Creative Communities of the World Forums

The peer to peer support community for media production professionals.

Activity Forums Live Events & Streaming Looking for a guru to give some input

  • Looking for a guru to give some input

    Posted by Robert Garry on March 2, 2006 at 5:18 pm

    Hi guys,

    I posted the following in the Broadcast Design forum as well but feel you gentlemen may also be
    able to give me some input into my conundrum. The original post folllows:

    I am being asked to try to incorporate an FX shot into a production I
    am doing for a client in the automotive industry. What I am being
    asked to do in my mind is rather complex so bear with me and my
    explanation.

    I want to map a car digitally so that I can get a wireframe image of a
    life sized car that I can track in 3D space. Again, wireframe image, no skin on the car.
    I understand there are ways to do this with lasers and tracking objects
    which I can translate to a 3D application to then create my wireframe.
    I am thinking mostly of motion capture or model tracing techniques when
    I say this.

    Here’s where it gets a little tricky though:

    This “broadcast” will be a LIVE show with a series of cameras shooting
    a real model of the vehicle on a stage.

    The clients are asking if I can shoot the show car on the stage and
    “map” the wireframe over the video of the show car in real time. They
    want to be able to highlight particular elements of the vehicle as they
    explain functionality and design elements.

    At this point I have persuaded them to not use moving shots as I felt
    this would add to an already growing production nightmare. I have very
    limited experience with “tracking” and “modeling” as my expertise lies
    in television production, editing and basic FX hence my posting to the
    community.

    So some straight forward questions:

    1. What kind of tech am I looking for to get this wireframe map?
    Perhaps someone could name drop a company, or a free lancer, that deals with this sort of technology?

    2. I would imagine I would need sensors on the showroom floor and a
    fixed camera with some sort of “computer eye” that would read the
    sensors and “fix” the wireframe to these sensors and hence to the real
    car for the video broadcast. What kind of tech should I be
    investigating?

    3. If, and that’s a big IF, we decided to move the camera (Ugh…) what
    kind of tech nightmares am I getting myself into? It is likely the car will be on one of those rotating stages so the wireframe would
    need to match the movement of the real car, again in real time.

    While I know there are many ways to do this in post I am unsure of how
    to translate this to a live broadcast so any suggestions that require
    conforming the elements and then spitting it to tape simply won’t work.
    I have already attempted to persuade them to go this route but they
    are stuck on this “real time” feeling for undisclosed reasons.

    The client’s know the budget for this could become large so no idea is
    too big at this moment.

    Any ideas or comments would be most appreciated. This project’s recent
    requests have obviously overstepped my knowledge, luckily we are still
    early in the development stages and I have plenty of time to learn!

    Thanks for any suggestions in advance.
    Bob Garry

    Robert Garry replied 20 years, 2 months ago 3 Members · 5 Replies
  • 5 Replies
  • Mickavid

    March 2, 2006 at 8:34 pm

    why don’t you convince them to script their dialogue and the do the performance in front of a large lcd screen that is playing back pre-produced elements. If it is elevated off the stage about 5 feet, they can have the real car in front of them while using the screen to show off the wire-frame-over-car images.

    hmmmmm,

    mick

  • Robert Garry

    March 2, 2006 at 10:00 pm

    I’ve tried to convince them of this but they really want the live feel and until they see that it either can’t be done or that it is astronomical in terms of expense I need to try to please the client’s dreams…nothing like pushing the envelope.

    Anyone else have any input?

    Thanks
    Bob

  • Bob Bonniol

    March 3, 2006 at 12:21 pm

    It might almost be easier to actually project the wireframe onto the real car (thus including the audience’s perspective on the trick) rather than dealing with pulling a post camera composite live…. But both really come from the same place in a way.

    First you must have a good model of the car in 3D. It would be quite surprising if the client didn’t already have detailed 3D data for the car. Then it’s a question of precisely locating your real camera angle vis the car, and camera matching that in your 3D renders. It’s ALL about knowing that camera position exactly, the focus and depth info precisely, and to avoid tremendous hassles it’s pretty critical to have the camera position locked off. Then you feed that data to your virtual camera in a good 3D app (say Softimage XSI, Maya, or Lightwave) It IS possible to derive camera position data for match moving in 3D from automated camera rigs, but the camera rigs are fairly big, and very expensive. Also you would have to have that move data before rendering, so you’d have to have the system in the venue, with the staging and the car, BEFORE you create the graphics.

