary people against the sometimes overwhelmingly challenging yet visually awesome world around them.  Question is how many movies have you seen lately whose stories/scripts merit that treatment?  And if we were to simply justify the use of and deploy the large format tool on the basis of technical merit or to force a visual effect that most times would only fit the story line (e.g. a drug dealer in NYC or in Colombia), would we in fact be fousting a visual perspective inappropriately for the sake of the effect alone?  

IMHO aside from the cost and staging implications that may have restricted the use of Vistavision in its day, there were also not that many movies that have actually merited that treatment. If you really want to test that theory, spend some time watching classic films on TCM, then ask yourself how many would have justified or benefitted from the use of large format cameras.

George Palmer
On the WNY Tundra with no spectacular vistas in sight


--
George Palmer

George Palmer
 

Sorry a typo in the post: “And if we were to simply justify the use of and deploy the large format tool on the basis of technical merit or to force a visual effect that most times would only fit the story line (e.g. a drug dealer in NYC or in Colombia),” 

I meant to say “that most times would NOT fit the story line or setting”

GEORGE Palmer
Composing text one fingered on a phone is tough on the WNY tundra

--
George Palmer

Scott Dorsey
 

What drove large format film systems was primarily the huge sizes of screens
and the desire to get a sense of complete envelopment in the theatre.

These days, far too few films are shown even on 25-foot screens in multiplexes
let alone actual big screens. People are watching them on their cellphones
where they can hardly see a damn thing.

The magic isn't in shallow DOF. Shallow fields can be a good thing and can
give a certain look that is great. But when you shoot for a big screen with
a large format medium, you compose completely differently. It's not just
that you don't have to fill the frame with faces.... it's that you can now
have a face way out in the distance in the tiny corner of the frame and the
viewer can see every detail in it.

Next time someplace near you runs Lawrence of Arabia or 2001 on 70mm, go see
it. It's a totally different experience, because of the size of the screen,
the fact that the screen fills up almost your whole vision, and the film is
composed for that.

(Some people may argue that what really sold large format projection was
decent multichannel sound, and that was a big factor. Lots of stuff shot
on 35mm was blown up to 70mm for distribution because 70mm mag sound was
so much better... but the films shot and composed for the format are
different).
--scott


Hell, go watch The World, The Flesh, and the Devil. It's just in ordinary
Cinemascope and it's composed for a movie palace.

Stuart Brereton
 

On Dec 2, 2017, at 1:49 AM, Thomas Gleeson <lensboy235@...> wrote:

Say you had a Monstro camera shooting a MS of a subject in a room you shoot a shot using the S35 format with a zoom lens set to 24mm then switch the camera to VistaVision. We could then zoom into a 35mm focal length to match the AOV and have exactly the same picture except the DOF would decrease. Now stop down around two stops and the two images are identical.

Now do the same test but instead of zooming in to match the shot we stay on a 24mm but move forward to match the MS on the subject. We match the MS but now the perspective on the background has changed as we have a wider AOV and we have moved forward. The Subject and background relationship has changed as the subject has grown larger in relationship to the background. The other effect in moving forward is we are focusing closer to keep the focus on the subject allowing for greater focus separation as DOF decreases. This is where the magic may lie?
By moving the camera you are adding another variable to the test, and all it’s really demonstrating is that a wide angle close up looks different to a long lens close up. The question here is whether there is any perceptual difference between s35 and FF given matching FoV and DoF. With film as an origination format, grain would be a giveaway, but would the same happen with digital sensors? As Mark points out, modern 4k sensors already out resolve what the eye can see at normal viewing distances.

I’m looking forward to seeing some side by side tests to show how difference a larger sensor actually makes, and how much of it is just “mine's bigger than yours”.

Stuart Brereton
DP, LA

John Brawley
 



On Dec 2, 2017, at 12:29 AM, lensboy235@... wrote:

For a over a century cinematographers have been limited to the 35mm format and its many variations and for many shooters smaller than 35mm has been their fate. 


Tom, I’m not sure I’d define not shooting on larger formats than 35mm as a limitation, it perhaps belies your own preferences for larger formats?  

Cinema seems to have done OK for a century or so mostly getting by with this smaller and smaller formats.

I have always enjoyed photography and have used many formats.

I like the convenience of the small formats like 135 and heretically, I also love the mFT format too.

But I also still today shoot medium format digital and film alongside those formats.  It drives my assistants nuts to have to carry all these cameras around but I love using them for different purposes.

There’s something very special about shooting the Leica S (where the Thalia lenses came from) outdoors for semi formal portraits and the Leica M when you want a camera on hand while you’re operating is hard to beat and a Mamiya 7 is great for candid landscapes.  

I think of larger formats as yet another extension to our tool set.  Just like there is no one camera that does it all, I don’t believe there’s one format that does it all. It comes back to story right?  I saw Sully at the cinema before knowing it was shot almost entirely Alexa 65 and I have to say, it didn’t stand out to me as being the right choice for that film.  I didn’t even pick that it was Alexa 65 till I read about it later.

I’m surprised more aren’t exploring Nolen / Pfister approach on The Dark Knight, mixing not only formats but changing aspect ratios mid narrative. I thought it worked very well for the work, as it also did with films like Mommy where they shot 1:1 with a single wonderful moment where it changes to 1.85 when a central character literarily pulls the frame wider in shot. There’s also a nice change midway in Hunger Games 2. The Grand Budapest Hotel does it with three different aspect ratios, all driven from story.

Aspect ratio to me has a greater affect on visual phsycology because it drives shot composition and staging, rather than the visual aretefacts of image / format size and the optical requirements of meeting these differing sizes.

JB

John Brawley
Cinematographer 
Atlanta Georgia






sergio@...
 

What we have to remember is that in spite of the magnificent acceleration of innovation in digital capture systems, the high end glass industry has a completely different rate of novelty and breakthroughs. The fact the ARRI has launched an 8 perf 65mm digital sensor less than a decade after their S35 digital sensor is insanity. 

We we are in unprecedented territory when it comes to the speed of innovation in terms of sensors, codecs, HDR capture, light sensitivity, all the things which really excites and drives the enthusiasm for embracing digital cinema. What we have to remember is that not all aspects of capture can catch up at the same rate. If I’m not mistaken this was the thinking behind Deakins staying away from the Alexa 65. What’s the point of a state of the art, shiny new 8perf 65 digital sensor if you still have to use 60 year old glass?

