Creative Communities of the World Forums

The peer to peer support community for media production professionals.

Activity Forums Adobe After Effects Mimicking real world dimensions and camera (realistic dimensions, meters vs pixels)

  • Mimicking real world dimensions and camera (realistic dimensions, meters vs pixels)

    Posted by Micz Kicz on March 30, 2016 at 9:35 pm

    Hi,

    I want to build a realistically measured 3D space in AE and use “real” camera settings to “film” it. And I hope someone here will help me understand the way AE is measuring space. Let’s say I have a photo of a 2 meters high gentleman, shot from 5 meters by the DSLR camera Nikon D300 with 50mm lens. Now I want to replicate, mimic this particular framing of this particular person with those particular lens in AE.

    I think I should start with the Camera Settings and set Film Size and Focal Length to match my Camera’s specs.

    Film Size

    Nikon D300’s “film size” is – as I understand – it’s sensor size. And web says it’s sensor is: “12.3MP DX CMOS, 1.5222x crop factor, 23.6 x 15.8mm. That’s 181 pixels per millimeter” (source: https://www.kenrockwell.com/nikon/d300/specifications.htm). Now, I presume it’s all in the math, but frankly I don’t know which values are important here and how to do the math.
    How the AE’s Camera Film Size is measured in comparison with how it’s measured in real life cameras? Can you somehow explain it to me and other confused out there?.

    Focal Length

    Then, when the Film Size problem is over, and AE Camera somehow mimics size of Nikon’s sensor, I presume that Focal Length should be set to 50 mm – as in my Nikon. Am I right?

    3D space

    Now, when the AE camera is “real” I should pose my 2m-high-guy Layer 5 meters in front of it and… here’s the dumb question: how many pixels equals 5 meters? How to convert those 5 meters into pixel value?

    Way of thinking

    Just to make sure You guys understand me, let me put it that way: This whole thing is quite important when you want to pre-visualize some camera and/or actor movements in “virtual film set” which should be build in After Effects IN MATCH with real-world dimensions. Let’s say I want to replicate in AE a room which measures 4x6x3 meters in real-life. With AE shapes as walls. What dimensions IN PIXELS should those shapes have?

    I would gladly welcome any help in this, thanks!

    Creative Creature 🙂

    Kalleheikki Kannisto replied 10 years, 4 months ago 4 Members · 6 Replies
  • 6 Replies
  • Micz Kicz

    March 31, 2016 at 2:41 am

    Hahaha, I’m not sure if that’s what you’ve meant, but, Dave, let me take it as a complement 😉

    First of all – thanks for the reply. But what you are writing is not exactly what I meant. The same situation as you write about occurs when you project a “real” movie on a wall in comparison with that movie’s projection on a smartphone. Objects APPEAR bigger on the wall, but in both projections you see a 2 meter high gentleman, shot with 50mm lens from the distance of 5 meters. Those were the FILMING CONDITIONS – same for both projections: five meters from the two-meter-high object, seen trough 50mm lens.

    Now I want to recreate those filming conditions inside After Effects. So it’s rather a question of relation between dimensions of objects and camera settings in AE.

    In other words: I want to build a room from shapes for my 2 meter high guy, and be sure, that when I put this guy-layer into this room, and view this scene with custom AE camera mimicking my Nikon, it will look the same as in real life in terms of framing and relative positions of the objects. It seemed easy at the beginning – just create some shape-walls. But when I started, it hit me that I don’t know what high the wall should be. I can assume that let’s say 1 pixel = 1 mm and on that assumption my 2-meter-high-guy Layer will have 2000 pixels in height. But how to set the AE camera to be real-like in that assumpted conditions?

    I think that what I’m looking here for is pretty obvious when we approach 3dmax, or CAD – you can build there some architecture models, and place there some people in the appropriate scale. And then render it with use of some lens. I’m pretty sure that here in AE this way of building space is possible to achieve. But maybe I’m looking at it in some dumb way…?

    Creative Creature 🙂

  • Kalleheikki Kannisto

    March 31, 2016 at 1:23 pm

    Since everything is mathematically calculated, you can decide what pixel count you consider one meter and build everything based on that scaling. A pixel is not a size-dependent unit. If you want it to represent one mm, you can do so. If you want it to represent one meter, you can do so. If you base everything on the same scale, it will be correct in relation to other things, no matter what one “pixel” is.

    To see this yourself, you can link everything in a scene, including the camera, to a single 3D null and scale the null. Nothing seems to happen although everything changes in size.

  • Kevin Camp

    March 31, 2016 at 4:13 pm

    I’ll preface my suggestions with the fact that I have not tried this before…

    but, for the film size, I’d try 35mm. it’s what the sensor is likely mimicking (if it’s not actually that size), and what the 50mm measurement on the lens should be in relationship to.

    for the conversion of pixels to meters. in the shot that you have, create a solid that is 100 pixels high. scale that solid to match the height of the subject. the resulting scale is the subjects height in pixels (100px * X%/100%, simplifies to Xpx). so if the subjects height is 2m, then his heighten pixels is X/2, and your distance would be 5*X/2. and you could try this for the height of the camera too.

    you might want to txt this set up before eat shoot, if you can…. just shoot a box that you know the dimensions for and do the process outlines above. then create 3D box in ae, and place it in the scene, where the box should be and see if the perspective and lens distortion looks correct. if not, i’d try adjusting the film size.

    Kevin Camp
    Art Director
    KCPQ, KZJO & KRCW

  • Walter Soyka

    March 31, 2016 at 5:45 pm

    In Ae’s camera math, a pixel is 1/72 of an inch.

    Walter Soyka
    Designer & Mad Scientist at Keen Live [link]
    Motion Graphics, Widescreen Events, Presentation Design, and Consulting
    @keenlive   |   RenderBreak [blog]   |   Profile [LinkedIn]

  • Walter Soyka

    March 31, 2016 at 6:51 pm

    [Kalle Kannisto] “To see this yourself, you can link everything in a scene, including the camera, to a single 3D null and scale the null. Nothing seems to happen although everything changes in size.”

    DoF is highly dependent on scale, the absolute scale matters a lot if you want the camera to behave physically.

    Remove the parent null on a scene like this, with shallow DoF, to see what I’m talking about.

    Walter Soyka
    Designer & Mad Scientist at Keen Live [link]
    Motion Graphics, Widescreen Events, Presentation Design, and Consulting
    @keenlive   |   RenderBreak [blog]   |   Profile [LinkedIn]

  • Kalleheikki Kannisto

    March 31, 2016 at 7:15 pm

    [Walter Soyka] “DoF is highly dependent on scale, the absolute scale matters a lot if you want the camera to behave physically.”

    True, I didn’t think of that. Camera is based on real world scale, not pixels.

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy