Jump to content


Photo

Difference between 8k x2 anamoprhic and 6k 4:3 Mode


  • Please log in to reply
25 replies to this topic

#1 Robert Daniel Martin

Robert Daniel Martin
  • Basic Members
  • PipPip
  • 20 posts
  • Director
  • New York

Posted 24 September 2018 - 02:41 PM

Referring to the RED DSMC2 MONSTRO camera

 

On there website I have noticed in the specs section they have an option of '2x ana' under the 8k resolution. What does this mean? What is the uncropped aspect ratio and this captures in camera for this 8k resolution setting?

https://www.red.com/...tity=1?sensor=1

 

 

Also see the highest resolution option that lists a 4:3 mode is at 6k for then setting it to 2x anamorphic squeeze.

(Note the 1 minute mark)

 

So what is the difference between these two options and what are the digital sensor dimensions for both?

 

Do either record the uncropped image at the exact dimensions of traditional 4 perf film at 24.89 x 18.66 with a diagonal 31.10???

 

I have noticed the Sony Venice is not actually an exact digital sensor equivalent at 4k.... it is 24.3 x 18.3

 

It'll be great to see the comparison of specs for this RED versus the Venice, since basically what I am trying to figure out is what camera (if any) grants the exact 4:3 digital equivalent when cropping in via the digital sensor, as well as which provides a highest resolution. So these are the two factors in play for my considerations...
I just want to make sure I am really capturing the true height and width field of view in the original film format, in camera of a digital sensor, when using x2 ana lenses. So whatever is the closest to 24.89 x 18.66 will be ideal

This all
does make me wonder why Sony and/or RED did not simply match the original format dimensions from the get go... since it's entirely possible to create that exact crop dimension when working down from a larger VV sensor, right?


Edited by Robert Daniel Martin, 24 September 2018 - 02:42 PM.

  • 0

#2 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 24 September 2018 - 05:32 PM

The pixel dimensions and the sensor size are not directly related -- you can have a 4K sensor that is the size of a Super-8 frame or the size of an IMAX frame, just depends on how large each photosite is, if there is scaling involved, etc.

 

Keep in mind also that 4-perf 35mm full aperture film is not what 4-perf 35mm anamorphic film uses, since it only needs a 1.20 : 1 area on the negative in order to capture a 2.40 : 1 image with a 2X optical squeeze -- and 4-perf 35mm full aperture is 1.33 : 1 (so the 2X squeeze would produce a 2.66 : 1 image unsqueezed if you didn't trim some of the width of the negative area.). A 24.89mm width includes the soundtrack area that anamorphic doesn't use.

 

If you actually recorded 24.89 x 18.66mm with a 2X anamorphic lens, your image would be 2.67 : 1 once unsqueezed, which is non-standard.

 

If you really want to match the area used for traditional 4-perf 35mm anamorphic, you're looking more at:

 

35mm anamorphic (contact) print projected area: 20.96mm x 17.53mm (2.39 : 1)

35mm anamorphic camera negative: 22mm x 18.59mm (2.37 : 1)

 

In other words, you could try and match your sensor area used to what a 4-perf 35mm camera would use for anamorphic, but what was seen in a theater was trimmed from that, so in some sense, what really matters is matching the area actually seen by people in the final product, which would be 20.96mm x 17.53mm.

 

Most digital camera except the Alexa have a widescreen sensor, so the limitations are generally not getting something 22mm wide or so, but getting a sensor that is between 17.65mm and 18.59mm in height.  Hence what these new "full-frame" digital cinema cameras are being used for anamorphic even though they waste a lot of the width of the sensor shooting anamorphic.

 

You can always record a slightly bigger area as long as you have frame lines for the actual area you plan to use.


  • 0

#3 Robert Daniel Martin

Robert Daniel Martin
  • Basic Members
  • PipPip
  • 20 posts
  • Director
  • New York

Posted 24 September 2018 - 06:32 PM

The pixel dimensions and the sensor size are not directly related -- you can have a 4K sensor that is the size of a Super-8 frame or the size of an IMAX frame, just depends on how large each photosite is, if there is scaling involved, etc.