Also, there are significant aesthetic differences between 5perf 65, 8perf 65 and S35 and FF sizes. 8K VV from Red and the sensor in the DXL is much more similar to S35 than the Alexa 65 6K is to the Red 5perf 65 sensor. 

The Alexa 65 sensor is at least 10 years ahead of its time (and glass technology). My guess is that people will slowly start upgrading their systems to accommodate FF (already well under way) and eventually VV 5perf and Arri’s 8perf 65. But I’m afraid the availability of these new technologies confuses people and gets them impatient to try them even though they do not yet tick all the boxes so to speak, and then they’re disappointed or confused by the results though they don’t account for the fact that they have to pair the cameras with inferior lenses and accessories. 

Leica and Cooke delivering modern FF high end glass is super encouraging indeed for those of us who have been waiting for this large format revival. Panavision has a whole host of vintage glass for VV and the Primo 70s can cover a 5K crop to the Alexa 65 sensor. But we’re certainly years (and millions of dollars ha!) removed from a Master Prime or Summilux calibre modern glass covering 8perf 65. There’s only so much you can do with the physics of light.
--
Sergio Montiel
Director  |  Cinematographer  |  Editor
New York  |  Latin America
+1 (212) 920-7150

Art Adams
 

>I think one issue here is that these are completely subjective terms, referring to qualities that some people will not see as important, and other people will not see at all.

If someone else sees something that I don't, the first thing I do is ask them to help me see it.

If we don't understand each other's subjective descriptions, then we should ask each other what we mean by them such that we can learn from each other.

Surely I'm not the only curious person on this list.

>If you choose focal lengths that match the FOV between s35mm and FF, and then shoot them both at apertures that match the DoF, at that point are ‘dimensionality’, or ‘roundness’ sufficient reasons to choose FF over s35?

Well, I guess that's an artistic decision. And we, as cinematographers, are paid to make those—many times a day.

It's fascinating to me that a background can seem more three dimensional by compressing depth, pulling it forward and reducing distortion. I then feel its actual shape because it's not artificially pushed into the distance, where any differences in depth are minimized. 

If backgrounds feel more real because they're a bit closer and possess a lot more detail, I'd gladly add that option to my artistic toolbox. It would also be nice to put a person in front of that and not see them look as if they've exploded from within.

As for my "favorite choice of camera," that's whichever will help me make the most appropriate pictures for the project. If a project looks best in a large format, and someone will pay for that, then a large format camera will be my favorite camera for that project.

Maybe I don't have access to all the lenses I want because they haven't yet been made. So what? Artificial constraints are one of the best ways to generate creativity.

Of course, maybe S35 is "good enough." There was a time when NTSC was good enough. Before that, Tri-X was good enough. Before that, tintypes were good enough. Before that, oil painting, wood carving and chiseling runes into stone blocks were good enough. Maybe scribes should have stuck with papyrus instead of moving to paper, because what's the point in giving up your favorite inks?

"Good enough" has never been a good enough reason for artistry to stagnant.

--
Art Adams
Director of Photography
San Francisco Bay Area

Art Adams
 

>There is obviously a clear distinction between the magnification properties of anamorphic glass and spherical glass (as Dan explained pretty clearly I think, in the video that lenboy235 linked to).

This is interesting. I looked for that effect in the anamorphic footage they presented and I didn't see it. I did see a lot of distortion, but I didn't see any difference in background magnification between the spherical lens tests and the anamorphic tests. How and where are you seeing it?

>But once we reach S35mm (and especially with the incredibly noise-free digital sensors we enjoy nowadays), the jump upwards to Vistavision and even 5-perf 65mm/Alexa65, really does become harder to discern at conventional viewing distances.

I wonder if that's true. I haven't done a good comparison. I thought the VistaVision vs. Technoscope comparisons were interesting, but I've noticed that the "roundness" effect tends to be more pronounced in formats with taller aspect ratios. There's a video on that same page of a guard walking through a prison checking on prisoners that has a very distinctive large format look but it's less pronounced than the VistaVision reference.

Then again, I took a look at some Lawrence of Arabia footage on Youtube and I could still feel that large format presence. It's the "epic" look that I remember from my youth that I could never quite put my finger on. 

 Add to this the fact that circle of confusion and screen viewing angles are constantly left out of these discussions (when they are so utterly fundamental to the perceptual differences rendered by these different-sized formats) and it becomes pretty clear why the whole issue has become as vague as the terms used to describe it (“roundness”, “natural perspective”, “dimensionality” etc.)

The difference in perspective vs. angle of view is visible regardless of the format. A shot that has the angle of view of a 25mm S35 lens but has no distortion and a closer, richer background should be a pretty obvious difference to us as cinematographers. Is everyone going to notice? Not consciously.  But most people don't consciously notice lighting either, but we still do it.

Now you’d need an Alexa 65 to test this, as it can do all three formats on the same sensor (So Brett and Sean at Arri Australia, hit me up if you want to get this done!), but I suspect that if we did, and then projected the results all on the same screen, we’d find that (assuming we have nice, neutral, distortion free lenses) there would be no overly apparent optical differences in perspective or compression.

You can see the differences in depth compression in the Panavision video. Look at the "person in front of the Xmas lights" clips and compare spherical S35 to spherical large format. The background looks to be half the distance away in the large format clips.
 

Are you going to be able to tell the difference between a noiseless S35mm 25mm @ f/1.4 image, an even more noiseless 38mm f/2.1 Vistavision image, and a 53mm f/3.0 Alexa 65 image (which would probably appear noiseless even if you didn’t rate it at the lower ISO)?

What you seem to be saying is that larger sensor cameras could be rated faster, which would compensate somewhat for their reduced depth of field.

You also seem to be making a case against, say, projecting 35mm film vs. 70mm film. Or viewing a 35mm print made from 70mm film. In either case, I think we'd all agree there'd be a noticeable difference. In order to display true 4K resolution you'd need to capture a minimum of 5K anyway.

And before anyone points me to Steve Yedlin's video again, yes, I know that resolution alone doesn't make a huge difference in every situation. But there is a valid argument that vastly increased resolution does result in much smoother images with more subtle transitions. The best example I've seen of that is here:


Will everyone notice? No. Are we in the business of making images that appeal to the lowest common denominator in every instance? No.

