11 replies to this topic

[pedrosura]

I have been working with VUE software to animate shots outside moviestorm. One technique used by studios is to film actors and record the camera movement while they are filming to then

output this camera position to the Computer Generated Software and then composite shots using green screens. I wish there was a way MS would output the camera movements files. 

 

For VUE Software the  Supported file formats are Max Script (.ms),MatchMover (.rzml) and MatchMover Ascii Camera 3D Track (.rz3).

 

I am not sure what they are for other software out there?? But I think that this would be a capability that would be worthwhile since it would allow using MS to composite MS actors with Computer Generated animation. 

 

Obviously this is information known internally to Moviestorm. 

 

 

 

[tfish11]

You're way ahead of me mate, I'm still trying to get to grips with Green screen at the basic level. However MS is a learning curve and each of my films is an improvement on the last I would think 

[pedrosura]

You're way ahead of me mate, I'm still trying to get to grips with Green screen at the basic level. However MS is a learning curve and each of my films is an improvement on the last I would think 

You are very right. I look at my first film and it is great because I can see how far I have evolved. But i know that I can improve still a great deal. Obviously, your film is never going to be good enough. Which is great.

 

I love working with green screens. If it is within Moviestorm, then it is very easy because by copying the movie you have the exact camera movement. My problem is when I want to use green screens to composite shots from outside software. Mimicking the motion is very tedious and, in a sense, impossible. Very hard to make it look right. 

[Ben_S]

At the moment the camera information is in the <keyframe> nodes in the xml of the saved movie.

 

Free shots are the easiest to interpret - they are shown as target shots that do not have targets. <targetOffset> has the camera position. Pan, tilt, roll, bearing I think are all derived from the <keyTransform>

[pedrosura]

At the moment the camera information is in the <keyframe> nodes in the xml of the saved movie.

 

Free shots are the easiest to interpret - they are shown as target shots that do not have targets. <targetOffset> has the camera position. Pan, tilt, roll, bearing I think are all derived from the <keyTransform>

So Ben,

 

Theoretically, if I shot a movie using 100% free shots, do you think it could be possible to read the xml file and make a translation to something else? I am going to look into this. Perhaps i could do it with Excel or something similar...

Thank you

[pedrosura]

I found the format for Matchmover. 

MatchMover creates an Ascii file as follows.

imageSequence "Sequence00"{                720    576f("\\Spirou\public\images\MatchMoving\sgi320.avi" )                                            b( 0 238 1 )            -even}Camera{0                F ( 943.095 )            Pr ( 1.06667 )            Pp ( 360 288 )            K ( 0 )                Oc ( -42.4206 67.3594 -19.604 )                Rot ( 0.793155 0.598043 -0.115111 -0.158493 0.38519 0.909125 0.588035 -0.702832 0.400301 )1F ( 943.095 )            Pr ( 1.06667 )            Pp ( 360 288 )            K ( 0 )                Oc ( -41.9334 67.8446 -19.764 )                Rot ( 0.795385 0.594972 -0.115633 -0.157956 0.387665 0.908166 0.58516 -0.704077 0.402323 )}PointTrack Track00 (4.79921 0.0732729 1.00537)PointTrack Track01 (-0.00894661 0 0)PointTrack Track02 (-0.00894661 1.37239 10.0013)

• The label imageSequence contains the name of the sequence in the Project window.
• The values 720 and 576 refer to the size of the image.
• f( "\\Spirou\public\images\MatchMoving\sgi320.avi" ) gives the full path name of the sequence.
• b( 0 238 1 ) refers to the begin, end and step values.
• -even defines the type of interlace which can be Upper field first or Lower field first. The default value is no interlace if no type is specified.

 

 

I guess they are callled REALVIZ asc camera 3D  tracks. I am going to see the differences between this and MS for the formatting. 

• Camera

  The first line gives the frame number.

  • F indicates the focal length in pixels, defined as focal length (mm) times image width (pixels)/film back width (mm).
  • Pr indicates the pixel aspect ratio.
  • Pp indicates the position of the principal point.
  • K indicates the distortion value.
  • Oc indicates the Camera position (X, Y, Z).
  • Rot indicates the three lines of the Matrix of rotation. The last line of the Matrix, the last three coefficients, also represents the direction along which the camera is pointing, expressed with respect to the reference coordinate system.
  • The remaining lines of code give the Point Track label and its X, Y and Z coordinates.

[pedrosura]

i would be happy if we could in the camera view specify x,y,z position of the camera in the timeline. this could be a great first start..

[dgarygrady]

I haven't tried this, but depending on what you want to do a possibility would be to use camera tracking, which is supported in HitFilm, AfterEffects, and Blender, among other applications. It's normally used to insert CGI objects and characters into a real-world scene or actors into a CGI world, but Moviestorm could be used in the same way as the real world.

 

Suppose for example you want to insert a complex animated object (a robot, dinosaur, spaceship, or whatever) into a Moviestorm scene without getting into modding. In that case you could create the scene in Moviestorm and then use a compositing/CGI program to track the camera motion. Blender and HitFilm would allow you to animate a 3D model that you could insert into the scene. I confess I've never looked into whether AfterEffects allows you to animate models as fully as HitFilm does, but it may.

 

Alternatively, suppose you want to to insert Moviestorm characters into a CGI world created in another program. In Moviestorm you would make the background a chromakey color and place flat objects on the floor and wall planes that would be different from the main chromakey colors (for example, blue versus green, or dark green versus light green). Then in your CGI software you would motion track the background objects in order to general a camera solve.

 

You will generally want to avoid zooming during a shot, and depending on the software it would help to make note of the lens focal length (the zoom setting under the lens button on the Camerawork view when you select a specific camera angle).

 

The sample lessons from Hollywood Camera Work's Visual Effects for Directors can serve as a decent introduction to the concept of camera tracking: http://www.hollywood...ampleclips.html

 

There are plenty of YouTube tutorials on camera tracking in various software. Try searching for

 

    <software name> camera tracking

 

Here's a good example using Blender (though the music is almost as irritating as a Fran Drescher rap album would be):

 

https://youtu.be/G8b4lmB2t68 

 

 

[pedrosura]

I haven't tried this, but depending on what you want to do a possibility would be to use camera tracking, which is supported in HitFilm, AfterEffects, and Blender, among other applications. It's normally used to insert CGI objects and characters into a real-world scene or actors into a CGI world, but Moviestorm could be used in the same way as the real world.

 

Suppose for example you want to insert a complex animated object (a robot, dinosaur, spaceship, or whatever) into a Moviestorm scene without getting into modding. In that case you could create the scene in Moviestorm and then use a compositing/CGI program to track the camera motion. Blender and HitFilm would allow you to animate a 3D model that you could insert into the scene. I confess I've never looked into whether AfterEffects allows you to animate models as fully as HitFilm does, but it may.

 

Alternatively, suppose you want to to insert Moviestorm characters into a CGI world created in another program. In Moviestorm you would make the background a chromakey color and place flat objects on the floor and wall planes that would be different from the main chromakey colors (for example, blue versus green, or dark green versus light green). Then in your CGI software you would motion track the background objects in order to general a camera solve.

 

You will generally want to avoid zooming during a shot, and depending on the software it would help to make note of the lens focal length (the zoom setting under the lens button on the Camerawork view when you select a specific camera angle).

 

The sample lessons from Hollywood Camera Work's Visual Effects for Directors can serve as a decent introduction to the concept of camera tracking: http://www.hollywood...ampleclips.html

 

There are plenty of YouTube tutorials on camera tracking in various software. Try searching for

 

    <software name> camera tracking

 

Here's a good example using Blender (though the music is almost as irritating as a Fran Drescher rap album would be):

 

https://youtu.be/G8b4lmB2t68 

 

 

i am familiar. the problem is that MS does not output camera position or allow you to input camera position in a manner compatible with software likemVUE. i wish it could take an x,y,z position for the camera.

[Ben_S]

Well, if you don't mind editing the movie.mscope for a saved movie in a text editor, the <targetOffset> node for a <keyframe> basically includes an x,y,z coordinate for the camera keyframe position.

 

The default starting camera has <targetOffset>0.0 0.0 2.0</targetOffset> for example, which means two metres above the origin (which is in the centre of the grid, about 10cm down - the grid is not at height zero).

[bongoman]

Ben offers the most precise solution.

However, I have used camera tracking in Blender successfully to augment scenes created in MS, by switching textures and rendering a second "duplicate" scene from MS for tracking purposes.

I made a movie demonstrating how here:


This is a few years old when I was learning camera tracking, so I made some mistakes, but it is a good overview of the process, and is applicable for any tracking software not just Blender.

Hope this helps.

[pedrosura]

Ben offers the most precise solution.

However, I have used camera tracking in Blender successfully to augment scenes created in MS, by switching textures and rendering a second "duplicate" scene from MS for tracking purposes.

I made a movie demonstrating how here:
140003225

This is a few years old when I was learning camera tracking, so I made some mistakes, but it is a good overview of the process, and is applicable for any tracking software not just Blender.

Hope this helps.

very cool. you just gave me the idea that I can go both ways. create in MS and then get the shot in VUE or get the shot in view and then in MS.

I really wished that MS could take an x,y,z, t camera position and pointing angle in addition to the user moving thecamera around. then it would be perfect.