 

Keep in mind also that 4-perf 35mm full aperture film is not what 4-perf 35mm anamorphic film uses, since it only needs a 1.20 : 1 area on the negative in order to capture a 2.40 : 1 image with a 2X optical squeeze -- and 4-perf 35mm full aperture is 1.33 : 1 (so the 2X squeeze would produce a 2.66 : 1 image unsqueezed if you didn't trim some of the width of the negative area.). A 24.89mm width includes the soundtrack area that anamorphic doesn't use.

 

If you actually recorded 24.89 x 18.66mm with a 2X anamorphic lens, your image would be 2.67 : 1 once unsqueezed, which is non-standard.

 

If you really want to match the area used for traditional 4-perf 35mm anamorphic, you're looking more at:

 

35mm anamorphic (contact) print projected area: 20.96mm x 17.53mm (2.39 : 1)

35mm anamorphic camera negative: 22mm x 18.59mm (2.37 : 1)

 

In other words, you could try and match your sensor area used to what a 4-perf 35mm camera would use for anamorphic, but what was seen in a theater was trimmed from that, so in some sense, what really matters is matching the area actually seen by people in the final product, which would be 20.96mm x 17.53mm.

David- great to hear from you! hmm i'm going to have to digest all that further, but nonetheless, what I affirmatively want to do is capture 2.66 in camera because I want to frame for 2.55 and then crop off the sides and retain the full height in post to present the image as such.
So basically what I believe they did in regards to 'Presented in Cinemascope' for La La Land... which was in fact shot on 35mm film


I don't know of a digital film camera which has the native aspect ratio to capture 2.55:1 in camera... if there is one then do let me know. I certainly will not use the 6:5 anamorphic aspect ratio for a precropped image option like what is available in the Sony Venice.  No 2 perf shooting for me!

For reference what is the uncropped aspect ratio that you are referring to of 22mm x 18.59mm and 20.96mm x 17.53mm?

what would it be for 2.55:1?

 

 

Regardless, what does the 'and Ana 2x' in 8k option refer to on the RED DSMC2 Monstro when it comes to the aspect ratio of the digital negative in camera?


  • 0

#4 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 24 September 2018 - 08:31 PM

"La La Land" basically exposed Full Aperture width (Super-35), lens centered for full aperture, rather than the sound aperture area for standard 1.85 and anamorphic where the lens is offset because the left side of the image will be covered by the optical soundtrack on the print.  So "La La Land" had to go through a D.I., it couldn't have made a contact print with a soundtrack on it, and besides, it was released digitally anyway.

 

Anamorphic / CinemaScope came out in 1953 as a response to 3-camera Cinerama, which was projected using three full aperture 6-perf 35mm frames side by side and the sound came from an interlocked 35mm full coat mag roll with seven discreet tracks of sound.

 

So the original idea for CinemaScope was to shoot and project 4-perf 35mm full aperture (1.33 : 1) with the sound on a separate roll run in interlock.  So with a 2X squeeze/unsqueeze, the aspect ratio would have been 2.66 : 1.  But that idea was ditched soon for getting Kodak to make a 35mm print stock with smaller sprocket holes (called CS perf) to make room for four tiny mag stripes, two on each side of the image on the print.  This shaved the width of the projected image from 2.66 to 2.55.  Early CinemaScope movies like "The Robe", "How to Marry a Millionaire", and "Twenty Thousand Leagues Under the Sea" are 2.55 with the image centered between the perf rows (which is why when you see some prints projected today, the image is slightly misframed side-to-side.)

 

After a few years, CS perf prints were abandoned and standard perf prints were made with standard optical tracks on the left side of the frame, shifting the center of the image to the right (just as with Academy and 1.85).  This shaved the width of the image from 2.55 to 2.35.

 

I think the ANA mode in a Red camera records a 1.20 : 1 (6:5) sensor area so with a 2X optical squeeze, the image aspect ratio is 2.40 : 1 unsqueezed.  If you want a 2.66 image, then you'd want to record more of the sensor area.  The advantage of using the 6:5 option is that besides recording only the amount of data you need for a 2.40 image, it can do an unsqueeze for the monitor out and viewfinder, though I think nowadays there are external options for unsqueezing the image, and you can create custom frame lines for 2.66 : 1.


