RED camera sensor size

Gian Claudio · September 21, 2009

Hi, I just began working as a 1st AC on a feature film shot with two Red cameras (4K, using Ultra prime lenses). I arrived when the job was already in its 5th week, so I didn't do any kind of camera test in the rental house. It's my first time with Red cameras.

I read on the Red site that the camera sensor sizes are supposed to be a 24,4 x 13,7(or something close to that); this to obtain a frame size that is the same of S35 standards (an Arri ANSI gate exposed area is 24.9 x 18.7mm; I realize it's higher cause it's a 1.33 aspect ratio vs the Red's 1.77 HDTV cmos).

Looking through the viewfinder I noticed that the field of view of any lens I mount seems to be much shallower than it would be in a S35 or even 35mm camera's groundglass.

It seems that the effective sensor area you are exposing is much smaller than the "super 35mm" frame they claim...

I confirmed this to myself also by leaving, by mistake, a 50-65mm hard top on the lmb5 mattebox, in front of a 28mm lens; when I realized my error I took a look in the viewfinder and absolutely no vignetting was noticeable! (obviously no "focus check magnifying" was engaged). The edges of the matte were not even visible in the safe area outside the 2.40 format!

Has anybody done (or heard of) a close comparison between Red camera and super 35 camera effective field of view? Assuming that I am wrong and they are the same, isn't the "out of frame lines but still visible" area (in the standard 2.40 "Red" evf markings) itself already much wider than the correspondent area that would separate any S2.40 frame leader test side edges from the ANSI (24,9mm wide) side edges exposed through the film gate, next to the perforations?

Thank you for any answer!

David Mullen ASC · September 21, 2009

Take a look here:

http://www.jfi.net/RED/REDFormats3.pdf

The gist is that while the sensor is Super-35 in width (approx. 24mm) , the recorded 4096 pixel image area is smaller in width, closer to Academy 35mm (approx. 22mm).

Gian Claudio · September 21, 2009

Take a look here:
http://www.jfi.net/RED/REDFormats3.pdf

The gist is that while the sensor is Super-35 in width (approx. 24mm) , the recorded 4096 pixel image area is smaller in width, closer to Academy 35mm (approx. 22mm).

Is also the safe area included in these 22mm?

David Mullen ASC · September 22, 2009

Is also the safe area included in these 22mm?

"Safe area" is whatever you want it to be. The dimensions of the 4096 pixel wide recording are documented but how much of that you want to use as a safe area is up to you.

Generally safe areas are not to show you what gets cropped out, they are to show you the area that you have to compose inside of in order to always be visible even in extreme cases of TV CRT monitor overscan. But you have to always protect the full transmitted area. "Safe Area" is partly a holdover from the notion of "Title Safe", i.e. the area that the titles have to stay within to always be visible even in extreme cases of TV overscanning. But the area outside of title safe may or may not also be visible. This is something I've had to explain many times to boom operators, that the safe area isn't an area that gets cropped and thus safe to put the mic into, it's not the outside area that's important, it's the inside area -- it's so you know that what's inside the safe area will always be visible, whereas the area outside the safe area may also be visible.

Generally you would use all the width of the 4096 pixel-wide RED recording for the image, whether for conversion to 16x9 HD broadcast or some cinema application. However, you may or may not also want a safe area in the groundglass / framelines. But the full recorded area generally becomes the TV transmitted area, which is slightly bigger than the title safe area.

John Sprung · September 22, 2009

With CRT's pretty much gone, it's time to re-think this safe area stuff -- and get rid of it. We really should be able to look at one rectangle that is the one and only actual picture. That would enable very precise and edgy compositions and let you use lots of negative space.

-- J.S.

Gian Claudio · September 22, 2009

"Safe area" is whatever you want it to be. The dimensions of the 4096 pixel wide recording are documented but how much of that you want to use as a safe area is up to you.

What about "Surround View, which is an additional visible area outside the actual recorded image (from the Red camera manual)"?

Does that mean the sensor has actually got more than 4096 horizontal pixels (even if they are not recorded)?

In other words; is this "not recorded but visible area", that I was by mistake referring as "safe area", included in the 22mm width?

Sorry David (and thanks for your patience!), I know you already answered me but today I tried putting the 50-65 hard matte in front of the 24mm and it still didn't cover the frame (it became visible only by closing the iris down to T11 and setting focus to the short end!) It's still hard to believe the recorded area is equivalent to the width of an academy 35mm gate when you look through the viewfinder!

Chris Kenny · September 25, 2009

What about "Surround View, which is an additional visible area outside the actual recorded image (from the Red camera manual)"?
Does that mean the sensor has actually got more than 4096 horizontal pixels (even if they are not recorded)?

Yes. The sensor has 4520 active horizontal pixels. (There's actually a mode in the latest beta firmware that lets you record 4480x1920.)

David Mullen ASC · September 25, 2009

Yes. The sensor has 4520 active horizontal pixels. (There's actually a mode in the latest beta firmware that lets you record 4480x1920.)

Yes, the active area is 4520 pixels wide, which is why you get a "lookaround" area on the eyepiece and monitor (one of the best aspects of the RED in my mind) the recorded area in 4K mode is 4096 pixels wide, unless you are shooting in "4K HD" or "Quad 4K" or whatever that format is called, then you are recording a 3840 x 2160 pixel area, even smaller than the normal 4K recorded area.

Dal Neitzel · September 26, 2009

With CRT's pretty much gone, it's time to re-think this safe area stuff -- and get rid of it. We really should be able to look at one rectangle that is the one and only actual picture. That would enable very precise and edgy compositions and let you use lots of negative space.

I would agree that there are likely fewer CRT computer displays than LCD computer displays but is the situation the same with CRT television sets? The "un" safe area also represents the 15% of the image area that is "worst case scenario" hidden behind the bezel on the world's tube TV sets. I suspect there are still millions and millions of SD CRT TV sets...possibly still more of those around than newer LCD screens. I think its a tad premature to start disregarding the safe area. But I certainly look forward to the day when I don't have to be concerned that TV viewers will see less than others.

Brian Drysdale · September 26, 2009

But I certainly look forward to the day when I don't have to be concerned that TV viewers will see less than others.

I'm not sure you can entirely ignore a "safe area" even cinemas crop on either side of their screens and the area that a LCD screen shows seems to vary depending if it's on "smart" "wide" "14 x 9", "4x3" or "zoom".