VistaVision and Monstro and cinematography?

[Tom Gleeson]

I am involved with the Technical Committee for the Australian Cinematographers Society and I am planning both a presentation and video looking into the nature of large sensor cinematography.

We already have the amazing Arri65, the imminent release of Sony’s Venice and now Red’s Monstro’s are rolling of the line. This heralds a revolution. From the dawn of cinematography and for over a century variations on the 35mm format have dominated. We have moved from an optical/mechanical world to ever increasingly powerful digital/electronic where camera “gates” can be whatever we want them to be and their growing bigger.

At the ACS we are most interested in what this means for the aesthetic. Apart from the obvious technical advantages I want to demonstrate and explore how the world photographs differently with large sensors.

We are looking for good and specific examples of cinematography that leverage big sensors to enhance the cinematography. We will also shoot some tests to demonstrate the changed perspective due to the larger FOV on large sensors. Possibly a test subject sitting in an environment with elements on multiple planes. We will photograph the subject using both 35mm and VistaVision sensors but at different focal lengths to maintain same FOV.

I also want to demonstrate the advantage of VistaVision when a very large FOV is required. VistaVision should not require as an extreme wide lens to capture the FOV.

Phil Holland has alluded to these questions around large sensor cinematography and I am very much looking to his input. I will admit that I have had little experience with large formats so presently I am looking at this issue only from an academic perspective. I would love to get input from others on what they see as the advantages of large format cinematography and I am equally interested in questions or even ideas for tests we could carry out. Here is your chance !

 

[Aaron Lochert]

We will also shoot some tests to demonstrate the changed perspective due to the larger FOV on large sensors. Possibly a test subject sitting in an environment with elements on multiple planes. We will photograph the subject using both 35mm and VistaVision sensors but at different focal lengths to maintain same FOV.

There's no difference in perspective. Steve Yedlin explains it better than I can. Scroll down to the "Angle of View" section:

http://www.yedlin.net/170504.html

 

[Phil Holland]

There's an additional "thing". I'm working on something to explain it and it's hard to explain. AOV is easy to match if you are just looking at side by sides, and certainly one thing is just how different focal lengths draw, but there's another allure to large format cinematography.

 

[Dan Kanes]

Destiny !

I mean... Density !

Optical design and imager size do have a lot to do with the way an image is drawn for our perception.

Certainly we can calculate angle of view mathematically, but the optical path is indeed important to perspective. In my mind perspective is more than just angle of view.

Anamorphic imaging for instance is special because it can allow a wide angle horizontal field of view on a lens with a telephoto focal length and otherwise telephoto properties.

Personally that’s one of my biggest excitements about Monstro use - the first Red sensor I’ll have access to that allows for full-height anamorphic imaging area.

 

[Tom Gleeson]

Aaron,

Thankyou very much for pointing at Steve’s amazing article. I’ll admit I struggle with the mathematical and the physics of optics and I will need to study this piece further. Steve seems to be demonstrating that there is no perspective change if you don’t move the camera but change the focal length to compensate for the changing sensor size. This will result in shallower focus but this too can be compensated by stopping down.

The perspective change would only come from moving a camera forward to maintain SUBJECT size as the sensor grows larger with the same focal length?

If there are other articles like Steve Yedlin out there please point me in their direction. Its a tough learning curve although taking delivery of a Monstro in a few weeks will help.

Phil your such a tease but I am looking forward to your thoughts. Can I guess the allure is the different motion cadence you have hinted at.

 

[PatrickFaith]

This is more from a financial side, but if audiences are going to "pay" they want an experience which isn't the same thing they get for free.
If the image doesn't translate to "high quality" they simply won't go to the theater, and they won't pay for something that doesn't gush quality on their phone.

I attached a shot from my monstro of me doing a basic focus test. To me this shot bleeds quality even when I'm just not doing anything. I'm not so sure I would focus on the monstro just being a large sensor. To me the "quality" can be judged by the audience when the entire chain is high quality: the sound is high quality, the sensor is large, the color science is perfect, the lens is high quality. I feel the audience can determine if the director/crew is devoted to quality. The monstro is an example of a camera that is devoted to quality, the same as the arri 65 is. The monstro I do believe is more accessible to medium budget films, so the ability to create compelling quality is something that medium budget films desperately need in order to have draw.

