Embedded composite shots and virtual cameras.

NullUnitNullUnit Website User Posts: 791 Just Starting Out
OK, lets see if I can explain my question well enough. I have been learning and now understanding the power of embedding composite shots but an aspect of it is making my brain hurt. I was working on a composite shot that had many layers and an animated camera move. I put some related layers into a separate composite shot and then embedded it. The new composite shot has the 3-d camera and the cameras animation data in it from the original composite. And, of course so does the original composite shot with the now embedded composite. Now, my question is: Does the camera info in the new composite have any bearing on the camera info in the original composite that contains the embed. Now my brain hurts again... @-)

Comments

  • RobinRobin Website User, HitFilm Beta Tester Posts: 1,671 Enthusiast
    I think I understand what you want to know... Sad truth: I have no idea :D Let's see if any of the HitFilm staff has an answer :D
  • AxelWilkinsonAxelWilkinson Staff Administrator, Imerge Beta Tester, HitFilm Beta Tester Posts: 5,242 Staff
    If you set the embedded comp to 3D Unrolled, then each of the layers will assume its original position in global space, so their position relative to the camera will remain intact. If the embedded comp is set to 3D, then its just a 3D plane, and everything it contains will be mashed into a single plane. The "Unrolled" aspect of the 3D Unrolled option is that each of the layers contained in the embedded comp will still behave as if it were a unique layer in the mater comp in which the unrolled comp is embedded.
    Composite shots can be tricky to explain. :)
  • NullUnitNullUnit Website User Posts: 791 Just Starting Out
    OK, that makes sense. The part I'm having trouble wrapping my head around is that I created my embedded comp after I created a camera move. And there are 3-d objects in my embedded comp, so there is a 3-d camera in my main comp and my embedded comp that both have the same camera keyframe animation data. does the camera keyframe data in my main comp take precedence over the camera data in the embedded shot? Is the keyframe data in the embedded comp ignored? I ask because I wonder what would happen if I need to change the camera move.
  • AxelWilkinsonAxelWilkinson Staff Administrator, Imerge Beta Tester, HitFilm Beta Tester Posts: 5,242 Staff
    So you animated the camera, and duplicated the camera into the comp when you embedded? And then you have another camera, in the master comp, performing the same movement? I would think you would want to have one camera performing the move, rather than two. Are there other 3D objects in the Master comp, in addition to the embedded comp containing the camera? If so, then the moving camera should probably be in the master comp. If not, then keep the camera movement in the embedded comp, and then use a stationary camera in the master comp. This isn't a situation I've actually tried, though, so I'm not 100% sure the best way to handle it.
  • NullUnitNullUnit Website User Posts: 791 Just Starting Out
    edited March 2012
    I'm breaking new ground! :D I didn't do it on purpose, though. Yes, there are 3-d particle effects in the master and embedded comps. When I created the embedded comp I had already made the camera move in the master comp. So, when HitFilm created the new comp (which contained 3-d particle effects), it automatically created a new camera and copied the cameras keyframe data from the master comp into the new embedded comp. I'm scared to play with the camera keyframe animation because I'm happy with the camera move, but after I'm done I think I'll play around with them and see how they effect each other...
  • AxelWilkinsonAxelWilkinson Staff Administrator, Imerge Beta Tester, HitFilm Beta Tester Posts: 5,242 Staff
    Well, playing back the master comp should tell you immediately whether there is an issue with the cameras, yes? It seems odd that HitFilm would use the camera movement dat in both places. IF you selected the camera along with other layers when creating the comp, then the movement should have moved into the embedded comp, and a new camera would be created in the master comp, as necessary, but with no animation. If you didn't select the camera layer, then a new camera should have been created in the embedded comp, with no animation, and all the camera keyframes should have stayed attached to the camera in the Master comp.
  • SimonKJonesSimonKJones Moderator Website User, HitFilm Beta Tester Posts: 4,450 Enthusiast
    Hopefully to clear this up a bit...
    If you have a 3D unrolled embedded comp, any camera information inside that embedded comp is entirely ignored. The 3D objects inside that comp are then subject to the cameras in the parent comp. If you're using a 3D unrolled comp, it's best to think of it as a convenient way to group layers together, rather than anything else.
    If your embedded comp is set to 2D plane or 3D plane, it then uses the cameras inside the embedded comp. With 2D/3D plane embedded comps, it's best to think of them as behaving like videos. In this case, the cameras inside the embedded comps are entirely separate to any cameras in the parent comp.
    If you want the same camera move in the embedded comp that's fine, you just need to make sure the cameras are the same. If you change the camera move at a later date it isn't a huge problem - simply copy the new camera and paste it into the relevant embedded comps. I actually have a really good example of doing just this in a tutorial we put up a few weeks back:

    Take a watch of that.
    Hope that helps!
  • AxelWilkinsonAxelWilkinson Staff Administrator, Imerge Beta Tester, HitFilm Beta Tester Posts: 5,242 Staff
    Thanks for the clarification Simon.
  • NullUnitNullUnit Website User Posts: 791 Just Starting Out
    edited March 2012
    Ah, I see. Thanks for that as well!
    Kind of a mind bending concept, at first.
  • SimonKJonesSimonKJones Moderator Website User, HitFilm Beta Tester Posts: 4,450 Enthusiast
    One way to think about it, which may or may not help, is that it's a bit like motion control rigs for real cameras. The whole reason they were invented back in the 70s was to enable VFX artists to film multiple shots of different miniatures, but all using the exact same camera move.
    So for Star Wars, they'd shoot the Falcon, then each individual TIE fighter, all separately but using the same motion controlled camera setup. The camera rig was controlled by a basic computer which was able to mechanically replicate the exact move each time. This meant that all the different plates of the different models could then be composited together to look like one single shot, with one single camera move. They kinda peaked with this in the asteroid field sequence in Empire and the Endor battle in Jedi.
    The multiple-cameras-in-embedded-comps concept that we're discussing in this topic is, conceptually, exactly the same thing. The only difference is that these are virtual cameras shooting multiple CG scenes, rather than physical cameras shooting multiple miniatures.
    I've no idea if that comparison helps or makes it even more complicated. :P
  • NullUnitNullUnit Website User Posts: 791 Just Starting Out
    It does clear things up.
    To throw more fuel on the fire: I was experimenting with how different layers interact with each other when they are imbedded or not and I ran into a similar issue with 3-d lighting. I noticed that 3-d lighting used in an unrolled 3-d embed does not work in the master composite, but if I switch the embed to just 3-d or 2-d then those lights will work. I'm guessing that thats because the 3-d lights in the master composite take precedence over the lights in unrolled embedded comps since when they are unrolled they become part of the over-all 3-d environment. I'm realizing that there is a bit of a method to knowing/ pre-planning how and what to put in embedded comps.
    Another twist: Motion blur. If I have an embedded 3-d particle effect I can turn on motion blur on the effect itself or on the embedded layer in the master comp. Which do you guys recommend. One or the other? Or both?
    I'll post the video clip that I'm working on to learn this stuff soon.
  • AxelWilkinsonAxelWilkinson Staff Administrator, