Any idea how to replicate that 3d digital camera look from the Radiohead video?

well, you probably know that no camera's was used in the making of the video,

so to imitate that, man, I can't think of another way than a 3D program, I mean AE's particle engines would be able to render the dot's and all, but how to position them to take shape of eg, a face, THAT's the question!

- Derrick
 
Maybe you could color correct your footage to be a super high contrast black and white and then use that as a displacement map to drive Trapcode Form?
 
Derrick_SA said:
well, you probably know that no camera's was used in the making of the video,

so to imitate that, man, I can't think of another way than a 3D program, I mean AE's particle engines would be able to render the dot's and all, but how to position them to take shape of eg, a face, THAT's the question!

- Derrick
Click to expand...

Thanks. I'm just using the term camera as they were taking images. I guess laser imagers would be a better term. Maybe it'd be a better idea to do the whole thing in a 3d program. I'm itnerested in even being able to take 2d footage and giving that dotted look to it.
 
grahamdunn said:
Maybe you could color correct your footage to be a super high contrast black and white and then use that as a displacement map to drive Trapcode Form?
Click to expand...

That's exactly what I was going to say. They kinda lied when they said no cameras and no lights. Because as mentioned above, it could have been done with a 3D model, but they still would have had to film him with a motion capture facial setup to get the movement.

I see two main sets of "dots" one is behind the head and is set up to respond to frequencies of the song (they are noticably bluer) and his face outline, in which the dots are greener.

That was filmed with him lit and with a camera pointing straight at him. That or motion capture. Then they Luminance channel if it was video or a model of a face set with minor tranparancy, and exported with a depth or Z map was created, and then applied like a displacement or normal mapping to give it depth.

So no, the video itself has no video or lights, but all of the reference material they gathered to create it did.
 
As I understood it, the video was filmed with a "wall" of sensors, each capable reading how far directly in front of it something was. Then they used this data to create how close each dot was, and gave it a corresponding luminance number.

I may be wrong but that's what I garnered from what little information is out there. Maybe we're talking about the same thing and I just don't understand you.
 
Danstin Studios said:
As I understood it, the video was filmed with a "wall" of sensors, each capable reading how far directly in front of it something was. Then they used this data to create how close each dot was, and gave it a corresponding luminance number.

I may be wrong but that's what I garnered from what little information is out there. Maybe we're talking about the same thing and I just don't understand you.
Click to expand...

We are talking exactly about the same thing. Go to Wikiapedia and look up Displacement mapping and Normal Mapping and you will see exactly what you just described.

Nice job on your part, especially if you're not familiar with the process. You described it perfectly. Good one.
 
I'm not sure if we're all talking about the same thing or not, but I think they used some sort of LIDAR or something like that to capture the footage. It's supposed to be used for surveying geography or something like that I think? I got the impression it was some sort of new technology and that this was a weird use for it, but the details are a little mysterious.
 
There could be a way to make this kind of effect with a camera...

But its a REAL dirty trick and you probably wont get as much detail as a laser or radar 3d recording.

I suggest you look up tutorials on how to get normal mapping from photos and adapt them for video footage of your actors face.

A normal map is basically a red green blue lit image to indicate shapes angles and height better on an object. Its used in 3d production and games to simulate texture and shape that’s not really on the mesh.

You can recreate this normal map [normally created digitally using high res models baked into a normal texture] using colored spot lights, or in still a single white light and several photos taken at key angles.

I cant remember what angle is what color so you need to check the tutorials on that one.

Once you captured this RGB lit footage you can use plugins in after effects to convert the normal into a grayscale hightmap. [Im pretty sure After Effects cant do this out of the box so you would have to look for a plugin.

I’m certain there are some out there. And there are some for photoshop [nvidia normal tools] so if you cant find a plugin for AE you can put the footage into Photoshop as a filmstrip and convert it there.

Take that animated map into a 3d package and apply it to a high poly plane mesh and apply displacement to the mesh using hightmap you created. Apply the dots texture to your mesh and render.

Again this wont be 100% perfect. The human face is complex you will likely get some shadows where the light won’t hit and you would loose height data for those parts. You might have to clean the sequence in ae rotoscoping and removing any unwanted anomalies before applying it as a hightmap in your 3d package.

Anyway i hope this wasn’t too confusing, Good luck.
 
not confusing at all, and very well described.

I would say you could even go so far as to put makeup on your actors face starting with white on the tip of the nose, and getting darker grays as you get further away from the camera. Now THAT would give you a height map from hell.
 
You can simulate the effect with Trapcode's Form. You would either have to use 3D renderings to drive the different layers, i.e., size, opacity, displacement, etc. Or do a hell of a roto job to create those from scratch on live footage.

On a personal note, I love Radiohead and all, but this video is just plain bad. The first time I saw it I couldn't make it past the first 15 seconds. I just tried to re-watch it to comment on this post and I didn't make it past a minute. Kudos to whoever for pushing the boundaries, but at least make it interesting.
 
You must log in or register to reply here.
Back
Top