    So a fixed position eliminates all that hassle. After that it’s just math. We have mapped alot of scenic objects with projected textures before in live environments with essentially the same technique. In that case you replace the camera position with a projector and project the render directly onto the object. It requires first projecting a guide grid on the object to have a map for predistorting the projected texture. Conversely you could use a 3D enabled media server (like The Hippotizer from Green Hippo, or the MBox Extreme from PRG) to map your footage directly onto the live object (in this case the car). Cirque Du Soleil is using this 3D capability in the Hippotizer to map video onto many moving scenic objects in their new Beatles show in Vegas (for instance), and the effect works well. We are using the Hippo capability to track multiple automated moving screen surfaces in Sinatra at the London Palladium. So this is not uncharted territory. The hard thing is the complex shape of the car, but again it’s been done.

    If yous want more info on this give our studio in Seattle a call (206) 399-8938 and we can discuss how it might happen and if we can help you. We might also be able to recommend the right vendors for providing the right gear to pull it off effectively live.

    Cheers,
    Bob Bonniol

    MODE Studios
    http://www.modestudios.com
    Contributing Editor, Entertainment Design Magazine
    Art of the Edit Forum Leader
    Live & Stage Event Forum Leader
    HD Forum Leader

  • Bob Bonniol

    March 3, 2006 at 12:21 pm

    It might almost be easier to actually project the wireframe onto the real car (thus including the audience’s perspective on the trick) rather than dealing with pulling a post camera composite live…. But both really come from the same place in a way.

    First you must have a good model of the car in 3D. It would be quite surprising if the client didn’t already have detailed 3D data for the car. Then it’s a question of precisely locating your real camera angle vis the car, and camera matching that in your 3D renders. It’s ALL about knowing that camera position exactly, the focus and depth info precisely, and to avoid tremendous hassles it’s pretty critical to have the camera position locked off. Then you feed that data to your virtual camera in a good 3D app (say Softimage XSI, Maya, or Lightwave) It IS possible to derive camera position data for match moving in 3D from automated camera rigs, but the camera rigs are fairly big, and very expensive. Also you would have to have that move data before rendering, so you’d have to have the system in the venue, with the staging and the car, BEFORE you create the graphics.

    So a fixed position eliminates all that hassle. After that it’s just math. We have mapped alot of scenic objects with projected textures before in live environments with essentially the same technique. In that case you replace the camera position with a projector and project the render directly onto the object. It requires first projecting a guide grid on the object to have a map for predistorting the projected texture. Conversely you could use a 3D enabled media server (like The Hippotizer from Green Hippo, or the MBox Extreme from PRG) to map your footage directly onto the live object (in this case the car). Cirque Du Soleil is using this 3D capability in the Hippotizer to map video onto many moving scenic objects in their new Beatles show in Vegas (for instance), and the effect works well. We are using the Hippo capability to track multiple automated moving screen surfaces in Sinatra at the London Palladium. So this is not uncharted territory. The hard thing is the complex shape of the car, but again it’s been done.

    If yous want more info on this give our studio in Seattle a call (206) 399-8938 and we can discuss how it might happen and if we can help you. We might also be able to recommend the right vendors for providing the right gear to pull it off effectively live.

    Cheers,
    Bob Bonniol

    MODE Studios
    http://www.modestudios.com
    Contributing Editor, Entertainment Design Magazine
    Art of the Edit Forum Leader
    Live & Stage Event Forum Leader
    HD Forum Leader

  • Robert Garry

    March 3, 2006 at 5:19 pm

    Thanks for the input Bob, this might be a good alternative soluttion. I’ll be discussing this with the clients and post back when I have more information.

    I posted this in the Broadcast Design forum as well and have been told that maybe the company Sportvision might be of some help. Anyone have any experience doing the types of things they do? (1st down marker in football broadcast, Nascar live updates, etc..)

    Anyone else got any input on alterante techniques or my original solution?

    Thanks for all the input so far and advance thanks to any future input!!!

    Gotta love the cow!!! 🙂

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy