[Doyle Lewis]

Correct me if I am wrong but I believe you are talking about spherical panoramas. like this:

You can take these and in a 3d software wrap them as a texture onto a sphere. For the most Part I have seen HDRI versions of them used as a GI lighting on an object in 3d scenes later used to be composted back onto a flat background, but there is no real reason that you could not use the sphere as a background to then move around in. The only immediate downside I could see to that is that you would not have correct paralaxing in your background so once you start orbiting around your scene it could break the illusion of your composite, unless everything in it is meant to appear far away. (e.g.: a plane in the sky could probably use this effect because everything else is so far away it would probably seem to paralax together) I believe there are ways to do this in AE although I am not extremely familiar with them. I do know though that AE’s raytracing options allow for you to use them as well as Video copilot’s element 3D.

Hope that Helps,
Doyle Lewis

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

It is really a matter of what you are more comfortable with and the style of your narrative. If you want a more gritty or saturated look CG will probably be fine, but if you want real world daylight shot then it is harder to pull off in CG. If your comfortable with CG enough to make it look realistic and it is easier then having to go out and find a tall building that will let you shoot then do it, if not then use footage. anyways if you want to do it CG, as always everyone’s main man Andrew Kramer already has a tutorial on how to do it. I copied the link below. This is before Element so he uses a separate 3D software but just do what he does but replace the outside 3D software with Metropolitan pack. Hope that helps.

https://www.videocopilot.net/tutorial/3d_ledge/

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

Hey chris if you have After Efects CS 5.5 or later, then the AE 3d camera solve might be a better solution than a 2D track for you since you want to attach objects that live in 3D space, Like a light and particle system. If that is not an option for you then check your z space for the light. Does the light have a value of 0 for its Z position? the farther forward or back you bring the light in Z space the more the light will exaggerate or lessen (respectively) the track from the parented null. Let me know if that helps or if you have further questions

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

Your guess is correct! you need to make the target of the track a null layer. You then parent your rocket footage to the null layer and then animate your rocket moving up in Y space using position keyframes. With the rocket parented that change in position will stay relative to the null.

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

So I have attached a project file that is the video you sent composited with some glows and atmosphere. I did a very crude track of your camera just by eye. You could replace this camera with an actual position information exported from the 3d software you used. I then created fog using a fractal noise and masking it into an odd shape. Set my blending mode to screen and lowered the opacity. I set the evolution to time*50 and scattedred duplicated layers in 3d space so that they would pass by the camera and scale as the camera moved towards the stadium. I also scrubbed the start times for them so that they all didn’t animate uniformely and changed random ones to an evolution of time*80. I then duplicated your footage added a multiplied glow to it and set the layer to darken. I played with the glow settings till i got one i liked. I added a black solid with a mask to create a Vignette to make it appear like it was shot with a lens and created a slightly tighter mask on an adjustment layer with a Gaussian blur to add a bit of fake focus. If you can make a zpass of your comp then ignore this layer and instead set the zpass to be a mask for a lens blur effect. To add to the fake focus I made a 3d adjustment layer and set it as the farthest back object in z space. Set a Gaussian blur and masked around the stadium so that the city and the sky would be out of focus. (once again a zpass could solve this in a much more precise way.) The one thing I didn’t add to the project file that I would if you have the ability is lens flares. If you have VC optical Flares I would suggest using the bright parts of the light as a luma matte and set the flares to track luma and make them subtle because there are a lot of lights.. Then use the lights on the stadium and create a luma matte and maybe set some brighter stronger flares to that. Let me know what you think or if you have any questions or suggestions

8017_introrenderexampleglow.aep.zip

Edit: I would also add some overall color grading to give it a feel and pull it into the same space. I did it after uploading this link and it added a lot to the believability.

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

what happens if you turn off your corner pin effect?

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

I would use the Autoscroll script that comes with AE. Just type it into the effects pane and it will pop up. I would create a a precomp that is long enough vertically to fit all the words you want to scroll. nest that inside a main cpmp and mask around the visible word so only it shows. then just animate the scroll speed parameter when you want it to do the “jackpot effect.” Lastly turn on force motion blur for the layer and the comp and you should have it.

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

I haven’t really done this myself, but I’m thinking if you want it to do rolls and such then you probably want to move the plane and not the camera. I would probably rig it to a stick. Paint the stick green if you don’t feel like rotoing it out in post. I think the big thing to remember with this sort of shot is making the shot look like you filmed it in real life some way. this meaning:

1. don’t let the shot look like macro photography. shooting miniature things tend to make your depth of field shallower. Try to keep the depth of field wide so that the plane looks bigger than it really is.

2. I would probably use a longer focal length lens because if you were shooting this for real you would have to be in a separate plane shooting out of the side looking at your subject plane and that would require your camera operator to zoom in towards the subject. This will make keeping the depth of field wide a little more difficult but what your are really looking for is how the lens will flatten the depth of the plane.

3. Also along the lines of thinking about if you were shooting this for real, a camera op would have trouble keeping the shot perfectly steady. You can do this by either moving the camera around while shooting the plane or in post by adding a wiggle to a 3d camera. I would probably do it in post myself just because it saves you having to do all the 3D camera tracking work in your final composite. That being said be sure that you leave plenty of space in your frame to crop in so that you can get the wiggle without having to crop too close into your plane. It also helps to shoot at a higher resolution than you want to deliver the video at so that you have no loss in quality due to the digital crop.

4. lastly add some character to the model plane you have. Give it some scratches and dents so that is seems like it is real and used in the real world.

Hope that helps

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

Thanks for the clarification Todd,

That makes sense. I think what he is trying to understand though is why that doesn’t (at least seemingly) save him time in the export queue.

Doyle Lewis, Assistant Videographer

thinkck.com

[Doyle Lewis]

Really the only thing you can do in a render is lower the bitrate. really the only thing the caching does is keep Ae from having to calculate the next frame, not from having to encode it in the final render.

Doyle Lewis, Assistant Videographer

thinkck.com