So at that point, with all of the obvious perceptual differences minimised

I don't understand why people keep saying this. THERE IS NO WAY TO ELIMINATE OBVIOUS PERCEPTUAL DIFFERENCES BETWEEN S35 AND LARGE FORMAT PHOTOGRAPHY. When you have to use a 50mm lens to capture an image in large format with the same angle of view as a 25mm lens in S35, you're going to see perceptual differences—because the 25mm is still a 25mm lens, and a 50mm lens is still a 50mm lens, and they have very different characteristics when it comes to depth representation.

--
Art Adams
Director of Photography
San Francisco Bay Area

Art Adams
 

>If I’m not mistaken this was the thinking behind Deakins staying away from the Alexa 65. What’s the point of a state of the art, shiny new 8perf 65 digital sensor if you still have to use 60 year old glass?

I've been using 60-year-old glass for my S35 work with great success. And there are people saying that old glass works really well in large format because it exaggerates the amazing tonal transitions that the higher resolution large format sensors naturally create.

Deakins loves Master Primes because they are crazy sharp and perfect. I don't. I'm not Deakins, but there are plenty of other cinematographers who are not Deakins and have different preferences to his. I don't think his opinion represents enough people in the industry that glass manufacturers should commence layoffs.

But I’m afraid the availability of these new technologies confuses people and gets them impatient to try them even though they do not yet tick all the boxes so to speak, and then they’re disappointed or confused by the results though they don’t account for the fact that they have to pair the cameras with inferior lenses and accessories.

So... echoes of RED in 2007? We survived that, and it only gave us more and better tools in the long run.
 
But we’re certainly years (and millions of dollars ha!) removed from a Master Prime or Summilux calibre modern glass covering 8perf 65. There’s only so much you can do with the physics of light.
 
If you have to give up half your paint brushes to work with a larger canvas, then maybe that's a decision you make when the opportunity comes along to work with a larger canvas. It doesn't mean that larger canvases are bad.

I used to think I was becoming old and curmudgeonly, but actually  I'm feeling pretty good at the moment.

--
Art Adams
Director of Photography
San Francisco Bay Area

Stuart Brereton
 


On Dec 2, 2017, at 10:20 AM, Art Adams <art.cml.only@...> wrote:

Surely I'm not the only curious person on this list.


You’re not. I am very curious to see side by side tests, but my curiosity is tempered by a certain amount of healthy skepticism.


Of course, maybe S35 is "good enough." There was a time when NTSC was good enough. Before that, Tri-X was good enough. Before that, tintypes were good enough. Before that, oil painting, wood carving and chiseling runes into stone blocks were good enough. Maybe scribes should have stuck with papyrus instead of moving to paper, because what's the point in giving up your favorite inks?

"Good enough" has never been a good enough reason for artistry to stagnant.

This argument makes no sense. With the exception of NTSC television, none of those mediums was used with the intention of faithfully recreating reality. As a case can be made that Digital cinema cameras are now at a point where we cannot perceive the incremental improvements, it’s fair to say that perhaps good enough is just that. To quote a certain Vulcan scientist, “A difference which makes no difference, is no difference."


"I don't understand why people keep saying this. THERE IS NO WAY TO ELIMINATE OBVIOUS PERCEPTUAL DIFFERENCES BETWEEN S35 AND LARGE FORMAT PHOTOGRAPHY. When you have to use a 50mm lens to capture an image in large format with the same angle of view as a 25mm lens in S35, you're going to see perceptual differences—because the 25mm is still a 25mm lens, and a 50mm lens is still a 50mm lens, and they have very different characteristics when it comes to depth representation.”

I believe that this is one of the points in contention. Do they in fact have different characteristics when FoV and DoF are matched? Some say yes, some say no. If these characteristics are demonstrable, then great, but if they’re just a matter of opinion, then the move towards larger sensors starts to look like something driven more by marketing than necessity.


Stuart Brereton
DP, LA


Luis Gomes
 

Well? Having short depth of field on wide angle lenses is surely interesting. 
As I always had laughed when my director says  “blur now” on a 24 mm 

Luís. 
Free day end. 
Xmas concerts 
Everyday from now. Till I vomit 
Finland. 
--
Gomes.luis@...
http://fi.linkedin.com/pub/luis-gomes/20/11b/335/
Freelancer video Professional. 
Finland. 

George Palmer
 

“If these characteristics are demonstrable, then great, but if they’re just a matter of opinion, then the move towards larger sensors starts to look like something driven more by marketing than necessity.”  Amen! The emperor’s new clothes weren’t visible, new or clothes.  But his marketing skills sure were.  Not saying this is exactly the case here, but if not quantifiable......maybe....

George Palmer
Quantifiable WNY Tundra Currmudgeon 
--
George Palmer

Stuart Brereton
 

Here’s a quick, and not very scientific test I shot. The first frame is from a Canon 6D Full Frame camera. 50mm lens @f4. The second is from a Fujifilm APS-C camera. 35mm @f2.5. According to my FoV calc, the actual equivalent should have been a 31mm lens, but they don’t exist… The images have been corrected to balance exposure, and resized to a web friendly resolution. The slight difference in framing is due to the tripod mounting thread being in different places on the cameras.

I am genuinely interested to hear thoughts on the differences in perspective and ‘dimensionality’, or lack thereof.

 

Stuart Brereton
DP, LA


George Palmer
 

OBTW, though this probably belongs in CML Glass ( like much of this thread ) some have alluded (and lamented?) on the 60 yr old technology embodied even in our modern lenses. Hate to be a realist, but beyond incremental improvements in glass formulations, grinding and polishing techniques, and ever diminishing worldwide rare earth coating resources, glass is glass.  AND you get what you pay for; cheaper is never better.  

BUT we are obligated as professionals to constantly be self reminded that film and that “60 year old ”lens technology IS the benchmark by which we have quantified the eagerly sought after phantasmic 6K-8K Film resolutions. And if sidebar off the record comments matter, my conversations with friends over the years at a WNY film manufacturer lead me to believe that few folks anywhere ever actually saw quantifiable true projected “plus” 6K film resolutions because of commercial projection lens technology. Lens technologist  friends however assure me that aside from minor coating tweaks, the film (and yes even TV) lenses produced in the last 30-35 years have been easily at least 4K-6K capable.