  • 0

#5 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 24 September 2018 - 08:34 PM

It's easy to calculate either sensor area or pixel ratio for 2.55 with a 2X squeeze. Just divide 2.55 by 2 and you get 1.275.  So the area used for 2.55 : 1 would be 1.275 : 1.  So if you know one dimension -- width for example -- you can calculate the height needed, or vice-versa.


  • 0

#6 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 24 September 2018 - 08:38 PM

Anamorphic (scope) gate dimensions were modified in the early 1970's to thicken the frame lines in order to hide splices better.

 

If you want to learn more about the early CinemaScope gate dimensions, you can look here but you'll have to convert the measurements to metric:

http://www.brianprit...c Sound Prints)

 

But I wouldn't get hung up over a few fractions of a millimeter. As long as your image uses between 17.5mm to 18.5mm height of the sensor for your 2.55 : 1 2X anamorphic image, you'll be in the same field of view range as anamorphic lenses had back in the day.


  • 0

#7 Robert Daniel Martin

Robert Daniel Martin
  • Basic Members
  • PipPip
  • 20 posts
  • Director
  • New York

Posted 24 September 2018 - 09:11 PM

So the original idea for CinemaScope was to shoot and project 4-perf 35mm full aperture (1.33 : 1) with the sound on a separate roll run in interlock.  So with a 2X squeeze/unsqueeze, the aspect ratio would have been 2.66 : 1.  But that idea was ditched soon for getting Kodak to make a 35mm print stock with smaller sprocket holes (called CS perf) to make room for four tiny mag stripes, two on each side of the image on the print.  This shaved the width of the projected image from 2.66 to 2.55.  Early CinemaScope movies like "The Robe", "How to Marry a Millionaire", and "Twenty Thousand Leagues Under the Sea" are 2.55 with the image centered between the perf rows (which is why when you see some prints projected today, the image is slightly misframed side-to-side.)

Yes, for me 2.55:1 has always felt the most balanced- and of course, most epic :)

hmm so in other words, right now in the digital cinema realm to attain the highest resolution and capture the uncropped 2.66 for a 2.55:1 presentation it would be:

a) RED DSMC2 MONSTRO at 6K in 4:3 with x2 sqeeze

B) Sony Venice at 4K in 4:3 with the x2 squeeze

c) Alexa mini at 2.8K in 4:3 mode

 

- d) Alexa LF???? perhaps-- sort of confusing at this point which sensor type and which resolution would result.

 

- And of course the Panavision DXL2 comes in the either rental flavor of the Venice or RED as well.... which either of these two options seem to be the best choices at the moment ~panavized digital cinema

 

 

Any other options?


  • 0

#8 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 24 September 2018 - 10:02 PM

You got the main choices, there are some other cameras with full-frame sensors, like the Canon C700 and ME20F-SH, some of the Phantom high-speed cameras have a tall squarish sensor, over 17.5mm in height.  Sony makes some cheaper video cameras with full-frame sensors.


  • 0

#9 Robert Daniel Martin

Robert Daniel Martin
  • Basic Members
  • PipPip
  • 20 posts
  • Director
  • New York

Posted 26 September 2018 - 10:04 PM

You got the main choices, there are some other cameras with full-frame sensors, like the Canon C700 and ME20F-SH, some of the Phantom high-speed cameras have a tall squarish sensor, over 17.5mm in height.  Sony makes some cheaper video cameras with full-frame sensors.

David-- I'm glad you've jumped on this thread. While I've got you here, I am very curious as to what tool you would use to go location scouting for a shoot in 2.55:1

What I mean is, what camera or tool is best for a director to go location scouting and be able to take stills and video clips of shots and compositions in this anamorphic aspect ratio? Doing so, I can create a lookbook with material in this aspect ratio ahead of time.

 

Using Magic lantern on my Canon 5Diii I have uploaded and use a 2.55 custom crop mark overlay when shooting stills and video in 16x9 mode. However, I then have to now create a workflow in photoshop to add a 2.55 letterbox atop all these stills to recreate the look in-camera for the final images.
 

I'm shooting with spherical full frame lenses and using the equivalent focal lengths to get the same HFOV of the anamorphic counterpart

(so I use a 12mm flat lens to represent the 25mm ana scope lens and so on) but I know shooting stills and video in this 16x9 mode, even on a physically larger sensor of 3:2, does not allow the camera to see the same VFOV height that a 4:3 sensor provides when shooting anamorphic. So I would like to be able to get that in camera and compose for it.

 

Yet using my same workflow, even if I shot in the 4:3 crop mode on the 5diii or even on the native 4:3 mode of a medium format Fuji camera, or something-- adding a letterbox atop or even simply cropping them to that ratio will be cutting off the height of the top and bottom (of which I want to retain) to great a widescreen image perspective out of a 1.33:1 image size.

 

So I suppose what I am asking-when a director or DP goes into the field way ahead of time for scouting, and wants to be able too see that true 4:3 height of anamorphic images ~ the uncropped 2.66 field of view -- What tool should I use to do this?


  • 0

#10 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 26 September 2018 - 10:42 PM

I’m not sure you understand field of view equivalence issues if you think you can’t put a lens on a full-frame camera without cropping into the view that an anamorphic lens would see. Just do the math. Why would you crop a full-frame spherical image to 4:3 first before then cropping it again to 2.66?

What do you mean that a 4:3 sensor shooting anamorphic has a bigger vertical field of view than what you can get on a full-frame camera shooting spherical? 35mm anamorphic is 17.5mm tall and full-frame still sensors are near 24mm tall — and the anamorphic squeeze only works on the horizontal plane. A 50mm anamorphic lens and a 50mm spherical lens have the same vertical field of view.
  • 0

#11 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 26 September 2018 - 11:08 PM

The early spec for CinemaScope projection said .715" for the aperture height, which is 18.16mm tall.  So for a 2.66 : 1 image with a 2X squeeze, thus a 1.33 : 1 gate, that would be 24.1528mm x 18.6mm.

 

Since the image has a 2X horizontal increase in field of view, you can pretend that the camera format therefore is a virtual 48.3056mm x 18.16mm.  In other words, instead of thinking of your 50mm anamorphic as a 25mm spherical, just think of it as a 50mm spherical lens on a bigger format similar to 70mm film.

 

So let's say you are limited to shooting stills on a Canon 5D Mk4 with a sensor that is 36mm x 24mm.  The crop factor between 48.3056 and 36 is 1.3418222X.

 

So a 50mm anamorphic lens on the camera with the 18.6mm tall format and the virtual 48.3056mm width (basically you are imagining a 50mm spherical on that larger size) would require you use a 37.26mm lens on a 36mm x 24mm wide sensor.  And you'd have to crop your image to 2.66, so the effective sensor area used would be 36mm x 13.533834mm.

 

Your full-frame Canon 5D still would be cropped to 6720 x 2526 pixels for a 2.66 : 1 image.

 

But the field of view of the 50mm anamorphic on the 35mm 4-perf format at 24.1528mm x 18.16mm would be the same as a 37.26mm spherical lens on a full-frame camera cropped vertically to 2.66 : 1.  Doesn't matter that your anamorphic format uses an 18.16mm height and your full-frame camera sensor is cropped to 13.533834mm if the view is the same.

 

Otherwise, I suggest carrying the real anamorphic lenses around and either getting a full-frame camera adapted to take them or using an Artemis Prime viewfinder set-up on an iPad Pro that has the adaptor to use the same lens mount. See:

http://www.chemicalw...rtemisPrime.php


  • 0

#12 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 26 September 2018 - 11:10 PM

Now if you have to shoot video on your DSLR, that's another issue, you just need to find out the sensor area used for 16x9 HD recording (or 4K if that's an option) to calculate your crop factors.  But you can always figure out the equivalent view both vertically and horizontally.


  • 0

#13 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 26 September 2018 - 11:21 PM

You could also just use the regular Artemis app on your iPhone and compensate mentally for the fact that 2.66 : 1 is 1.108333X bigger than 2.40 : 1.  So if you set your Artemis up to simulate the view of a 50mm anamorphic lens for a 2.40 image, assuming no custom aspect ratios are possible, you figure that you'd have to use a 55.416mm lens in 2.66 to get the same horizontal view, or you'd have to input a custom focal length of 45.112795mm into the Artemis for 2.40 to see what a 50mm lens would see in 2.66 (of course your vertical views would be off).

 

Not sure why you can't scout though with finders designed for 2.40 and just imagine that at 2.66, you're just going to see 1.108333X more horizontal picture.


  • 0

#14 Robert Daniel Martin

Robert Daniel Martin
  • Basic Members
  • PipPip
  • 20 posts
  • Director
  • New York

Posted 28 September 2018 - 02:08 PM

Been studying your posts -- please point out any false statements or contradictions below so I can get this squared away. I'm just going to read downwards to build a conclusion.

35mm anamorphic is 17.5mm tall and full-frame still sensors are near 24mm tall — and the anamorphic squeeze only works on the horizontal plane. A 50mm anamorphic lens and a 50mm spherical lens have the same vertical field of view.

 

Regarding the VFOV height perspective of an image on traditional 35mm negatives: I always thought you do capture a taller height/VFOV in camera when shooting with scope lenses because instead of 3 perf s35 or 2 perf techniscope, the camera is now printing on the taller full 4 perf aperture of the negative? So therefore, inherently a normal 50 spherical lens would not be capturing/printing a taller image since these flat lens are not used in conjunction with the silent aperture 4 perf 4:3 mode
-- So in other words, when we watch a film in 1.85 or even 1.78, their VFOV, though a taller looking perspective then 2.55 / 2.40 / 2.35 films fit into a 16x9 container, are not actually as tall, since the Field of View captured with that 50mm anamoprhic lens is printing onto an additional 4th perf and therefore taller negative. Please confirm?

 

Now- when working on a physically larger (wider and taller) negative of a FFVV 3:2 camera (like a canon 5D or the Cine cameras mentions above):
I now understand that either the spherical or anamorphic 50mm lens would and does capture/print the same VFOV height, even taller to begin with, because 24mm is obviously taller then the 17.5mm of 35mm negs. Please confirm.
 

It is worth noting that the sensor does crop in (doesn't use the full available height of the sensor) when turned to 16x9 mode for still photos.... I wonder what this mm height actually is then in this mode, since it could help when exactly approximating what new height the camera is working at so as for us to compare that to the 17.5mm film standard

16x9 stills are 5760 × 3240 while the 3x2 stills are at 5760 × 3840

I don't know what the math equation for how to figure this crop factor out. Please advise

 

 

The early spec for CinemaScope projection said .715" for the aperture height, which is 18.16mm tall.  So for a 2.66 : 1 image with a 2X squeeze, thus a 1.33 : 1 gate, that would be 24.1528mm x 18.6mm.

 

Ohhh-- So then lets figure that crop factor above with 18.6mm instead then, yeah??
So I am presuming the 18.6 is the uncropped 2.66:1? They cropped the sides (not the height) to get to the original cinemascope standard of 2.55 and I presume are still keep cropping off the sides to get to the new standard of 2.40 and 2.35?? ~you don't crop height when having shot with anamorphic-- this only occurs when shooting with spherical to attain the same perspective....
So apologies, but where did 17.5mm come from? Is that the uncropped 3 perf height of 1.85 or is the 2.40 / 2.35 or something else??

 

 

Since the image has a 2X horizontal increase in field of view, you can pretend that the camera format therefore is a virtual 48.3056mm x 18.16mm.  In other words, instead of thinking of your 50mm anamorphic as a 25mm spherical, just think of it as a 50mm spherical lens on a bigger format similar to 70mm film.

 

A virtual 48.3056mm x 18.16mm --- aaa yes, an 8:3 image. Dearly wish it were easy and this crop factor was natively available on a camera to start with: Medium format cameras like the Fuji GFX50s are just short on the horizontal end at 43.8mm x 32.9mm... but then of course the height is even larger. This camera does have a 65:24 mode which is 2.71  -- this the closest I've found to my 2.66 uncropped ~2.55 ideal....but for an $5000 upgrade and a different crop factor arithmetic now to compare to 4 perf 35mm it is just getting to be an alternative hassle.

 

So as for my question above regarding the 3:2 to a 16x9 crop on the FF canon 5D (and while using FF lenses for that matter): the VeriticalFOV is a mystery to me. Again, I do understand that normal cinema 35mm lenses (either flat or scope) when put on a larger sensor of FFVV, will provide the exact same height) but the fact I am using FF lenses in the first place, is now giving me an arithmetic blind spot. 

 

So let's say you are limited to shooting stills on a Canon 5D Mk4 with a sensor that is 36mm x 24mm.  The crop factor between 48.3056 and 36 is 1.3418222X.

 

So a 50mm anamorphic lens on the camera with the 18.6mm tall format and the virtual 48.3056mm width (basically you are imagining a 50mm spherical on that larger size) would require you use a 37.26mm lens on a 36mm x 24mm wide sensor.

 

Okay I gotcha -- I definitely had my focal length equivalents incorrect since i'm on a larger shooting format to start with.
So here is the breakdown if I were to get a set of say...Cooke xtal express anamorphics for the film, and I wanted to get used to those focal length horizontalFOV,  but while using a FFVV still camera (the canon 5D3), with FF spherical lenses---

25mm Cooke Anamorphic -->18mm FF spherical

35mm --> 26mm

40mm --> 30mm

50mm --> 37mm

75mm --> 55mm

100mm --> 75mm

200mm --> 150mm

 

Mmmm so this is why they make x1.33 anamorphic adapters for the dslr world. I've only found one x2 adapter out there but I'm stumped when finding any, regardless of 1.33 or x2, that fit my front thread size lenses of 82mm.... I use a tamron 24-70 and various zeiss ZE primes

https://www.bhphotov...angefinder.html

https://www.bhphotov...ic_adapter.html

-- Lets go down a quick rabbit hole regarding tools for me to directly fix this problem but still usingmy same FF camera: I've noticed that these adapters, when used on a 16x9 mode turn the image into 2.35 when it's a x1.33 type-- This would be useless for my intentions because that doesn't capture wide enough HFOV.... or is it a tall enough VFOVfor my ideal framing of 2.55??? Or both?
 

The top one, though x2 squeeze for using on 4:3 mode, it doesn't fit onto the front of my lenses...it states: "It mounts to lenses with a non-rotating front and a front thread diameter of 62mm. SLR Magic recommends that the Anamorphot-50 2.0x be used with lenses that have a front element that is physically no larger than 50mm in diameter." Also, "Supports focal lengths between 35 and 100mm for APS-C and MFT format sensors, and 70 to 100mm for full-frame format sensors"
--- only being able to use my 70mm FF lenses and above seem pointless because it hardly covers my Cooke rental package range... So... what a waste.
And this one isn't compatible to fit on my lenses anyway.
 

For the bottom choice: the specs actually state the 82 is a rear thread and that it's only up to a 65mm front thread.... it was designed for the fuji zooms lenses... but those are for apsC format lenses....

--Goodness maybe I should just buy a  ApsC camera??... but ehhh that sensor's negative is 23.5mm×15.6mm.... So its's 3 perf s35 and fundamentally unable to capture the same aspects of the bigger 24.1528mm x 18.6mm 4 perf film format used when shooting anamorphic....
So now we see this is my current conundrum: even if I try to do it natively on a smaller format DSLR and get an adapter-- it's still not quite the same and then those adapters are on apsC lenses, they can't support the widest anamorphic focal length of 25mm anyway.... they top out at 35mm. I'm a wide angle user-- and if i can't quite get to the widest of 25mm equivalent on apsC or even close to that on FF, then what's the point.

 

 

And you'd have to crop your image to 2.66, so the effective sensor area used would be 36mm x 13.533834mm.

 

Your full-frame Canon 5D still would be cropped to 6720 x 2526 pixels for a 2.66 : 1 image.

 

But the field of view of the 50mm anamorphic on the 35mm 4-perf format at 24.1528mm x 18.16mm would be the same as a 37.26mm spherical lens on a full-frame camera cropped vertically to 2.66 : 1.  Doesn't matter that your anamorphic format uses an 18.16mm height and your full-frame camera sensor is cropped to 13.533834mm if the view is the same.

ummmm I think my ideal solution and the answer to all the confusion I seek, is right here, but it needs to be unpacked a bit further with your help.

 

Aaaa for starters, lets stick with the 5Dmark3 at 2.55 which does 16x9 stills at a lesser resolution of 5760 × 3240.  And lets avoid cropping the top/bottom in post to a further 2.66  -- At 2.55 it would be 5760 × 2260

So then what would be the the effective sensor area used for 2.55 on the FF?

