I don’t know if this is done before, or if it makes sense to post something about a WIP without even having a screen to show, but here goes:
I’m making a short movie entitled “The Fall” which will be as accurate a simulation as possible of what its like to fall from the top of a building 1200 feet tall. Now…I’ve always wondered what it would be like. So I’m going to try and find out without the cuts and bruises normally associated with plunging headlong into the firmament and decelerating abruptly upon reaching the ground below.
The format is going to be a first person view of the decent. This shapes up for a few things I’m going to need to do:
Build the skyline of a city
Have a realistic looking street below
A nice sky
Some reflective skyscraper glass
This could be a big project, but its only going to represent a few seconds of video. I’m looking for creative suggestions about what to add in, and keep in mind I’m not really looking to make this video funny. Its supposed to be exhilerating.
Once again…I don’t know if its been done before, but if it has, I’d like to see it. Discuss.
OO, I don’t really know what to keep in mind, but I do have something to suggest, look at skydiving videos… or watch a movie that has skydiving… or watch the matrix? with the phone falling scene? just thoughts and ideas…
Well the difficulty is that in every video out there, someone gets rescued by a parachute. There’s no video tape of something just flying headlong into the pavement onto the busy street of a city. 1) You’d kill innocent people doing it 2) Youd ruin your camera. So I have no reference material other than instinct and what I’ve seen of things falling. What I have going so far is coming along well though. I’ll post a short animation as soon as it doesn’t suck. %|
maybe too “artsy”, but as the person’s falling, the speed of the animation get slower, so that by the time they are within a few centimeters of the ground, it is very slow and then ends, and then the person watching is left wondering “did they ever hit the ground”? based on the idea that if the person keeps getting closer to the ground by half the remaining distance, they will never actually get there (because there will still be another half of the distance to travel)… but i’ll let the physicists and philosopher’s battle it out over whether or not a person jumping off a building really hits the ground
'nother idea:
or maybe never actually show the person in the camera. only show their reflection in the skyscraper’s window, or reflection in a bystander’s sunglasses, and then show the shadow getting bigger as the person approaches the ground.
Actually I never intended to show the person at all. Or the reflection. Its going to be a bird’s eye view of the ground until the last moment. Thanks for the ideas…that slowing down one is intriguing…
maybe have the camera point in the other direction, as if the person is lying on their back. at first all you’d see is sky, then the top of the building in the peripheral vision, then other buildings, etc… until the ground.
or maybe slowly rotate it, like the person’s rolling while falling. or would that make too many people nauseaus?
I was thinking about having it curve off before it hits the ground and weave between the buildings or something like that…keep the ideas coming. The rotating thing is cool…perhaps looking left and right as if “Faller”, as we’ll call the character, is absorbing the view or trying to figure out a way out of the situation.
Well, here’s what I’d do. Start with the character walking to the edge of the building. Then the character will look down, showing his feet and maybe some body, plus at the same time shoing how far of a drop it’ll be. Then he looks back up and slowly the view starts to rotate from looking forward at another building, to looking straight at the ground while you plummet.
Then for the hitting ground part, just have the ground get close and then it goes black. So it ends just as quick as it began.
Plus I’d say you do some math. Figure out how high this building will be. Then figure out an approximate weight of the character. Figure the feet per seconds he’ll fall, and then you’ll have a rather accurate length animation.
It’ll take 9 seconds for pretty much anything to fall from that height. Its going to be a 1300 foot structure. The difficulty will be having enough detail on the side of the building to give the impression of speed. I think I’m going to leave out the character body for simplicity.
On the contrary… If you want to give the impression of speed, then your difficulty would be not to be having enough detail on the side of the building…
Have you ever tried to take a picture of a motorcycle passing fast in front of your camera???
Do you have enough detail on the background objects, in the picture???
Erik: I don’t want this to be dramatic…so much as…raw unadultered vision of what it would be like to plummit to your death. Thanks for the idea though because I can use SOMETHING like that…like perhaps on “impact” I could just flash a picture of a baby on the screen or something bizarre like that. I’ll work with it.
