Tutorial Suggestions!

yeah, I know that for jpeg.. I just said it as an example.. had no idea it could be done (noob).. png would be my choice normaly. I'm just happy to learn all this new stuff.. it certainly demistifies all the bullshit one has to take sometimes from uncooperative editors lol.

I just saw hypes entry at vfxtalk on multiplicity chalenge (congrats on 3rd place)



now what I wonder about is the flow of this guy

http://www.vfxtalk.com/forum/showpost.php?p=33811



not so much about his end result, but I do wonder what does this picture represent

http://www.vfxtalk.com/forum/attachment ... 1166949372


he speaks of setting layers in 3d space and tracking camera to that.. I didn't really get it.. could anyone throw a look at this and maybe elaborate if not make a tut on what he's talking about.. It's seem useful and relevant, I just don't know crap about it...

I'll check this tonite.. gotta run.. I have a date ha ha ha

I always hear a lot of people talking about these 'cards', and using them in Maya or Max, what is it exactly? What I've got in my head is;

They film someone against a bluescreen, or whatever, and then they roto out the actors so they're left with just the actors. Then export these out as an image sequence with the alpha channel. Then open the 3d package of choice and whack these image sequence onto a correct size plane, so when you scrub through the time line, you'll be able to see exactly what the actors doing, so then you can model everything around them etc? I saw in the making of King Kong, that they used these 'cards' in the scene where everyone was chanced in that massive dinosaur scene - where jack black runs with his camera over his shoulder and a massive stampede - they were able to actually move these planes as well, as in animate its location, and it still fitted in perfectly with the scene - I would have no idea at all how to even start about doing this! Have I got it right in my mind what these things mean?

well, it's really not that specific, but yeah, that's pretty much the idea. it doesn't have to be actors on a greenscreen, it can be layers of a mate painting, so you could do a pan on a matte painting and be able to get some perspective shifting in there. Zach uses it perfectly in his latest video tutorial:

http://www.simplycg.net/viewtopic.php?t=2397

in that second picture zach posted, you can see the little planes representing the ground and background. Zach used them as nulls to parent an image to, but you could very easily size it however you like and stick an still image or video footage on it. In fact, it's alot like the old techniques of front-projection and rear-projection, except you're not limited by putting up a screen that's perpendicular to the camera lens, you can position them anywhere in 3d space.

here's another example bzudo and zach worked out.
http://www.simplycg.net/files/video_wal ... le_104.jpg
(actually, i'm don't think they used AE 3d on this, but this pic perfectly presents the idea anyway.)

It's sometimes referred to in FX houses as 2.5D, and is used alot for background things.

here's an example so complex someone wrote an AE expression to do it automatically.
http://www.aenhancers.com/viewtopic.php?t=160

here's another, in blender.
http://orange.blender.org/wp-content/th ... enshot.jpg
that one, you can see all the "layers" as those black boxes all stacked up in front of the 3d part. they're all distributed in 3d space, even though they're all 2d paintings, so if you put a camera in front of this and pan it slowly, it'll have tons of depth to it, and should pass for a full 3d scene.

you could also think of it as camera projection, but using only planes instead of building actual geometry that matches the scene.

Wow man - thanks for a nice detailed explanation! But I'm still confused about the whole situation.

Say if I had a locked camera filming me running for the distance towards the camera, and then passing the camera. Would you simply roto me out, whack me on a plane in 3d, match the camera's location in 3D, build the scene around me, and then animated the position of the plane going from the background to actually passing the camera?

The link to the AE Enhancers does work man - would be intresting to see that one.

no. that scene would be built around you and you would just be composited into the scene. you wouldn't need the plate of you running to make the 3d scene.

like, say you had to make this shot (from SW Episode 1):
http://www.avoe.org/URBANLOVERS-7/naboo-4.jpg

looking at that image as it is, the water is a real element (something they went out and shot somewhere), the little creatures are entirely CG, the submarine is CG or a miniature (probably CG), and everything else in the scene is a matte painting. The buildings, the far background, the sky, and the trees close to the camera on the left. All matte paintings on seperate layers in the compositing program.

sure, perhaps the foreground buildings started as 3d, or maybe a combination of real building elements and CG elements. Either way, in the compositing program, it's a flat 2d image.