So you're confirming that when I add a 2.55 letterbox atop this still photo or simply crop its top and bottom off to 2.55 outright, when using a 16x9 crop mode on a 3:2 sensor, it will retain/output the exact same height of a 35mm 4 perf neg??
Do please confirm-- because if so, my workflow is giving me what I want to accomplish in the realm of my virtual larger sensor method approach.


  • 0

#15 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 28 September 2018 - 07:33 PM

If your still camera is recording a 16x9 (1.78 : 1) image and you want a 2.55 : 1 image, you have no choice but to crop in post, unless you find a still camera anamorphic lens that perfectly squeezes a 2.55 image onto a 1.78 area (approx. a 1.4X squeeze needed).


  • 0

#16 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 28 September 2018 - 07:51 PM

Here's the history:

 

1953

4-perf 35mm 2X anamorphic with smaller CS perfs to make room for magnetic stripes on each side of the frame:

23.16mm x 18.16mm (2.551 : 1)

 

1957

4-perf 35mm 2X anamorphic with normal perfs and optical soundtrack on the left side of the frame:

21.31mm x 18.16mm (2.346863 : 1)

 

1970

4-perf 35mm 2X anamorphic with shortened gate height to hide splices better:

21.31mm x 17.78mm (2.3971438 : 1)

 

1993

4-perf 35mm 2X anamorphic with gate changed to standardize all sound formats to being 20.96mm wide:

20.96mm x 17.53mm (2.3913042 : 1).

 

2007

DCI specs for digital projection of scope images:

(2K) 2048×858

(4K) 4096×1716

(2.3869463 : 1)


  • 0

#17 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 28 September 2018 - 07:59 PM

Digital Photography Review states the specs of a Canon 5D MkIII as having max resolution of 5760 x 3840 and sensor size of 36mm x 24mm.  Assuming that 5780 x 3840 uses the entire sensor edge to edge, then 5780 pixels = 36mm wide.  So for a 2.55 : 1 aspect ratio using spherical lenses, divide 36 by 2.55, which is 14.12mm.

 

You don't need to match the height of 4-perf 35mm, you need to match the field of view by picking a focal length that gives you the equivalent view and by matching the aspect ratio.  You could match the height and width of 4-perf 35mm anamorphic back in the days of 2.55 : 1 by cropping a full-frame sensor to those dimensions but you'd need to use the same 2X anamorphic lens to match the field of view.  Since you are prepping / scouting with a spherical lens, then matching the height of a 4-perf 35mm anamorphic gate is irrelevant, you just need to match the field of view with the equivalent focal length and crop to match the aspect ratio.


  • 0

#18 David Mullen ASC

David Mullen ASC
  • Sustaining Members
  • 20300 posts
  • Cinematographer
  • Los Angeles

Posted 28 September 2018 - 08:01 PM

When I say that a 50mm anamorphic and 50mm spherical have the same vertical field of view, I'm assuming the same format size vertically.  Obviously the vertical view of a 50mm is trimmed when masking / matting further to a shorter height or using a camera that pulls down fewer perfs.


  • 0

#19 Robert Daniel Martin

Robert Daniel Martin
  • Basic Members
  • PipPip
  • 20 posts
  • Director
  • New York

Posted 29 September 2018 - 02:05 PM

One clarification:

Since 2.55 is not as wide as 2.66, and therefore as less HorizontalVOF (because the sides are cropped off a bit to get from 2.66 to 2.55),

 

I would no longer want to work with the uncropped squeeze factor of x1.3418222. Makes sense but then why would I squeeze more using x1.4 to find my new equivalent focal lengths to see the same field of view of a 2.55 ? Wouldn't it be opposite ~ less of a squeeze than 1.34?

 

Ex) 25mm anamorphic goes to 18.6mm spherical lens with a1.34 squeeze.... but then goes even wider to a 17.85mm focal length using 1.4 squeeze.

So the equivalent focal length would be not as wide to gain the 2.55 Hvof in this precropped methodology- correct?

 

What would be the proper new squeeze ratio?


