Maximising sensor data, retaining compositing restrictions?

Grant Hodgeon

Hi all,

First post here and apologies if sent to the wrong area. I'm on the pinnacle of understanding how to truly work with capturing composite imagery as high a resolution as possible whilst trying to reference a different image captured on a different sensor/lensl

Here's the example of intent and where I've understood so far;

Reference/Plate image = 5DSr with 17mm lens. EXIF information says the Angle of View is 93.3 deg.

So the reference data is a 35.8mm sensor width, 17mm lens and sensor area calculated angle of view is 93.3 deg (width is 92.9 deg)

The reference image is a 35mm still that is wide angle shot of a hospital and the intent is to introduce two characters to the scene.

The two characters are to be shot in a studio environment and composited in. The intent is to photograph them on a 100mp sensor that is 53.7mm wide. But we'll want to maximise our resolution of our models, so we'll shoot vertical. For reference the height of the sensor is 40.40mm

Now this is where thing start to become muddy in terms of priority and what get's lost with what compromise, call it learning on shaky theory.

Do we crop our original 'sensor gate'/reference image to the size in which we'd like to place our model -- at a scale in which we think is best? For example our reference image scene is 30 feet apart and our models would be taking up only a small portion of the scene, so then work out Angle of View from our new determined sensor 'height', and find the equivelent lens for our larger sensor/camera and shoot at the same REFERENCE subject-camera distance reference which in our original data was read to be 2.53 meters...

Do we calculate our horizon line prior to doing this?

Is the Horizon Line the only priority? Am I conflating/confusing different methodologies on an otherwise simple process?

I feel like Angle of View RELATIVE to Distance can change how a subject *scales* and I'm not sure if that thinking is correct, I'm worried this might start a confusing discussion and I'm hoping it is just a clear statement! Considering we know in our original scene the focusing distance was set to 2.53 meters, and we know our subjects would sit best at that distance for our composite, shouldn't that Angle of View & DISTANCE value need to relatively match on the larger sensor in order for the scale to be correct?

I'm flying blind with half intuition and half baked theory, statements like 'I feel' are no good, but given I've gotten this far, any direct input from the daily drivers of this kind of information will help me a long way...

DIT with an F in TRIG

Gavin Greenwalt

If they are small in frame, a good distance from camera and on a chroma screen you only have to consider two things 1) relative physical position of cameras (height and distance of camera from subject) 2) lighting.

Just be sure to document what everything was shot on for you the compositor to align the plates.

Gavin Greenwalt
VFX Supervisor
Seattle, WA

Grant Hodgeon

> you only have to consider two things

It almost seems too reductionist to even say physical position of 'camera'. What is your definition of the camera? The sensor plane? The focal plane? The sensor plane is physically at a different location depending on the lens and sensor combination, so I assume we're looking to match the focal plane height and distance.

Height and distance doesn't include angle which is important, no?

And we're trying to maximise the detail captured by the sensor so we're concerned about what is almost an argument in other areas of this forum -- does distance vs focal length affect the 'shape' of the subject -- but more importantly, does it affect it enough to be noticeable? If not, cool. We're pushing from a FF 35mm Sensor to a 100MP one at 57mm and this larger sensor & large format lens shift alone has been argument enough with the folks on the other side of the camera. I might be adding to my own confusion.

DIT with an F in TRIG - still.