So then George lucas changes his mind and decides he wants the camera to be pushing in closer to the sub as it surfaces, or maybe dollying along as if on the bank next to the river.

if those are all just elements layered on top of each other, when you move the "camera" (the viewport, be it 2d or 3d), there will be no perpective change. it will look like you're panning the camera over a piece of paper with that picture on it.
But if each element is on it's OWN "card" (layer), and you space them out in the Z axis, then when you move the "camera", the far background won't move very much, but the trees close to the camera may go entirely out of frame. Thus, you've just cheated a 3d shot using only 2d. And, if George wants more changes, like make a building roof a different color, it's much easier and faster to do it in compositing or send it back to the matte department rather than having the 3d guys re-render it. what's faster to render - a bunch of detailed, high polygon models with high-res textures and some serious lighting and global illumination, or five single polygon planes? It saves countless hours in production. And that's why you can do it in AE and other compositing programs, cuz they don't even want you to leave your compositing program to do such a simple 3d thing!

of course, if you have a big camera move, this technique doesn't hold up, because it's not actually 3d and no obscured parts of the buildings can become unobscured, because they're not there. So, this technique is typically used for background stuff and set extensions, things like that.

Does that help a little?

Hi guys,

I'm literally pulling my hair out. I'm learning about matte painting and camera projection techniques. There's one thing that really puzzles me so much and has done for some time now.

When creating a camera projection in Max and you encounter stretching you need to project from a second camera, i just don't understand this process of how you setup the materials to project two images onto one object and then somehow blend between the two.

Take this example, a recent test:

http://www.ptarmiganpr.co.uk/video_tests/flash/c1.mov

original plate:



The grand canyons, notice at the end of the cam move you see the right rock face, there's lots of stretching in this area. To correct this i need to project from another cam.

Also the furthest back plate tiles, you can make it out a little to the top right, this is my second problem.

Some one pleeeese help

- Dave.

Trust me mate i've spent the last few days on a mini quest trying to figure this one out, i really thought it was quite simple at first

I started working on another test yesterday, you can download the files here:



The main problem i'm having is setting up the materials to work with mutiple camera projections using per pixel maps to patch the stretched areas.

No luck as yet, hope someone can ccome with something

- Dave.

Ok, I'm at work and not anywhere near Max right now, and certainly don't have all that much experience doing camera projection, but here's something I thought...

So, camera mapping is, basically, putting a texture on an object from a specific angle, right? throwing a flat picture across sculpted geometry. This same technique can be accomplished using the plain old UVW Map set to Plane. Then you basically just have to move the planar gizmo to match the focal range/target of your projection camera. Then, what you have is camera projection without the camera. Make sense? good...

So, now that you have the same result only using traditional mapping methods, it might be easier to paint a small extension and add it into the texture, using whatever you want - material IDs, you could turn a top/bottom material 90 degrees so its a left/right material, something like that. Basically, when you take the camera out of the mix, you might be able to start thinking of those blank faces as just a few faces that need textured, not as places the camera can't reach.

I hope that makes sense to you guys. it does to me, but maybe I didn't explain it clearly.

Like to see an AE video tutorial of a digital clock like this one:
with a screen tracking placement ą la Hype.
Sean, maybe this could be something for you?

there should be a preset for number animation like that. If you're using After Effects 7.0 or higher, look through the Text Animation Presets. You'll see a bunch like that. No tutorial necessary.

Thnx Hype for the tip. I'll wait for your wire tut anyway. You're really helpfull.
I am not on the 'pro' side but I am pleased to see you want help us and take some time for it.
Thnx m8.

I would like to see an advanced compositing Tutorial. Something that shows how to put CG in Live action. I would like to see some compositing with AE. It would be nice to see a transformation from one object to another. If anyone can make one or point me in the right direction, it would be greatly appreciated.



Thanx

definately mouse exploding tut (like the fake lactaid ad)

I am DESPERATE for a tutorial on UVW mapping.... ive read a load of tutorials and.... I have no idea what there talking about...anyone who finds a really good one gets an invisible chocolate (its non-existent flavoured)

Hey rrateme95. Its not really a quick solution. But i can recommend 3ds max fundamentals <---- (http://www.youtube.com/watch?v=jvyL3Mc77uk) This training set has a pretty good video tutorials on texture mapping. They texture F14 plane in those videos. I got a lot of knowledge from these videos. But it is costing 100 dollars to buy this training videos.

hey guys...

i have a request for a tutorial...

i have a Sony HDR-HC7, uses tapes...

i managed to transfer video from it on the pc and compress it, but i know i did not do it correctly because i have interlacing and the output was 32 gig for 10 minutes of video... basicaly i was just guessing where to go and what to click all the way though...

i would apreciate if anyone of you could show how to transfer video from your cam corder in premiere pro for example and then set an output format to export it as an avi or what ever...

in exange i can share anything i know...

ok kinda confused rich you want a tutorial on how to get your extract video files on your machine to get smaller and to be deinterlaced and have a format that works with primer or after effects.

i'd apreciate a tutorial that

1. shows how to tranfer video from the camera into premiere, or whatever software.

2. how to desinterlace footage if needed and get the best quality out of it.

3. export to avi, or whatever format so that i can save it on my pc, track it or composit it whenever i feel like it...

for the pull out of camera, that depends on the camera, usually camera comes with a software if not then at least give us the extension name (avi, vob, mov) that the camera gives, so i can research it for you.

2- deinterlace that i will cover in my next quicky. (hopfully tommorow i will start recording it i am still in quicky two for bhnh right to left text in all packages).

3-usually use mov or avi (mpeg4 codec) these two can be imported in AE for compositing and tracking, i will cover this quickly just a small touch with the deinterlacing one.

ps: please tell me your camera movie extension, your camera name model, and the size of the recorded footage using that camera (size not mb or gb i mean resolution and aspect ratio). THANK A LOT I WILL TRY MY BEST

if you want, I'll make a short one to show you what I do. But, honestly, alot of mine is automatic - meaning I don't do anything, but it works out anyway.

For example, I shoot my footage on miniDV tapes, nothing special. I capture in Premiere by just going to the list of cameras and picking my camera model out of the list. Then it can talk to my camera, and I open the capture window, make my logs, and batch capture. Premiere automatically makes them AVIs, since I am on a Windows machine. I never specified, when I installed Premiere, that I want AVIs, and I've never told it a compression to use. I don't even know where those settings are. I just use firewire. I always thought if I bought some fancy video card, like Blackmagic one, that then I would have compression options and settings. But using firewire, I have no clue where they are, or even IF they are.

So I capture to AVIs. I know the codec is Microsoft DV, even though I didn't select it. I'm fine with that. Then I bring in my footage to AE. I never, ever see that interlacing stuff. AE seems to somehow automatically deinterlace my footage, or its at least displaying it on the monitor as deinterlaced. Any of the footage I've ever posted here has been done this way, and I've never had those interlacing lines in it. I don't know why.

So thats what I do. I just capture it, and then that captured footage is already a format I can open in any program I want. I don't do any codec picking, exporting, rendering, none of that stuff. I wish I could explain it more, but that's seriously all I do.

As for getting the best quality footage, well, that's a whole other topic. I would recommend reading The DV Rebel's Guide. I think there's a whole chapter on how to save the 10% of info that editing programs clip out of footage. It's a great read, and explains how to save that 10% in detail!

A simply video tutorial about camera matching workflow please...