Perhaps the combination of pixel rich (6K-8K) digitally captured imagery and laser scanning have or will remedy part of the projection part, but our cameras will still most likely continue, going forward, to be using glass lenses for photon capture.
--
George Palmer

Art Adams
 

>As a case can be made that Digital cinema cameras are now at a point where we cannot perceive the incremental improvements, it’s fair to say that perhaps good enough is just that. To quote a certain Vulcan scientist, “A difference which makes no difference, is no difference."

Which cameras? HD cameras? 4K cameras? 8K cameras?

I can see those differences. And while they might be subtle to some, they are still useful. There are plenty of other subtleties that we employ in our work that are less subtle than sensor size.

Since when do we try to talk people out of letting us use better tools for imaging?

Do they in fact have different characteristics when FoV and DoF are matched?

Yes. I guess the part of my last email where I pointed out that the Panavision videos show this in the comparison between spherical S35 and spherical 70mm didn't make it through. (It's in part 1.)

--
Art Adams
Director of Photography
San Francisco Bay Area

Art Adams
 


Here’s a quick, and not very scientific test I shot.

In the second image, the background camera appears noticeably farther away to me. I measured it, and indeed it's smaller.

The super soft background doesn't help much in this kind of comparison as the differences become harder to judge.

--
Art Adams
Director of Photography
San Francisco Bay Area

George Palmer
 

A small correction to my previous lens technology post “aside from minor coating tweaks, the film (and yes even TV) lenses produced in the last 30-35 years have been easily at least 4K-6K capable.”  Should read   “(and yes SOME even TV) lenses”

George Palmer
--
George Palmer

Stuart Brereton
 


On Dec 2, 2017, at 6:30 PM, Art Adams <art.cml.only@...> wrote:

In the second image, the background camera appears noticeably farther away to me. I measured it, and indeed it's smaller.

it’s not. There’s no difference that I can see.


I don’t have a dog in this fight, and truthfully, I would love there to be some aesthetic difference, but I just don’t see it.


Stuart Brereton
DP, LA

Merritt Mullen
 

On Dec 2, 2017, at 5:41 AM, George Palmer <drvid@...> wrote:

Just rewatched the Vistavision “How the West Was Won”.
Just to make a correction -- "How the West Was Won" was shot in Cinerama, which consisted of three 6-perf 35mm frames (captured with three 27mm lenses). It was not a particularly shallow focus image because of this, though it was very high in resolution. You might be thinking of "The Searchers", which was shot in VistaVision.

I agree with Stuart and my own tests show this, if you shoot with a 50mm lens on an S35 sensor and a 75mm lens on a VistaVision sensor, and you match distance to subject and match depth of field by stopping down the 75mm by 1.5-stops, you get rather similar images.

But since people often use the same shooting stops whether they work in S35 or larger, they generally get a shallower-focus look with the larger formats. Besides that depth of field difference, and any increase in resolution that a larger format might have, the unique characteristics of the optics that some large formats lenses have due to design or age will also add to the look. With large format film, there was also a difference due to the reduced grain size.

David Mullen, ASC
Los Angeles

Stuart Brereton
 

To explore this further, I’m thinking of doing another test. I have a 6x7 camera here, and looking at PCam FoV calc tells me that an APS-C 35mm and 50mm will exactly match the Vertical FoV of the 6x7 127mm and 180mm lenses. If there are differences in perspective and compression, they should be plainly visible between two such different sized formats. 6x7 is more than double the size of the Arri 65 sensor. I’ll try to compare FF as well.

Stuart Brereton
DP, LA

Jessica Gallant
 

There are too many variables in the image. I would suggest doing the next test with a simple solid
colored sphere against a neutral background.

Since one of the descriptives is “roundness”, let’s see if the sphere is perceived as slightly rounder
in FF than in S35mm. Post both photos side by side, but don’t say which is which, then ask which
sphere appears rounder than the other. Options would be 1. sphere A appears rounder, 2. sphere
B appears rounder, or 3. both spheres appear to be equally round.

My guess is the most likely outcome will be that the FF sphere will be perceived as slightly rounder
then the S35mm sphere by a small majority of test subjects. This result would indicate further testing
is necessary to determine why a sphere appears rounder in FF than S35mm.

Next most likely is an even split between spheres A and B, or no difference between A and B. This
result would indicate that other factors are involved, such as viewer expectations or perhaps an easier
to obtain shallower depth of field which would be the next variable I’d want to test.

I doubt anyone would perceive the S35mm sphere as rounder.

Jessica Gallant
Director of Photography | Los Angeles | CA
http://jessicajgallant.com
http://wb.imdb.com/name/nm0002680/
cell: 818-645-2787
email: jessicajgallant@...
Skype: jessicajgallant



On Dec 2, 2017, at 7:31 PM, Stuart Brereton <ssbrereton17@...> wrote:

To explore this further, I’m thinking of doing another test.

Stuart Brereton
 

On Dec 2, 2017, at 8:14 PM, Jessica Gallant <jessicajgallant@...> wrote:

There are too many variables in the image. I would suggest doing the next test with a simple solid
colored sphere against a neutral background.
What do you mean by too many variables? The two shots were taken from the same position, with matched FoV and DoF. The subject matter is exactly the same in both images. The purpose of the test was to see if there was any discernible difference in perspective and field compression. i couldn’t see any, but others may differ.

I think you are taking the ‘roundness’ thing too literally. In The context it was being used in was not to describe the physical shape of things, but rather an intangible sense of ‘dimensionality’.

Stuart Brereton
DP, LA

Can’t believe I just typed "an intangible sense of ‘dimensionality’.”

Steve Oakley
 

thanks for redoing this. there was a similar test shoot on petapixel a month or two ago showing the same results.

there is one difference, the FF sensor’s bokeh starts a little sooner and a little more intensely a few inches behind plane of focus. however the lens itself could easily cause this so that too might be a non-issue. likewise the FF has slightly softer blur of the BG causing something of a softer contrast. 

all very subtle and generally not noticeable to the average viewer. all into the  95+% no difference category.

that said, 4K reduced to 1080 generally tends to make for  more contrasty  / smoother edges image thats discernible regardless of distance. hence canon’s use of a 4K sensor in the C300 down sampled to 1080. worked out for them pretty well and people still really like the image from that camera.

having seen 8K projected @ 8K, its amazing if you need to deliver WALL sized images that will be viewed reasonably close. the detail can be amazing and leave you asking for a little more res. thats a very specific and ideal viewing situation. 

 back in the real world I really get the sense that 4K as a deliverable shot with a really good 4K-6K camera will suffice. Why ? well because our eyes aren’t getting updated with higher res ! and 8K sensor sampling to 4K-6K may well be good enough because its all our eyes, in **average** viewing environments and situations will ever need or use. most of us have eyes good enough to see the difference between 4k & 1080, but not all. So while there is some arguement to be made about cleaner edges, cleaner transitions on significant changes of contrast, I’d much rather process that down in camera to a managable size rather than deal with all the extra data thru the rest of the process.

S


On Dec 2, 2017, at 8:43 PM, Stuart Brereton <ssbrereton17@...> wrote:


On Dec 2, 2017, at 6:30 PM, Art Adams <art.cml.only@...> wrote:

In the second image, the background camera appears noticeably farther away to me. I measured it, and indeed it's smaller.

it’s not. There’s no difference that I can see.

<split.jpg>

Thomas Gleeson
 


On 3 Dec 2017, at 1:43 pm, Stuart Brereton <ssbrereton17@...> wrote: it’s not. There’s no difference that I can see.



I agree with Stuart there should be little difference. If the two cameras with different sized sensors had matched nodal points (in the same position) and the AOV was matched with the focal length I would not expect there to be any difference in the geometry of the two shots.  Geometry or dare I use that word “Perspective” is determined by the cameras physical position in relation to the subject and background. Changing the focal length of the lens does not flatten or exaggerate the perspective as that is locked in by the geography. Check out Steve Yedlin’s great paper at http://www.yedlin.net/170504.html#AoV

Conceptually this can be a struggle to understand as we use long lenses to compress an image but of course this “compression” only works for objects far way but a long lens allows us to “zoom” into and single out these far away objects. What each focal length does is bring with it its DOF and by going with a longer focal length lens the larger format camera would have more focus separation. Stuart has stopped down to match the DOF but its one clear point of difference is that larger formats allow focus separation on wider AOV shoots. This is a biggie as the perceived 3D nature of a wide shot is greatly enhanced with some focus separation. Check out another of the fabulous Panalab videos particularly the Hateful Eight Camera tests https://vimeo.com/167052302

Are the DP’s of just recent big format productions such as the Revenant, Doctor Strange, Ghost in the Shell, Dunkirk, Spectre, Rogue One, Allegiant, Assassins Creed, The Hateful Eight, Murder on the Orient Express and Guardians of the Galaxy 2 helmed by DP’s such as Emmanuel Lubezki, Bob Richardson, Phedon Papamichael, Hoyte Van Hoytema, and Greig Fraser just drinking the cool aid? Maybe but I suspect there is more?

I have initiated this conversation from a genuine position of naivety. I largely avoided using the 5D Canon camera as it was such a poor tool but its Full frame images intrigued me. I watch films such as the Revenant and I see a unique image and it galls me I can’t fully articulate what makes them so interesting.

John I did not mean to suggest that the 35MM has been a limitation only as an industry we have not had many economically practical choices to go larger and its certainly not my proposition that large formats are going to take over. I agree with your statement " larger formats as yet another extension to our tool set." Yes I am showing a personal preference for larger formats as I grew up with 16mm then 2/3” cameras and it was a revelation moving to the larger 35mm. Even using those silly adaptors that allowed you to shoot 35mm lenses projected on to a ground glass using a B4 2/3” camera could capture a powerful “essence” of 35mm on video. What the hell was the name for them?

As of next week via the ACS we have full access to a Red Monstro and limited access to Sony’s Venice camera for testing so suggestions for tests to demonstrate principles or photographic effects are very welcome.

Tom Gleeson 
Sydney DOP

Art Adams
 

Panavision seems to have pulled the videos I referenced from their DXL/Panalab web pages. They do speak of large format photography's reduced perspective distortion here, on page 8:


As best I can tell, the "roundness" happens in some backgrounds because depth is compressed enough in wide shots that textures and shapes are more easily discerned, whereas in S35 a wide lens exaggerates distances and effectively flattens backgrounds.

I grabbed some stills from the missing videos. The VistaVision still below illustrates the "roundness" effect very well, as the lens is clearly somewhat wide but the background is close enough that I can feel the different degrees of depth.


I don't think this is the greatest comparison. The VistaVision image has a tremendous amount of subtlety due to the hazy sky. lower contrast, and large negative. The Technoscope image doesn't have much subtlety, and is much more contrasty, although that probably worked better in a lower-res format. Still, the background in the VistaVision image feels like it has more depth, even though the Technoscope image looks to have been shot with a wider lens.

This "roundness" seems to work better in tall frames. I've seen it in the 5D and I see it above. It's somewhat present in my go-to references—David Lean films—but there's something about a taller frame that exaggerates this effect.

--
Art Adams
Director of Photography
San Francisco Bay Area

Stuart Brereton
 

On Dec 2, 2017, at 10:42 PM, Thomas Gleeson <lensboy235@...> wrote:
Stuart has stopped down to match the DOF but its one clear point of difference is that larger formats allow focus separation on wider AOV shoots.
I’m not convinced that they do. This admittedly crude, test shows very little difference in focus separation between the formats, and what there is could easily be attributed to variables in lens design, and actual aperture, as opposed to indicated.



Are the DP’s of just recent big format productions such as the Revenant, Doctor Strange, Ghost in the Shell, Dunkirk, Spectre, Rogue One, Allegiant, Assassins Creed, The Hateful Eight, Murder on the Orient Express and Guardians of the Galaxy 2 helmed by DP’s such as Emmanuel Lubezki, Bob Richardson, Phedon Papamichael, Hoyte Van Hoytema, and Greig Fraser just drinking the cool aid? Maybe but I suspect there is more?
I think it’s a mistake to include film originated movies, as no-one is denying that having a larger neg is a benefit when shooting with film. It’s whether ‘large’ format digital origination is a benefit is the question here.

Stuart Brereton
DP, LA

Stuart Brereton
 

On Dec 2, 2017, at 10:55 PM, Art Adams <art.cml.only@...> wrote:
As best I can tell, the "roundness" happens in some backgrounds because depth is compressed enough in wide shots that textures and shapes are more easily discerned, whereas in S35 a wide lens exaggerates distances and effectively flattens backgrounds.

I don't think this is the greatest comparison. The VistaVision image has a tremendous amount of subtlety due to the hazy sky. lower contrast, and large negative. The Technoscope image doesn't have much subtlety, and is much more contrasty, although that probably worked better in a lower-res format. Still, the background in the VistaVision image feels like it has more depth, even though the Technoscope image looks to have been shot with a wider lens.
It’s very difficult to tell anything from those small frames. I’d expect an 8 perf Vistavision frame to have greater amount of subtlety compared to a 2 perf Techniscope image, which, after all, is not that much bigger than s16. As both frames are of different scenes, it’s comparing apples to oranges.

Stuart Brereton
DP, LA

Stuart Brereton
 


On Dec 2, 2017, at 10:55 PM, Art Adams <art.cml.only@...> wrote:

Panavision seems to have pulled the videos I referenced from their DXL/Panalab web pages. They do speak of large format photography's reduced perspective distortion here, on page 8:


That article is misleading, and sounds very much as if it was written by marketing people, rather than people who knew what they were talking about.

For example:

 "One way to pose the problem is to ask: which lens is used for a wide shot? Master director Alfred Hitchcock shot most often with a 50mm. Master photographer Henri Cartier-Bresson shot all of his larger-format photographs with a 50mm. The late Gordon Willis, ASC, favored the 40mm.”

This statement seems to suggest  that Hitchcock used a 50mm for wide shots, which is not the case. It then makes an equivalency with Cartier Bresson also using a 50mm, even though he was shooting FF 35mm, where the FoV would be different. And what exactly are ‘larger format photographs’?

Or this:

"Let’s assume that the central Field of View of human vision has an angle of about 45 degrees, with Perspective corresponding to roughly a 40mm lens. To get the same Field of View in Super 35 you need to shoot with a 27mm lens, or thereabouts, but then the perspective will feel a little exaggerated. To get a similar Field of View with a Dragon 6K, you would shoot with a 35mm, which is closer to 40mm and would yield a less exaggerated, more natural perspective."

A 40mm lens corresponds to human FoV only when attached to a s35 camera. Put one on a 6x7, and it’s very wide, put one on a 2/3” camera and it’s telephoto. This whole article is littered with factual inaccuracies. It may have come from Panavision, but it certainly wasn’t written by anyone who understood the subject.

Stuart Brereton
DP, LA

Geoff Boyle
 

I definitely prefer the lower image, the smaller sensor…

This has triggered off a really horrible thought pattern for me.
I’ve got to test this.
As I’m getting close to testing the Venice and it’s a camera that has “standard” frame sizes ie 3 perf, 4 per and FF, I need to shoot some tests with lenses that have matching fields of view.
This is the only camera that I can do this on easily and there is only one manufacturer that makes S35, anamorphic and FF lenses as far as I know.
So, Venice with Cooke S4, St and anamorphic…

Paging Richard Lewis, Peter Crithary and Les Zellan...

Cheers
Geoff Boyle
Cinematographer
EU Based
www.gboyle.co.uk
+31 (0) 637 155 076



On 2 Dec 2017, at 21:25, Stuart Brereton <ssbrereton17@...> wrote:

I am genuinely interested to hear thoughts on the differences in perspective and ‘dimensionality’, or lack thereof.

Argyris_Theos_cml
 


Stuart,

I am totally convinced that there is no rounding effect between formats and that all differences are being caused by the geometrical parameters/differences/errors associated with specific lenses.

However if you are going to test again, allow me to suggest that you use a sphere like this:

In this way the distances between meridians and parallels will be measurable.
​All you need is to place two spheres in different distances.

​I would also suggest that you allign the nodal points and that the foreground sphere is not too close because this would make the size of the front element matter (big front elements MAY show excessive roundness in objests placed too close - could one of the attending lens experts verify this please?)

​Best Regards

​Argyris Theos, gsc
DoP, Athens Greece
​+306944725315
​theos@...
​skype: Argyris Theos

Michael Brennan
 

Not to labour the point:)

With a subject at 5ft, the 162mm diameter front element of a Optimo 24-290, “looks around” more than the same focal length of a Prime with a 50mm diameter front optic.

This is an extreme example as most tests don’t mix primes and zooms.

Mike Brennan
DP
Melbourne

Thomas Gleeson
 

Argyris,

Can you repost your image as I cannot see it within my email client. Maybe another format? I would be happy to include it in my testing. Presently the plan is to use the Monstro camera and window it to S35 and then switch to VistaVision for comparison. Hopefully we will have a set of Leica Thalia’s to use. The thinking is to use the same camera/glass combination as to remove variables just leaving the sensor size.

Stuart Brereton wrote "I’m not convinced that they do. This admittedly crude, test shows very little difference in focus separation between the formats, and what there is could easily be attributed to variables in lens design, and actual aperture, as opposed to indicated.

Stuart I am confused by this did you not say the Full Frame camera had a 50mm lens @f4 and the APS-C camera 35mm lens @f2.5 Clearly the Full Frame camera had more focus separation? Am I misinterpreting?

Stuart Brereton wrote  "I think it’s a mistake to include film originated movies, as no-one is denying that having a larger neg is a benefit when shooting with film. It’s whether ‘large’ format digital origination is a benefit is the question here."

I will qualify I was not a party in any of the discussions that took place on any of these films but I would question wether any film originated project is concerned enough about grain to move to a Large format. This certainly was the case in the past with older grainier stocks and lossy generational printing but modern Vision 3 stocks going straight to a scan? If your willing to shoot with deeper stops required for a reasonable DOF on 65mm you may as well just shoot 200T or even 100T or 50D on S35. I have no doubt some productions shoot big format 'cause they can and they have the budget but is there more? I don’t know if there is so I am arguing with no one but I would love to get a clear understanding. So if you have thoughts on tests to prove or disprove or just to illustrate or even illuminate throw them the way of the ACS and we try out best to make it happen.

Tom Gleeson
Sydney Based DOP

Martyn James Bull
 

At the foot of all email messages is the link to the whole conversation on the groups.io website. Use that and you can see all the images properly in a web browser if having difficulty with email. 

On 3 Dec 2017, at 10:57, Thomas Gleeson <lensboy235@...> wrote:

Groups.io Links:

You receive all messages sent to this group.

View/Reply Online (#270) | Reply To Group | Mute This Topic | New Topic


--
Martyn James Bull
Filmmaker
Oxfordshire, UK

martyn@...
+44 7528 522137

Scott Dorsey
 

By "looking round" are you talking about subtle barrel distortion here?
--scott

Scott Dorsey
Kludge Audio
Williamsburg, VA.

Argyris_Theos_cml
 

Hello Tom,

One such sphere can be seen here
It is just a sphere like earth with meridians and parallels but without the continents. 
The idea is that the relative distances between "meridian & parallel crossing points" changes in relation to the exhibited roundness. 
Best

Argyris Theos, gsc 
DoP, Athens Greece,
+306944725315
Skype Argyris.Theos
via iPhone

3 Δεκ 2017, 12:57 μ.μ., ο/η "Thomas Gleeson" <lensboy235@...> έγραψε:

Argyris,

Can you repost your image as I cannot see it within my email client. 

Michael Brennan
 

The contents of this email are confidential
On 3 Dec 2017, at 11:38 pm, Scott Dorsey <kludge@...> wrote:

By "looking round" are you talking about subtle barrel distortion here?
--scott
No, I’m referring to the effect that others describe as roundness of the subject.

Could also call it 3 dimensionality of the subject.
For instance a mid shot where the face, in the fore to aft plane, appears to have more depth, or more shape, “rounded”.
At least that is my interpretation of what others are seeing:)

It is the difference of using a small probe lens versus a lens with a front element that is physically, far larger in diameter.

Most noticeable when a subject is close to the lens.

A Forced perspective (?) is the end result, the by-product is the subject appears less round.

I’m not aware of a technical term in common use that describes the optical physics.

Put another way, the ants-eye perspective is achieved by using a probe, endoscope or pinhole size lens.

It’s obvious when shooting miniatures and I suggest that it could be very, very subtly in play with every lens test where the subject is within a certain distance (a few meters) from camera and shot with lenses with different size front elements.


Could this be the cause of the very minor difference in roundness described by others? Therefore, not directly related to the size of the sensor?


Could also have a false memory of this effect:)

In any event, lacking the technical vocabulary to express it!

Mike Brennan
DP
Melbourne

Argyris_Theos_cml
 

Scott
No he means something totally different
e.g.
take a zoom with a big front element, add lots of diopters, zoom in (aka no "wide lens distortion") and shoot macro. The closer an object the weirdest it appears. This is caused by the fact that the front element is bigger than the object itself. So the far right of the lens sees a little bit "behind" the object, just as does the far left of the lens. This is the "looking around" and it exists even if we are not shooting macro, but in standard shooting it is so subtle, almost invisible.
Think about it, it is rational.
A.

Argyris Theos, gsc
DoP, Athens Greece,
theos@...
+306944725315
Skype Argyris.Theos
www.vimeo.com/argyristheos
via iPhone

3 Δεκ 2017, 2:38 μ.μ., ο/η "Scott Dorsey" <kludge@...> έγραψε:

By "looking round" are you talking about subtle barrel distortion here?
--scott

Stuart Brereton
 


On Dec 3, 2017, at 2:57 AM, Thomas Gleeson <lensboy235@...> wrote:

Stuart I am confused by this did you not say the Full Frame camera had a 50mm lens @f4 and the APS-C camera 35mm lens @f2.5 Clearly the Full Frame camera had more focus separation? Am I misinterpreting?

Sorry Tom,

I think I misunderstood what you were saying. Yes, a larger format will have less DoF at the same stop as a s35 imager. It’s roughly 1 1/3 stops difference between s35mm and FF.

Stuart Brereton
DP, LA

Art Adams
 

I wanted to try this very thing on a Venice recently, but the one that showed up in San Francisco was not ready for full frame prime time. I'm told it's returning in January, when it should be full frame capable.

The test should be very simple:

Frame two people, one in a medium shot and one medium wide or wide.

Put two C-stands just inside the edges of the frame, on either side.

Run through each frame size variation without moving the camera, keeping the C-stands at the same positions at the edges of the frame, preferably resizing with a zoom. (I think there's only one zoom that might be able to do this, the new Angenieux 25-250.)

--
Art Adams
Director of Photography
San Francisco Bay Area

Jessica Gallant
 

(The following is super nerdy and outside the realm of traditional cinematography. Feel free to disregard it
unless you like esoteric discussions.)

I have an undergraduate degree in perceptual & physiological psychology and this is the kind of research
I was both trained to do and did on a limited level.

Ideally, in any experiment or test, you’d remove all the variables from the test except for one. Roundness
is something that’s been mentioned by more than one person plus it’s a variable that can be easily tested.
You don’t want any other variables in the test (unless you’re testing one variable’s effect on another variable
which is getting way ahead of ourselves here), so placing the object in limbo or with a neutral background
eliminates the effect of depth of field and the size of the background.

There are a lot of other ways you can test this, but this particular way jumped out at me, maybe because
I perceived the lens on the camera in the FF shot as looking ever so slightly less flat and rounder than the
same lens on the same camera in the S35mm shot.

I suspect there is something akin to a optical illusion going on here in as much as the brain actively constructs
reality (in this case, what we’re seeing) based on limited information, subtle cues, input from other senses, and
expectations based on previous information.

Optical illusions generally occur when the brain is given ambiguious or misleading information (the famous
Ames illusion where the room constructed so that perspective is misleading so people of the same size appear
to be different sizes) and misconstructs reality based in misleading depth cues…

https://youtu.be/gJhyu6nlGt8

… or is asked to process information in a way different than it was intended to be processed such as the Hollow
Face illusion the inability to see an inverted face is hardwired into the brain

https://youtu.be/Rc6LRxjqzkA

BTW, I’m not saying FF looking “rounder” than S35mm qualifies as an optical illusion. However, I do suspect there
may be the same or similar mechanism of action occurring. The FF image may contain some subtle cues absent
or minimized in the S35mm image, and the brain can use these subtle cues to construct a more three dimensional
image.

Jessica Gallant
Director of Photography | Los Angeles | CA
http://jessicajgallant.com
http://wb.imdb.com/name/nm0002680/
cell: 818-645-2787
email: jessicajgallant@...
Skype: jessicajgallant


On Dec 2, 2017, at 8:25 PM, Stuart Bremerton <ssbrereton17@...> wrote:

What do you mean by too many variables?

Art Adams
 

>The FF image may contain some subtle cues absent or minimized in the S35mm image, and the brain can use these subtle cues to construct a more three dimensional image.

Two things:

I think you're on to something. And...

I think I was wrong—mislead by the missing Panavision videos.

Stuart and David, I think you're absolutely right: for two lenses with the same angle of view, perspective should match between the two. Perspective is, after all, a characteristic of angle of view combined with distance to subject. If the angle of view is the same, and the distance to the subject is the same, the only thing that should be different is depth of field.