This is my test shot with monstro and a arri master prime anamorphic. Note I think it's also important for that quality to translate all the way from IMAX to cell phones, and the monstro with high quality lenses I think does a good job on that (i.e. creates a obvious quality picture that works on 720p all the way to 4k dcp, this picture is a highly compressed jpeg yet feels high quality).

 

[Tom Gleeson]

Dan,

Not being negative but I don't really follow your train of thought? How does "Density" and "optical path" fit into the discussion of larger sensors? Could you elaborate further? Although I have no trouble understanding and celebrating full height anamorphic coverage on the Monstro:drool5:


Patrick I agree the larger sensor is only one aspect of the Monstro camera* but I was planning on the Committee's focus being the question of "why larger sensors and how they affect the image." There are a lot of assumptions about that bigger is a different or better aesthetic but I have not seen an articulate argument of why this is so. I "feel" that it is a correct assumption but I want to better understand why.


*The Monstro will still be at the heart of our research as its the only VistaVision sized professional camera we can get our hands on. Sony has managed to give us only a few hours with the Venice and Arri have been very helpful but there are no Arri65s presently in Australia.

 

[Dan Kanes]

By density I mean that for a given scene the sampling density is different.

If we trace the same rays of the scene using an 8mm optic for a super 16 format, an 18mm optic for super 35 format, and to borrow the note from the Yedlin article a 50-52mm optic in imax format, and we dontmove the camera, the density of photons rendered onto the surface area is quite different. We can correct against these differences but the angle of incidence at which the rays strike the sensor or film plane is different as well as how they fill the photosites or interact photochemically with the negative. Film by nature has more “crosstalk” - digital crosstalk is often ugly. The larger the surface area of the photoreceptor the lower the density - which will lead to human perceptual difference in contrast - we can correct against this, but it will still be visible.

Depth of field will be different - which will offer a different perspective - angle and field of view are only part of perspective IMO.

I’m sure there are details I’m forgetting or missing here, I’m not someone who says “it’s not cinematic” unless the sensor is bigger and the depth of field is so shallow you can only see wisps of hair, but indeed it does offer new perspectives - there is a lot to explore here.

I personally like the multi aspect large sensor because we can use the full sensor for plates or special scenes and don’t have to use it all the time, and can switch down to 5k or 6K for day to day operations - we don’t have to use the higher pixel density of Helium.

Fun side story - I was a lighting technician on several of Steve Yedlin’s films early in his career - awesome to see how far he has come. I truly love what we do and I love the people who do what we do.

 

[Nick Morrison]

Having tested Monstro recently (with a full set of vintage glass ranging from 7.5mm fisheye all the way to 300mm focal length), I can confirm the leap from a 30mm sensor to a 41mm sensor is HUGE...literally.

Anyone who tries to say otherwise by using vageries like "K's don't matter" and "FOV is the same" is dancing around the practical, day in/day out implications of working with a sensor this large.

Remember how the industry FREAKED OUT in 2011 when RED launched the Epic with an MX sensor that was 27.7mm across (4mm wider than Alexa s35), and then got even more bent out of shape with Dragon's 30.7mm bump (7mm larger than Alexa s35)?

Now add another 10mm across.

Alexa at Super35 is 23.8mm

Monstro at full gate is 40.9mm.

When you sit back and realize Academy 35 is 21.95mm across, it becomes clear Monstro is basically two Academy negatives side by side.

Monstro's connection to Academy becomes even more interesting when you consider Monstro's height - exactly 21.6mm (almost an EXACT match to Academy's width).

Monstro's "doubling" of an Academy sensor becomes even more fascinating when you realize ANAMORPHIC for decades did a similar thing - a 2x anamorphic element would "double" the fov of given lens and squeeze it into an unmasked Academy frame (and in so doing, the 21.95mm x 18.6mm negative would get "doubled" into a 43.9mm x 18.6mm experience, if you will.)

MONSTRO = 40.9mm x 21.6mm
DESQUEEZED 4PERF FOV = 43.9mm x 18.6mm

Pretty close right?

Anamorphic doubled the Academy negative optically (at the expense of light loss, distortion, and other aberrations).

Monstro doubles Academy spherically, with no loss of light, no aberrations, and in fact the added K's in resolution help improve lowlight performance, boosts the ability to stabilize a shot in post, and adds a perception of smoothness (per Cioni).

It's as if David Fincher, who's hatred for Anamorphic is legendary, designed his DREAM SENSOR.

All the benefits of anamorphic (wider digital canvas), and none of the pitfalls.

Anyone who thinks a sensor this big is NOT a big deal must also then be non-plussed about anamorphic, for the reasons listed above.

It's as if Monstro, as designed, is a hyper conscious attempt at spherical anamorphic - if that makes any sense?

And it terms of lens uses, yes the implications are huge.

A 35mm is no longer your "normal" lens, a 50mm is. CU's are no longer on a 85mm, they are on a 135mm. Wides are no longer on a 21mm, they are on a 28mm or 35mm.

Super wides don't need a 16mm anymore - a 21mm or 25mm will do.

What does that mean? Well longer lenses in general use means all your footage will be flatter and more compact - for faces and people, this is often a great thing.

Wides will also have noticeably less distortion. A 35mm as a "wide" is HUGE.

In general, using longer lenses will give your images a more medium format compactness than we're used to, and the longer glass also brings more selective focus and shallow depth of field into play.

On set, what we noticed right away was a muscularity and grandeur to the image - less distortion, and a flatter FOV did wonders to human faces.

Also, on a practical level IN POST, being able to shoot 8K VV and easily pull 5K/Super 35 reframes out of numerous parts of this large canvas should be a major boon to many productions. We pull 4K/Academy reframes from 6K Dragon all the time. Monstro will add a whole other dimension to this experience.

Hope this helps.

 

[Patrick Tresch]

Well are you running behind the Canon 5D train?

:sarcasm:

 

[Timo Alatalkkari]

Comparison of how different angles affect the image when using the s35 sensor as a reference point.

http://www.whitepointoptics.com/#ts70 -> go to: Image Circle And Comparison

TIMO ALATALKKARI
Chief Optics Technician
Whitepoint Optics

 

[Peter Lyons Collister ASC]

I remember when Fincher and I worked at Propaganda Films (music videos and commercials) and we had a conference room called the Silo. It was decorated with George Hurrell Hollywood portrait photographic prints. Hurrell shot these with an 8X10 inch view camera. The focus fall off on the faces was so creamy and beautiful that many directors there wanted the same thing in their videos. It was the larger "sensor" that created this. Fincher even joined with John Schwartzman to shoot a Nike commercial in 65mm to start to create this falloff.
My point being is that Nolan doesn't shoot in IMAX for the resolution but for the "feeling" the larger format helps form.
So the math and physics of optics and circle of confusion exist but it is our visual perception of the result that matters most......

 

[PatrickFaith]

...
Anyone who thinks a sensor this big is NOT a big deal must also then be non-plussed about anamorphic, for the reasons listed above.

It's as if Monstro, as designed, is a hyper conscious attempt at spherical anamorphic - if that makes any sense?

And it terms of lens uses, yes the implications are huge.

A 35mm is no longer your "normal" lens, a 50mm is. CU's are no longer on a 85mm, they are on a 135mm. Wides are no longer on a 21mm, they are on a 28mm or 35mm.

Super wides don't need a 16mm anymore - a 21mm or 25mm will do.

What does that mean? Well longer lenses in general use means all your footage will be flatter and more compact - for faces and people, this is often a great thing.

Wides will also have noticeably less distortion. A 35mm as a "wide" is HUGE.

In general, using longer lenses will give your images a more medium format compactness than we're used to, and the longer glass also brings more selective focus and shallow depth of field into play.

On set, what we noticed right away was a muscularity and grandeur to the image - less distortion, and a flatter FOV did wonders to human faces.

Also, on a practical level IN POST, being able to shoot 8K VV and easily pull 5K/Super 35 reframes out of numerous parts of this large canvas should be a major boon to many productions. We pull 4K/Academy reframes from 6K Dragon all the time. Monstro will add a whole other dimension to this experience.

Hope this helps.

100%, a good example is my A camera is now a monstro while the B camera is still a dragon weapon 6k. When I switch out the A camera with the B camera at the same blocking point to maintain the same ratio of foreground to background field of view (so the background objects dont change size), I need to change the lens to by 25% wider and loose about 2 stops of depth of field (wide angle lenses I have atleast have significantly less depth of field at the same shooting distance as medium angle lenses). So if I want to keep my depth of field the same, I need to shoot my monstro with like a t4 50mm in order to switch it out with a t2 35mm. This also is the wrong direction in T stops for telling my story, often I want to isolate the foreground object from the background, so want a shallow depth of field with the background dropping off and feeling smaller. Another way of thinking about it is I'm actually gaining 2 stops of shallower depth of field with the monstro at the same shooting distance and field of view - HUGE.
[the concept of backing the camera up so the crops match is flawed, since it also changes perspective in the wrong way I need to go from a story point, most of the time]

There's some shots I just cant do with my 6k, for example if I want to do a medium-wide shot with the foreground object popping a bit out in ratio to the background, yet still have a lot of the background still in frame. This type of shot is real nice in that it establishing the viewers attention while still giving context. I think people will start talking about foreground/background depth compression a lot more after they shoot a while with monstro, and really notice a bunch of story compromises they where making with 35mm.

One thing I have noticed which doesn't matter as much on my A vs B camera's which is somewhat counter intuitive from at least the publicity of large frame, is that on establishing wide shots I really don't notice much difference between my A and B camera. If the depth of field is wide(foreground and background are all in focus), and if the foreground object is ok to be compressed to the background, that type of shot can be handled easily with lens choices on my B camera. This is good for me, in that I just can allocate my B camera on wide shots and then A monstro camera is allocated on everything else. Why I think this is counter intuitive is I use to hear all the time in the old days when I actually watched VV in the theaters, that VV large expanse establishing shots were the total rage in the publicity of that format. [i guess lens tech has passed that up, where the new wide angle lenses on smaller sensors I think do a great job if one is shooting for wide depth of field] When I see 16mm films that really work "story wise", I see that the story has the main character somewhat emotionally overwelmed by thier background, so the lens foreground/background compression of 16mm camera makes sense for that, so aligning the story to the creative technical choices don't mean in everycase wider sensors is better though.

 

[Dan Kanes]

My point being is that Nolan doesn't shoot in IMAX for the resolution but for the "feeling" the larger format helps form.
So the math and physics of optics and circle of confusion exist but it is our visual perception of the result that matters most......

Could not agree more with this.

I half-jokingly call what we are doing here “The Quantization of Human Visual Perception.”

It’s true. The feeling and perspective do change, and real optical math actually backs up this assertion.

 

[Michael Tiemann]

Comparison of how different angles affect the image when using the s35 sensor as a reference point.

http://www.whitepointoptics.com/#ts70 -> go to: Image Circle And Comparison

TIMO ALATALKKARI
Chief Optics Technician
Whitepoint Optics

Your RED VV and ARRI65 columns are swapped.

 

[Timo Alatalkkari]

Your RED VV and ARRI65 columns are swapped.

Ok, could you specify?

 

[Michael Tiemann]

You show 30mm as equiv to 18mm on VV and 15mm on A65. But A65 is larger than VV.

 

[richard peterson]

... shot with monstro and a arri master prime anamorphic.

Hey Patrick, Do the MPA cover the full height(21.6mm) of the monstro sensor ? What is the max anamorphic settings on the Monstro 8k ? Do the MPA cover ?

the published image diameter for the MPA is 29.26mm; since height of the sensor seems be the important spec in relation to anamorphic lenses I was wondering what options the MPA will give us with the monstro

 

[Timo Alatalkkari]

You show 30mm as equiv to 18mm on VV and 15mm on A65. But A65 is larger than VV.

Ok, I understand what you mean but it's a comparison to the angle so the A65 sensor is wider so the angle of view is larger. Since the angle of view is 2x higher to the A65 cell than the Super35 cell, it is equivalent to the WPO ts70 30mm 15mm angle to the A65 sensor compared to the s35 sensor.

 

[Nick Morrison]

Your RED VV and ARRI65 columns are swapped.

Was thinking the same thing - Alexa 65 has a larger sensor than Monstro, and should have a wider FOV