Timonides: If I don’t have enough detail, at least enough horizontal lines, there’s nothing to give you a sense of scale, and hence, motion. If you look out the window of a moving car, yes, everything is a blur, but if you look out the front of the car, its quite clear (assuming you’ve got your glasses on and haven’t been using any foreign substance to excess, or it isn’t allergy season). Thus, I’ll need a degree of detail in the windows of the building to give the impression of motion. Here’s a test for you: Ever see a 747 overhead? Looks like its going pretty slow right? But if it were close to the ground where you can see the scale of the object, it’d be a different story entirely.
If your camera is following the 747 you don’t need that much detail on the “surroundings”…
Plus this would give some degree of realism on your scene… One of the reasons that most works are not that “photorealistic” is, that people spend to much time to add detail on their models… This thing is very bad, cause actually real natural shapes lack sharpness… Too much sharpness is very bad, especially when you go for photoreal… And ofcourse it increases the poly count and thus making the scene too “heavy” (you’ll understand that if you try to animate a very “heavy” scene…)…
On the other hand, if your camera is “fixed” and the “747” just passes in front of it, then you necessarily need A LOT of detail on the side of the building, while the moving object, doesn’t require that much of detail ;)…
In anyway, if you need much detail, try not to model it… Better “texture” it… Make a wise choice of texture, on your model…
ouupppss, sorry, I didn’t read your initial post, very carefully… I skipped the part were you mention it’s going to be a “first person view of the decent…”
O.k. that means that the camera are going to be the guys eyes… That changes things a bit…
Well this is interesting…
I tell you what… My opinion is that if the person’s “eyes” are focused on the ground (if he looks down…), coming closer and closer, then you don’t need much detail on the view’s “perimeter”, cause it consists of objects “moving” (from the guy’s point of view…) out of his field of view, very fast.
But if he raises his eyes and looks at the side of the building, yes you need detail on the building’s side, cause it is supposed that his eyes are a bit more focused on it…
Just try to imagine where his eyes would be focused to add (or texture ) the detail and where it won’t be focused to “neglect” the detail and give a slightly more realistic “tone”…
Perhaps you may need to fake some “depth of field effect” as well, if you are as perfectionist as I am…
The way I look at it: Life is extremely detailed. What I’m trying to duplicated is something realistic. If I don’t point detail in the scene, the human eye will notice when they’re watching. BTW…before the end of tonight I’m going to have my first proof of concept run posted as an avi file. Be here for it!
one thing to keep in mind, is ‘terminal velocity’ remember that as a mass falls it ‘speeds up’ its decent until it reaches ‘terminal velocity’ and is traveling as fast as it can go. I think wind resistance is either a crutial part of this equation or something that effects the outcome…
anyway, to make it as realistic you’d have to calculate…
I don’t remember how to, seeing as I haven’t taken a physics class in about 7 years.
I don’t think there’s enough time falling for the person to reach terminal velocity. In fact I think you’d accelerate for a long time after you started your fall. You have to be going over 100 mph to hit that speed I’m pretty sure. If someone knows otherwise let me know and I’ll just adjust my IPO curve so the decent rate is linear after a certain height.
Shiver me timbers thats more math than I did for it. I just went with x = (at^2)/2. Thanks for the work on that. The swap to feet/miles would be done way back at the acceleration and the height. The building is 1300 feet and the rate of decent is -32 m/sec^2
EDIT:
v at which you splash = -9.8 t = -9.8 sqrt(s0/4.9)
The velocity increases each second parabolically. So the velocity when you slash would not by -9.8 t, it’d be -9.8 t ^ 2
Velocity is linear. Displacement is parabolic. (v=at, s=1/2at^2)
After 9sec of resistance-less falling, you will be travelling at 317kph (197mph) and have fallen 297m (1302ft) which is about 115-120 stories.
The detail on the building probably won’t be of much use - 6sec down the building you pass nearly 24 stories every second - that’s nearly one story per frame - Motion blur will remove all detail from the building. What you need is sufficient detail of the ground below that you notice it getting bigger, very quickly.
just thoughts and ideas…
