Jump to content

Shooting anamorphic with digital cameras, which lens desqueeze factor is the right one?


davide sorasio

Recommended Posts

Hi everybody. I have a doubt concerning shooting anamorphic on digital cameras. I know the different camera menus (I'm talking from my Epic and Alexa experience) give me the option to chose between 1.3x, 2.0x and 2.0mag (this last one I have no idea of what it is) anamorphic desqueeze factor but I have no idea of what the rule of tumb is and which one to chose according to what.

Feel absolutely free to add anything else that you thing might make things clearer.

Thanks a lot,

Davide

 

 

Link to comment
Share on other sites

  • Premium Member

The majority of anamorphic lenses have a 2X horizontal squeeze, unless you rent the 1.3X Hawk anamorphic lenses, designed to squeeze 2.40 onto a 16x9 (1.78 : 1) sensor without cropping. Standard 2X squeeze means that for a final unsqueezed 2.40 image, you're using a 1.20 : 1 area of the sensor.

 

So besides knowing what lenses you are renting, you have to know the area of the sensor you are recording and you have to unsqueeze the image for the monitors and EVF, unless you prefer to look at a skinny image. The Epic has a 6:5 "ANA" mode for 2X anamorphic photography, basically just recording a 1.20 : 1 area of their sensor since that's all you need, saving you a little on the amount of data recorded. It also unsqueezes the image for the viewfinder and monitors in that mode. With an Alexa, you'd want to rent the 4x3 Alexas so you can use the taller sensor area. With the other cameras, which almost all record a 16x9 or even 1.9 : 1 area, you'll be cropping the sides to get down to a 1.20 : 1 area of the sensor, probably cropping in post.

  • Upvote 1
Link to comment
Share on other sites

6:5 is gonna be new 4:3 in future.

from ALEXA SXT Faq;

 

6:5? What the hell? I thought you needed 4:3 for anamorphic?

The 4:3 aspect ratio was originally left over from the film days, and now has been put to good use by VFX-heavy feature films. Anamorphic lenses don't actually need the full width of the 4:3 image; they squeeze the 2.39:1 image by a factor of 2, and the resulting area used on the sensor has a 1.195:1 aspect ratio, which is roughly 1.2:1, which is 6:5. When shooting with anamorphic lenses in 4:3 sensor mode, you always have to crop the extra area to the left and right of the image in post, an extra processing step that is avoided by shooting in 6:5.

In the ALEXA XT cameras with SUP 11 we have introduced this mode, however, there it was still called '4:3 Cropped', but that is the same thing as the ALEXA SXT/SXR 6:5 sensor mode

 

 

 

 

http://www.arri.com/camera/alexa/learn/alexa_sxtsxr_faq/

Link to comment
Share on other sites

  • Premium Member

I know the different camera menus (I'm talking from my Epic and Alexa experience) give me the option to chose between 1.3x, 2.0x and 2.0mag (this last one I have no idea of what it is) anamorphic desqueeze factor but I have no idea of what the rule of tumb is and which one to chose according to what.

 

It's pretty simple. Use a 1.3x de-squeeze setting for anamorphic lenses with a 1.3x squeeze factor - basically only the Hawk 1.3x lenses.

 

Use a 2x de-squeeze for pretty much all other anamorphic lenses - Panavision, Hawk, Cooke, Arri, Lomo, JDC, Kowa, etc.

 

Use 2x Mag(nify) if you are using 2x lenses and want to blow up the image to fill the 16:9 frame vertically - basically if you want the anamorphic artifacts like horizontal flares and squeezed bokeh but want to record a 16:9 image.

  • Upvote 1
Link to comment
Share on other sites

  • Premium Member

The stretched bokeh of the 1.3X lenses is more subtle, plus the Hawk lenses in general don't easily give you a big blue horizontal flare. But if you look at that trailer for "The Hateful Eight", shot with 1.25X anamorphic lenses on 65mm, you can still see some stretched bokeh, same goes for the 1.5X squeeze of Technirama.

Link to comment
Share on other sites

  • 4 weeks later...

thank everybody, onlu one last question to see if I got it right: so using "normal" 2X anamorphic I'm recording only the 1:1.20 portion of the sensor? and does that mean that some part of the sensor is not "used"? while the 2Xmag gives me the option of using the whole 16:9 sensor size?

Link to comment
Share on other sites

  • Premium Member

Yes, the whole sensor is not used when you are recording only a 1.20 : 1 area of it for 2X anamorphic photography, if the end goal is a 2.40 unsqueezed image. Why would you want to record a wider sensor area? I mean, you could if you wanted to, but what's the point?

 

Just think about it, your 2X anamorphic lens is squeezing the image laterally by 2X, meaning it has to be stretched back put laterally by 2X in post. Max sensor size in the MX Epic is 5120 x 2700 pixels, a 1.896 : 1 ratio, so if a 2X anamorphic lens even filled that entirely (it might not being 27.7mm wide, and most anamorphic lenses were designed for a film frame that is 22mm wide), you'd have 3.79 : 1 aspect ratio image once unsqueezed.

 

If you want a 2.40 image but want to use most of the sensor, you'd have to use an anamorphic lens with less of a squeeze ratio, like the 1.3X Hawk anamorphics.

 

Since the Red camera doesn't support a 1.3X squeeze, you'd have to watch the image that way on your monitors, full-frame but with a 1.3X squeeze to it.

 

The 2X Mag function is still cropping the sensor, if Satsuki is correct in its function: in fact it is even cropping more than the normal 2X function does, because instead of cropping the unsqueezed image down from 3.79 : 1 to 2.40 : 1, it is cropping it even more down to 1.77 : 1. If you want to end up with a 1.77 : 1 image (full-frame 16x9) but your lens has a 2X optical squeeze to it, then you only need to use a .885 : 1 area of the sensor.

Link to comment
Share on other sites

  • Premium Member

2X anamorphic lenses were designed originally for 4-perf 35mm photography and projection. The shape of the negative and print area is roughly 1.20 : 1, so when the image is unsqueezed by the anamorphic projector lens, it becomes 2.40 : 1 on the screen.

 

Therefore a digital camera would likewise have to have a more square-shaped sensor if you want to use 2X anamorphic lenses to squeeze a 2.40 : 1 image onto the sensor with little waste.

 

On the other hand, the resolution of the sensor is also a factor. For example, 4x3 Alexa sensor records 2880 x 2160 pixels (1.33 : 1 aspect ratio). So there is little waste to crop 1.33 : 1 to 1.20 : 1 for anamorphic photography, you'd end up using a 2592 x 2160 pixel area for 1.20 :1, expanded in post to 2.40 : 1. So because there is little waste, is that better than using, let's say, a 6K Dragon Red camera where the sensor is more widescreen, and therefore more is wasted for 1.20 : 1?

 

With the 6K Dragon sensor, full-frame is 6144 x 3160, and in 2X "ANA" mode (6:5 mode, i.e. 1.20 : 1), you are only recording 3792 x 3160 pixels. But that is still more than 2592 x 2160 pixels on the 4:3 Alexa for a 2.40 : 1 anamorphic image.

 

On the other hand, on the 6K Dragon, you'd be putting that 1.20 : 1 image in an area that is 18.95mm x 15.79mm. On the Alexa, that 1.20 area of the 1.33 sensor is 21.38mm x 17.82mm. On a 4-perf 35mm camera, the negative area for anamorphic is 22mm x 18.59mm, cropped in projection to 20.96mm x 17.53mm.

 

So in terms of field of view, the anamorphic lens on the 4x3 Alexa would be closer to what you'd get with that same lens on a 4-perf 35mm camera, while on the 6K Dragon, it would be cropped a little tighter in view.

Link to comment
Share on other sites

  • 1 year later...
  • Premium Member

If you put PV mounts on them (the Alexa 65 uses a XPL mount and usually the Sony F65 has a PL mount), then the Sony F65, being a Super-35 width sensor instead of a 65mm width sensor, is the closest to the 35mm 4-perf anamorphic format.

 

I said the negative area for 35mm 4-perf anamorphic is 22mm x 18.59mm.

 

On the Sony F65, the sensor is 24.7mm x 13.1mm. So if your maximum sensor height is 13.1mm instead of 18.59mm, then you'd be cropping the sides of the sensor to get a 1.20 : 1 area that when expanded by 2X horizontally becomes 2.40 : 1 because you are using 2X anamorphic lenses. If you do that math, that's an effective 15.72mm x 13.1mm area of the sensor. You'd lose some field of view so your lenses would be a bit more telephoto unless you used shorter focal lengths to compensate.

 

This is the same issue with most 35mm digital cameras because they don't have 4x3 sensors, unlike the Alexa, the sensors are more widescreen in shape, shorter than 4-perf 35mm film in height (more like using 3-perf 35mm film).

 

The Alexa 65 has the opposite problem -- the sensor is 54.12mm x 25.59mm (6560 x 3102 pixels). A 35mm 2X anamorphic lens was designed to fill 22mm x 18.59mm but many will project an image circle a bit larger outside that area, but I don't know if it would fill a sensor height of 25.59mm. But let's say that it does, since it can fill a width of 22mm. If you use a 25.59mm height and will only be using a 1.20 : 1 area (again because the image has a 2X optical squeeze so expanded becomes 2.40 : 1), then that means you'd be cropping the sides of the Alexa 65 sensor from 54.12mm to 30.71mm (again, I'm not sure if the anamorphic lens can fill a 30.7mm x 25.59mm area) -- you'd be cropping the 6560 pixel width to 3727 pixels.

 

At some point, with all of that cropping, especially if you have to crop all around because the lens won't even fill the 25.59mm height, you might as well just use the regular 4x3 Alexa, after all, it's the same sensor just cut smaller than the 65mm one. You'd just be wasting money using the Alexa 65 at that point.

 

The Alexa 65 sensor recording is already 2.20 : 1 just using spherical lenses.

  • Upvote 2
Link to comment
Share on other sites

  • 2 weeks later...

Thank you so much for such a detailed reply. :)

 

Oh, dear. So what kind of approach do the cinematographers use these days when wanting to diminish the incredible sharpness of these cameras?

 

The reason why I asked this in the first place is that I was re-watching Magic in the Moonlight, a film whose look I really, really like, even though I kind of think about it as “wrong” – meaning that even though at times it evokes the sunny feel of summer, I can’t say that it looks or feels like the French Riviera in the summer. There is something “cold” about it. (In fact, I think that Khondji somewhere stated about this film that this is where they first attempted to move Woody Allen away from all the warmth, though I have to check this. Odd, considering that summer indeed should be “warm”, on the face of it. Later the distancing from warmer looks continued with Irrational Man.)

 

Anyway, these lenses give really gorgeous looks to out-of-focus backgrounds. There is this wonderful appearance of the fuzziness of the background and this (semi)circular pattern in that fuzzy background, which I’m sure has a name, but which I don’t know.

 

So I was wondering if you can mount these lenses onto a supersharp digital camera and create something similar.

 

Perhaps there are other lenses more suited for this purpose?

 

These lenses used for Okja, as per IMDb, the Panavision System 65 and Primo 70 Lenses and Panavision Super Speed MKII and Ultra Speed MKII Lenses, what kind of look do they give?

 

The sharpness of these sensors is sometimes

– just look at those sharp trees in front of Hagia Sophia in the background around 1:01 or how sharp Istanbul in the background looks around 2:23. In reality, I don’t know which camera was used to shoot this, but somehow I think it was a digital one.
Link to comment
Share on other sites

  • Premium Member

You can use filters to soften the image, or you can use softer spherical lenses, you don't have to shoot in anamorphic just to get softness or less depth of field.

 

Look at "Arrival", shot on Zeiss Ultra Primes with the coatings changed to resemble old K35 lenses, and flashbacks on Zeiss Super Speeds shot wide-open, which are fairly soft at that aperture.

 

People have used old remounted optics such as Cooke Panchros, B&L Baltars, Canon K35's, Panavision Ultra Speeds (now remounted and called P-Vintage) or shot on some modern lenses with their coatings changed/removed. And of course older zooms are not super-sharp.

 

And Vantage makes a T/1.1 lens set that gets pretty dreamy wide-open.

 

And there is a huge variety of diffusion filters on the market.

