Is 2:1 for shooting with anamorphics?

[Ryan Hamblin]

I am sure this sounds like a amateur question, but I am about to do my first music video on red...just wanna get all my ducks in a row.

thanks in advance.

(r)yan
www.alrrp.com/multimedia.php

 

[Dan Hudgins]

Anamorphic sensor height

Anamorphic sensor height

I am sure this sounds like a amateur question, but I am about to do my first music video on red...just wanna get all my ducks in a row.

thanks in advance.

(r)yan
www.alrrp.com/multimedia.php

Standard 35mm movie Anamprphic lenses squeeze 2:1 but the image on the sensor is not 2:1 it is about 1.175:1, almost square.

So you want to shoot with the sensor mode that gives you the most image height.

If you shoot on the RED ONE with 2:1 or 16:9 the image will be more "telephoto" than on a 35mm movie camera I think because the sensor height is smaller than the film 'Scope frame.

At any rate you would need to crop the TIF frames made from the R3D files so that you get about 1.175:1 images then de-squeeze those and letter box them into 16:9 frames with black on the top and bottom. In my system there is a "menu" the user can enter the values into to crop, de-squeeze, and letterbox in one step.

If you don't want a letter boxed 2.35:1 (2.39:1) image as the result, you can just crop the scope image to use half the width of the 16:9 frame and desqueeze that gives you an image with 2K "resolution" running 4K mode, because you make the pixels twice as wide.

2:1 mode on the sensor is 4K wide and 2K high, that is a non-standard ratio since movies are framed for 1.85:1 which is a little shorter than 16:9 HD ratio, so 2:1 mode needs to be cropped or letter boxed for release as 16:9 HD. 2:1 is bad for filmout since some screens may be 1.85:1 and you would get black on the top and bottom, something that is not normally tolerated for release prints, in that case the 2:1 would be cropped and squeezed and made into a 'scope print for 2.35:1 (2.39:1) projection.

The 2:1 mode was in the camera to save bandwidth since it has fewer pixels than 16:9. You can ask if the bugs are worked out of the 16:9 ration yet at 4K, it is better to shoot 16:9 if you do not want letter boxed results.

You can shoot 2:1 and crop to 2.35:1 (2.39:1) then letterbox into 16:9 to make it look like you shot scope ratio (super 35).

 

[David Mullen ASC]

Technically modern scope projection has a 2X squeeze and a 2.40 : 1 (2.39 : 1) aspect ratio, so the shape of the aperture itself is 1.20 : 1 (I'm rounding off these numbers.)

So if you shoot with 2X anamorphic lenses, ultimately you'll be cropping the sensor / recording horizontally to around 1.20 : 1 (with a 2X squeeze to the image, later unsqueezed to 2.40 : 1). I haven't worked with the ANA mode on the RED camera to tell you the dimensions of what gets recorded.

 

[Dan Hudgins]

Full frame height

Full frame height

Technically modern scope projection has a 2X squeeze and a 2.40 : 1 (2.39 : 1) aspect ratio, so the shape of the aperture itself is 1.20 : 1 (I'm rounding off these numbers.)

So if you shoot with 2X anamorphic lenses, ultimately you'll be cropping the sensor / recording horizontally to around 1.20 : 1 (with a 2X squeeze to the image, later unsqueezed to 2.40 : 1). I haven't worked with the ANA mode on the RED camera to tell you the dimensions of what gets recorded.

When Cinemascope came in the 'Scope camera height was full frame, but later changed the height to less than the max frame height. Some optical printer movements are projector full height (rather than camera full height or 0.1866" * 4) which would give you more like 2.40:1 (2.39:1) rather than the original 2.35:1, do you know the details of who decided to crop the scope frame down?

Looking at the film pitch numbers, at some point it must have been 0.1875*4 to get 0.75 frame stroke or MAX frame height? Then somehow became 0.1870 for print, and in order to use bell and howell D printer size printing drums the negative needed to be 0.1866 since the radius of the negative is smaller than the print stock on the printing drum.

 

[Ryan Hamblin]

So for music video purposes, it will be fine to shoot 2:1 for final delivery in 16:9, no letter box for internet, and letterboxed for broadcast?

(r)yan

 

[David Mullen ASC]

The scope frame is still close to being 4-perf 35mm Full Aperture height, Academy width, but yes, the dimensions were shaved down slightly vertically.

Basically the original idea for CinemaScope, to compete with Cinerama, was to project 4-perf 35mm Full Aperture (1.33 : 1) with a 2X squeeze, thus 2.66 : 1, with the soundtrack running on a separate 35mm mag fullcoat reel, like Cinerama did. But before the first CinemaScope movie was even finished, they changed their minds and created the "CS" perf print stock, smaller perfs to make room for mag striping on both sides of the print image. This shaved the width down to 2.55 : 1 when projected.

Then they decided to use standard optical track sound on the left side only, as all other movies did, which shifted the projected image to the right of the print and reduced the width to 2.35 : 1. If you've seen prints of early CinemaScope movies like "20 Thousand Leagues Under the Sea" you'll often notice the image is miscentered because it was shot to be cropped equally on the left & right side to 2.55 : 1, not on the left side only to 2.35 : 1.

Due to flashes appearing during projection at the top & bottom of the frame due to poor negative splicing, since the projected image was nearly Full Aperture, a decision was made by SMPTE/ANSI to change the specs for the projector gate and make the framelines slightly thicker. This reduced the height of the image and made the aspect ratio a little over 2.39 : 1 instead of 2.35 : 1 (most people round 2.39... up to 2.40). This happened in the early 1970's. The projector aperture was changed again in the early 1980's to standardize the width of all 35mm sound projection to .825". This changed the overall dimensions again of the anamorphic projector aperture plate but they kept the proportions to around 2.39-ish again.

So it hasn't technically been 2.35 since the early 1970's. However, many movie theaters still have anamorphic projector masks dating back from that time and thus still show 2.35. Plus screen masking (curtains) and projector throw, etc. all further act to alter the shape of what gets projected.

--

Yes, anything 2.35 or 2.40 that is broadcast would require letterboxing, whether on a 16x9 or 4x3 master. For internet distribution, you can crop out the letterboxing area and just show the picture area. In post, you'd start out by removing the 2X optical squeeze to the image.

 

[Ryan Hamblin]

Thanks so much for your insight. Growing up only on digital and 35mm adapters has me lacking in alot of true film related issues. Again thanks for sharing your info.

(r)yan

 

[davidtorres]

Ryan thanks for the question ...

 

[davidtorres]

Will be great shoot again with anamorphic lenses + Red One.

Thanks David Mullen + Dan Hudgins for your light.

David Torres / AMC

 

[Alberto Caprioglio]

Standard 35mm movie Anamprphic lenses squeeze 2:1 but the image on the sensor is not 2:1 it is about 1.175:1, almost square.

In order to use anamorphic lenses, a case in which the paradox is only more apparently severe but still apparent - sensor proportions are not important, as long as the sensor is more than enough in both directions. Nothing new. Sensor proportions are not more important than they normally are in non-anamorphic shoots. What is important, is to streatch the image again in post production, to make it natural again, of course this time along a 90 degrees direction compared to shooting.

And no, I don't think 2:1 was meant for use with anamorphic lenses, but you can try to play with all these toys, if you have the occasion. You will get extreme wide screen shoots.
I believe 2.1 is made to direclty shoot 2:1 - simply an effective pre-crop - and save data comparing to shooting 16:9 when you know from the start that you would crop it to 2:1 anyway.
2:1 may be uncommon among the hundreds of film standards, but apart from being (introduced??) in the past supported and sometimes used by director of photography Vittorio Storaro in pre-digital era, is a simple proportion with very simple numbers, so it isn' completely unreasonable. It's not less reasonable than the strangely adopted 16:9 proportion, which is a complicately derived proportion, a sort of mix of different previous systems.. well at least this is what I have read about 16:9.

I don't like anamorphic lenses very much when they are used not to make better use of the sensor's pixel count nor to make wider screen shoots but only because they make out of focus elements elliptical instead of round. To me it's as childish as making your mom's face all blue in photoshop, I don't care if this can help getting an Oscar or impress at a local film festival.
I beleive that if it were just for this result, anamorphic lenses would never be made.
They are made to overcome the limitation of a vertical running film (and today, for 4:3 sensors), for the important market of higher resolution movies and/or of wide screens, the latter two being the two very aspects of the same "problem".

 

[Dan Hudgins]

When to unsqueeze?

When to unsqueeze?

In post, you'd start out by removing the 2X optical squeeze to the image.

Thanks for the info.

About when to unsqueeze your 'scope 4K frames for post, although it is nice to unsqueeze the frames at the start, it makes the frame files twice as big, so can make post processing twice as long, and can require twice as many harddrives and backup tapes.

If you crop the squeezed anamorphic frames to on film ratio of 1.175:1 to 1.200:1 and work with the squeezed frame all the way through you can save thousands of dollars when making a feature or short that has a high ratio.

In my system I just turn down the CRT monitor height all the way and the width up all the way, and it is not hard to work with the squeezed footage through color correction. Or you can make downsized "proxy" frames to work with that are unsqueezed and conform the squeezed frames later.

If you are going to film 'scope print, unsqueezing then resqueezing the frames can degrade the images depending on how the sampling is done.

If you are going to internet or HD you just unsqueeze after the color correction and editing from the finished 'scope output frames when you downsize to the sub-4K format.

So the decision as to when to unsqueeze can have a big impact on your post time, like weeks or more of frame processing for an 4K uncompressed 48bpp DI, and the cost of post since the unsqueezed storage is double.

 

[David Mullen ASC]

Well, wouldn't a lot of people edit using unsqueezed lower-rez proxies or conversions, letterboxed SD or HD, then go back to the R3D RAW files for the final 4K or 2K D.I.? In that case, the quality of the first conversions for offline editing are irrelevant, and unsqueezing of the 4K RGB files would just be for monitoring purposes during the color-correction session in the D.I. theater, you wouldn't be unsqueezing and then resqueezing the actual files themselves -- you'd just leave the optical squeeze in and send that to the Arrilaser in the proper pixel dimensions.

As for something meant for broadcast, you'd be working towards an HD master, so at some point in the chain, earlier probably than later, you'd want to convert the image to 1080P 16x9 with a 2.40 letterbox. So you wouldn't be dealing with 4K files that were doubled in width anyway.

 

[Dan Hudgins]

Unsqueezed proxies...

Unsqueezed proxies...

Well, wouldn't a lot of people edit using unsqueezed lower-rez proxies or conversions, letterboxed SD or HD, then go back to the R3D RAW files for the final 4K or 2K D.I.?

When you do CG with the frames you may need to color correct the 4K TIF before you make the proxy frames with the CG elements added for editing, so you would have part or most of your frame processing workflow in 4K squeezed (or unsqueezed) before final editing is done.

The simple conform from the R3D files can get more complex if you were trying to conform the whole workflow with added elements after editing, it seems easer to do the composites with color corrected 4K frames, then make the proxy frames from those (complex) composites for editing?

My system starts with making "proxy" low rez frames for editing, but you "replace" the whole 4K camera shots that get composites or titles etc and make new "proxy" frames to show the end result, conform is not quite the right word for how my system works since it can operate at many resolutions at once, you just need the smaller "proxy" frames so that you can get sync speed playback, it works on the full resolution frames using the same commands.