Chris Smith
Forum Replies Created
-
Let’s look at one of the masters of speed on camera: Michael Bay. In Bad Boys (1994) at the end of the film, the gang is chasing the plane on a tarmac in that old 911 993 Turbo. Bay wanted the sense of extreme speed. So he backed waaaay off and shot it with a 1000mm lens. Plus he purposefully added barracades between the camera and the car as midground elements to accentuate the speed as the car flew passed them.
In AE, without a real BG, a 1000mm lens is quite a bit too much, but in the real world, the cam op could never completely keep up with the car smoothly and the cam is not only shaky, but sometimes couldn’t keep up as the car would almost leave frame and the cam had to jerk over faster to catch up with it again. Instead of barricades, I think you should place other graphicle elements, like say more of those circles halfway between the cam and your subject circle so it really accentuates the speed as the objects fly by.
Obviously turn on motion blur and maybe think about some soundFX as well which does half the work for you.
Chris Smith
https://www.sugarfilmproduction.com -
Under your comp window, is the alpha channel (whire button) the only thing engaged?
When you play this quicktime just from the QT player is it still white?
Chris Smith
https://www.sugarfilmproduction.com -
I gotta say that it IS a suicide mission if you’re starting 3D from scratch. I personally have been playing with 3D for 10 years and I wouldn’t take that (but I do despise modelling and texturing though).
Any 3D program is right for it. C4D is a great one. Learn here at the Cow, the books at Amazon.com, C4Dcafe.com, and 3Dattack.net among others.
Chris Smith
https://www.sugarfilmproduction.com -
Treat it like AE. Most motion graphics don’t use AE’s lighting and shadows and people do tons of incredible graphics with it. So instead of do the ‘ol 3D thing’ which is worry about materials and lighting. Do neither. On your C4D material, only engage the luminence channel to give your objects unlit flatness like a Flash/Vector look. Kill your auto light and don’t add lights. No shading, no specular. Then create cool worlds that grow and animate. It’s a cool look because people think of it as a 2D look but the 3D (a lot more 3D than AE can give you) really stands out. The comedy central bumpers from last year were doing this.
For other inspiration from a guy who only does his mograph in 3D (except a few overlays) go here:
Chris Smith
https://www.sugarfilmproduction.com -
Chris Smith
September 8, 2005 at 2:28 pm in reply to: TP – Matching Particle rotation to surface position normalsIf you’re using a PSurfacePosition node to attach the particles to a surface, pipe the ‘alignment’ out into the ‘alignment’ input of a PsetData node. Obviously pipe the particle stream in there as well.
But the objext axis of what you are attaching needs to be oriented correctly for the object to stick out in the right position. You may want to prcatice with a cone to get the axis right first.
Chris Smith
https://www.sugarfilmproduction.com -
Brings up a good point why there isn’t a start AND end ‘grow’ parameter instead of just ‘end’.
Chris Smith
https://www.sugarfilmproduction.com -
Other quick note. Bring a good Digital SLR cam to set with you (like a Canon 20D or the like) and take ref pictures as well. Some of these cameras have remarkably film like response and will give you a great idea of what it’ll look like. We use it all the time. If I’m uncertain how the DP lit something, I’ll take pictures of the shot from the main cam’s POV with the 20D and we’ll discuss changes. Works very well and looks extremely close to what you’ll get.
What 35mm cam did you use? It may have had a tap on it.
Chris Smith
https://www.sugarfilmproduction.com -
Search the archives. In the last month or two we had a thread about it and someone pointed to a plug-in that would do it. What versions it used I don’t know.
Chris Smith
https://www.sugarfilmproduction.com -
First of all DaVincis have a keyer in the form of a ‘kilovector’ the same way they grab a color and tweak it they can grab the green and if nothing else turn it all pink for example to see that it’ll work.
Video Assist: Almost all film cameras except really cheap ones have a video tap. When light comes into the cam it hits a splitter and the tap (CCD cam built in) gets some of it and outputs a video image for framing reference. They usually look like poo, but that’s what clients watch to know what’s being shot. On almost all sets a guy records this into a computer or tape deck for playback. For example if I’m working with an actor and we did a handful of takes, I’ll go back and watch the tap off the computer to see how they looked and if we need changes. We even pre edit the spots now on the computer while shooting to make sure timing and continuity are working well. Then we output this edit for the post house for ‘inspiration’ for doing the final edit from the film transfer.
When we use a bluescreen, we use the tap to balance it. Since most sony monitors have a button that shows just the blue channel as a greyscale image, we hit that and drop the brightness down. Then it’s like magic, it’ll show you where your hot and dark spots are clearly as a grey scale map on the screen. Too bad they don’t have the button for the green channel.
Yeah, lots of folks get flat transfers knowing they’ll color it later. I do this a lot as a backup on a second deck. So when they do the dailies to one deck, I take the direct output of the Spirit to a second deck. This way I can do my own cut and color it in Shake later.
Chris Smith
https://www.sugarfilmproduction.com -
Well technically DigiBeta compresses video 2:1 but it usually never effects a key. Ppl make great keys every day from the Film to Digi pipeline. If it looked better in the transfer suite it was probably that they: 1. had a much better monitor than you’re looking at 2. They were looking at their HD monitor while the transfer was to SD (unlikely).
It looks like they didn’t do much in the transfer. I’m guessing you asked for it flat?
As far as lighting, there are so many ways to do it. We usually use skypans to light the grnScreen, then if it’s one person, in close-ups, we’ll use all Kinos for key and back light. If it’s a larger setup or more than one person, then it’s 12x Muslin with 9-lights behind them.
As far as light metering, you don’t HAVE to be so precise. Often just on your video assist you can tell there is unevenness. But yeah, just stand near the cam and take some spot meter readings of various points on the screen and make sure they are relatively even. Especially around the talent. The screen doesn’t need to be good at all elsewhere in the shot, just at the edges of the person. But for example, because of the paper’s sheen, where it bends you have quite a bright highlight. Then on the floor, it’s very dark.
With some work this is still very keyable. However the closer your key is to even, the less work you do on the post side. Ideally if it’s all even then you can key it equally and not break it up into multiple keys with different settings for each area of the shot.
Also next time you’re in telecin