 

And if you want to shoot anamorphic, you can use them on most of the 35mm sensor digital cameras though you may end up cropping the sides of the sensor if it is not close to 4:3 in shape, but if you're attempting to reduce sharpness, then high resolution isn't your goal anyway so who cares if you end up cropping your 6K Red camera horizontally to 3.8K?

 

It's only the Alexa 65 where there isn't much reason to use 2X anamorphic lenses because it's already a widescreen sensor, though "Rogue One" used the old 1.25X anamorphic Ultra Panavision lenses on the Alexa 65 camera.

 

 

 

Oh, dear. So what kind of approach do the cinematographers use these days when wanting to diminish the incredible sharpness of these cameras?

Link to comment
Share on other sites

  • 4 months later...

On the Sony F65, the sensor is 24.7mm x 13.1mm. So if your maximum sensor height is 13.1mm instead of 18.59mm, then you'd be cropping the sides of the sensor to get a 1.20 : 1 area that when expanded by 2X horizontally becomes 2.40 : 1 because you are using 2X anamorphic lenses. If you do that math, that's an effective 15.72mm x 13.1mm area of the sensor. You'd lose some field of view so your lenses would be a bit more telephoto unless you used shorter focal lengths to compensate.

 

 

While using spherical lens on the cameras like F65, Red Dragon, Does the loss of field of view exists?

Link to comment
Share on other sites

  • Premium Member

In the future, all this will become a floating point. You have 1.3x anamorphics, the 2x anamoprhics, but you also have the 1.25x Panavision anamoprhics for the DXL (they're developing new set of Ultra Panavision lenses) and there are also some older Bolex 1.5x anamorphics floating around.

Link to comment
Share on other sites

  • Premium Member

 

While using spherical lens on the cameras like F65, Red Dragon, Does the loss of field of view exists?

 

 

Compared to what? All formats crop the projected lens image since it's a circle and most aspect ratios are rectangles. So when you say "loss of field of view" what is your frame of reference for a "normal" field of view? 3-perf Super-35? A 24mm wide gate?

 

The 6K Red Weapon camera has many different recording options that crop the sensor in different amounts -- and some of the larger options show more of the lens image than Super-35 does, not less.

 

You just have to compare the dimensions of sensor area recorded for your format of choice in the digital camera to whatever you consider your standard frame of reference for field of view.

 

For example, the 1.85 area of 3-perf Super-35 is about 24mm x 13mm. The Sony F65 active sensor area is 24.7mm x 13.1mm (1.88 : 1) so probably a 1.85 extraction for theatrical would also be 24mm x 13mm.

 

The Red Dragon sensor can be used for many different resolution formats, see:

phfx_redDragon_FormatKey_01_4k6K.jpg

 

So any format that records a wider sensor area than 24mm would see a wider field of view than Super-35 (unless you run into vignetting problems because the image circle isn't big enough), anything close to 24mm wide would see a similar horizontal view, and anything smaller than 24mm wide would have a cropped view compared to Super-35.

Link to comment
Share on other sites

 

 

Compared to what? All formats crop the projected lens image since it's a circle and most aspect ratios are rectangles. So when you say "loss of field of view" what is your frame of reference for a "normal" field of view? 3-perf Super-35? A 24mm wide gate?

 

I missed '4-perf 35mm/ Super-35'.

 

While using spherical lens on the cameras like F65, Red Dragon, Does the loss of field of view exists compared to 4-perf 35mm/ Super-35?

 

Only RED WEAPON 8K VV(VistaVision Dragon Sensor) - 33.60 mm x 21.60 mm

and RED WEAPON 8K VV Ana(Anamorphic) - 33.60 mm x 21.60 mm having height of the sensor greater than 18.59mm.

 

Does the height comparison of the sensor with 4-perf 35mm is relevant only when shooting in anamorphic?

 

Link to comment
Share on other sites

  • Premium Member

4-perf Super-35 has a 1.33 : 1 aspect ratio -- if you need a 1.33 aspect ratio or the 1.20 : 1 area that 2X anamorphic uses for a 2.40 aspect ratio, then the height is relevant to field of view, but if you are shooting spherical for 1.78 (16x9), 1.85, or 2.40 then using a 1.33 : 1 format or sensor matters less.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...