I think one of the things that threw me was this image from Panavision's lens video, part 1:


The assumption here is that the lenses used are the same angle of view, but they likely aren't. For one thing, the spherical shot has a plant in it, and the patterns in the background lights are different. The charts are also different, which implies that the spherical test was shot at a different time. They don't say what format they used for the anamorphic test, and there are multiple options for 2:1 in PCam, so I compared the angle of view between an Alexa in 4:3 2.39:1 anamorphic mode to the Panavision DXL in spherical 8K mode, both with 50mm lenses:


A 50mm on both cameras would see roughly the same angle of view, and yet the spherical large format image above looks to be shot on a lens of twice the focal length, given that the light strands appear to be half the distance apart and twice as soft.

So... that's a bit deceptive, especially as Panavision didn't document (or at least communicate) the lens, format and distance choices for any of those shots.

That, combined with the statement about the next frame—that the larger format, in combination with the increased magnification of the longer focal length lens, brought the background closer— threw me:


I'm not sure how wide a lens this is. It doesn't look very wide. Still, it has a certain something that feels beyond the reach of S35. It might have to do with the shallow depth of field. Whatever it is, it feels almost like a painting out of the Hudson River or Luminous schools.

Lastly, the comparison between VistaVision and Technoscope was deceptive as the sense of increased dimensionally has a lot to do with the taller VistaVision frame, which no one will ever shoot on a DXL as 1.5:1 is a dead aspect ratio in everything but stills.

Still, there is something going on with the larger sensor size that affects dimensionality. I don't think it's so much "roundness" as it is the fact that the perspective characteristics of wide angle shots on longer lenses feels more natural, and that in turn makes close objects feel more real—because they are actually "rounder" based on slight distance-to-subject distortion rather than unrealistic S35 wide angle distortion. I pulled some frames from Lawrence of Arabia, found on Youtube so the quality is not so great, that show some of this characteristic in spite of the small image size and strong video compression:


This is clearly a fairly wide lens, possibly along the lines of a 25mm in 35mm film (so possibly a 50mm lens combined with 65mm film) and, while the angle of view is wider than normal, there's very little distortion on Peter O'Tooles' face. It does feel "rounder" to me simply because there is some distortion, but not so much that it feels unreal.

Here's a comparison to a classic close-but-wide Spielberg 27mm closeup:



This is a great looking closeup, but the distortion is really noticeable. Still, this frame has the desired emotional effect, possibly because of the abstraction caused by the distortion.


Depth of field plays a role as well. I recently shot some talking heads for a corporate project on an Epic Dragon, in 6K, on a 100mm Ultra Prime at 8' or so, which resulted in a closeup a little wider than this. We shot at T2, and just the face and eyes were in focus. (We were using an Interrotron so the person was looking directly at the camera.) The effect was stunning: because their face and eyes were sharp, but nothing else was, it felt as if their faces were literally pushing out of the screen.

The shot above doesn't have the "roundness" feel to me. The lens choice feels like it's on the longer side.


Once again, clearly a wide angle shot but not much distortion. The guy on the far left looks like he's pincushioning a bit, but I think that's a combination of stance and overeating along with maybe a light touch of lens distortion. Otherwise, this is a remarkably "flat" image with wide angle perspective, which I think is part of the large format magic. I feel like the guy in the foreground exhibits some of that large format "roundness."


This is a longer lens than the others, but it still feels like it's a wider-than-"normal" perspective lens. The foreground face does feel "round" to me due to a combination of the wide angle of the lens combined with the flatness of a longer focal length, resulting in less distortion. The lack of distortion seems to tell my brain that this is a more realistic image than if I saw obvious close-but-wide distortion on, say, a 40mm lens on an S35 sensor at the same distance.  Presumably the large format lens would be in the 75mm range when paired with 65mm film.

What's wonderful about these Lawrence of Arabia frames is that they are stunningly composed, with judicious use of forced perspective, but without unrealistic distortion. Because of this, every shot feels to me as if it is a painting.

To sum up, I think Panavision missed the mark on their videos in a few ways:

-They oversold the dimensionality of large format imaging by making an unfair comparison between two completely different formats with widely varying image characteristics (contrast and aspect ratio being the biggest offenders)

-They made statements about lens magnification in regard to perspective that don't appear to be accurate or, at the very least, aren't clear and/or well illustrated, and

-The spherical and anamorphic comparisons weren't done in a scientific manner that disclosed all the relevant information, such that we knew what we were looking at.

I get that this stuff is hard. It's really difficult to articulate things that fall into the artistic realm rather than the technical realm. I also get that not everyone knows how to do scientific comparisons. But... I'm surprised that a company like Panavision put these videos out. I certainly made some incorrect assumptions based on information in these videos.

If someone from Panavision would like my input in revising these videos, please do get in touch.

Meanwhile, I think the following video, which is still on the Panalab website, does a great job of showing this "wide but not distorted" feel that I equate with epic 65mm filmmaking:


--
Art Adams
Director of Photography
San Francisco Bay Area

Adam Wilt
 

Art says,

Run through each frame size variation without moving the camera, keeping the C-stands at the same positions at the edges of the frame.

I would add one thing: change the f-stop to match depth of field between the two shots: about 1.3 stops different, as Stuart says. The Depth of Field Equivalents calculator at http://www.cambridgeincolour.com/tutorials/digital-camera-sensor-size.htm can be handy for both lens and f-stop values.

Cheers,
Adam Wilt
technical services: consulting / coding / camerawork
Vancouver WA USA (no, not that Vancouver, the other one)

Argyris_Theos_cml
 

This is clearly a fairly wide lens, possibly along the lines of a 25mm in 35mm film (so possibly a 50mm lens combined with 65mm film) and, while the angle of view is wider than normal

Here's a comparison to a classic close-but-wide Spielberg 27mm closeup

No it is not

It is shot with a field of view equivalent to a 50mm in Academy 35mm film. This is obvious, despite the visual cues that make people think this is a wide shot.

Addfitionally Peter O’Tool carries much brighter make up and he is lit with Brutes. What you confuses you most is that you can see the “bounce from below” board in his eyes. This gives a bit of dimensionality

I think I will slide back to anoninimity now.

 

Argyris Theos, gsc

DoP

tel. +30 6944 725 315

skype: Argyris.Theos

www.vimeo.com/argyristheos

Previous Topic Next Topic