  • 0

#20 Robert Daniel Martin

Robert Daniel Martin
  • Basic Members
  • PipPip
  • 20 posts
  • Director
  • New York

Posted 29 September 2018 - 03:08 PM

Also here is some math I did for this specific camera:
 

- Dividing the height of the pixels of the 16x9 mode by the pixel height of the taller 3:2 FF mode to find the crop factor the camera uses

3240 / 3840 = 0.84375 is the crop factor going from 3:2 to 16x9

- multiply this crop factor by the original 24mm height to get the sensor height the camera is using in this 1.78 16x9 mode

0.84375 X 24 = 20.25mm

 

20.25mm is still larger than 18.16mm in 4:3 4perf. This is a good start!

 

- Then now do the same to find what the sensor height is that you retain after adding the 2.55 letterbox mask in post by dividing the pixel height of 2.55 by the larger uncropped pixel height of 1.78 16x9

2260 / 3240 = 0.69753086 and then multiply by X 24 to get the mm height

0.69753086 X 24 = 16.74mm

~ So to use the 16x9 mode and then add a 2.55 letterbox mask in post (or just crop the stills outright) results in a shorter vertical sensor height than the original 18.16mm film format... in other words 36x16.74 is too far of a crop --going into a smaller format then the 23.16 x 18.16mm's height for starters which results in less Vfov captured in camera? Or Not?

 

Alternatively: If I shot simply 1.50 stills on the 3:2 mode and then added a letterbox or outright crop (after doing the same math steps) it would be 36x14.12mm to get 2.55

 

 

You could match the height and width of 4-perf 35mm anamorphic back in the days of 2.55 : 1 by cropping a full-frame sensor to those dimensions but you'd need to use the same 2X anamorphic lens to match the field of view.  Since you are prepping / scouting with a spherical lens, then matching the height of a 4-perf 35mm anamorphic gate is irrelevant, you just need to match the field of view with the equivalent focal length and crop to match the aspect ratio.

 

Aaa so to get the VerticalFOV is impossible using either a 3:2 or 16x9 mode spherical lenses but to attain the HorizontalFOV is entirely possible? Please clarify


Yes, my whole MO is to match both the width and height field of view.
I can get the width by using the proper crop factor multiplier (please provide) to discover correct spherical lenses equivalent. And in my case, I am not looking for a virtual 48.3056mm width -- it'd be a tad but shorter (what measurement exactly?) because I'd already be pre-cropping the width in camera on a 36mm sensor size.

In addition, are you saying that if I shoot stills in 3:2 mode or 16x9 mode, and add the appropriate 2.55 mask in post (to either one of these), they do retain the exact same vertical field of view height that is captured with anamorphic lenses on a 4:3 4 perf negative even with adding the mask?

In other words, 36 x 16.74mm (16x9 mode) and 36 x 14.12mm (3:2 mode) are the exact equivalent to the vertical field of view captured in camera as is attained in the 23.16 x 18.16mm original film format? So the the physical mm height of these two digital negatives are shorter then 18.16 but they capture the same Vfov just onto a smaller printing area.

Please confirm

 

And yes, I did just basically ask the same question in three different ways haha but I just want to make sure we're not mixing our vocab up with coming to an incorrect conclusion that the 2.55 perspective I am going to create by adding these letter boxes, will also result in the same VerticalFOV retained on screen for viewers, when perhaps it is not.

 

Sort like how we all know 2.40 masked spherical shot images are not the same Vfov as a 2.40 anamorphic shot images are when shrunk into a 16x9 container

Cool-- we're at the finish line


 


  • 0


Tai Audio

Wooden Camera

Willys Widgets

Paralinx LLC

FJS International, LLC

Metropolis Post

Abel Cine

Broadcast Solutions Inc

Gamma Ray Digital Inc

Media Blackout - Custom Cables and AKS

CineTape

New Pro Video - New and Used Equipment

rebotnix Technologies

Aerial Filmworks

The Slider

Ritter Battery

Visual Products

Rig Wheels Passport

CineLab

Technodolly

Glidecam

Visual Products

Tai Audio

Rig Wheels Passport

Abel Cine

New Pro Video - New and Used Equipment

FJS International, LLC

CineTape

Gamma Ray Digital Inc

Broadcast Solutions Inc

Willys Widgets

CineLab

rebotnix Technologies

Metropolis Post

Technodolly

Wooden Camera

The Slider

Paralinx LLC

Media Blackout - Custom Cables and AKS

Aerial Filmworks

Glidecam

Ritter Battery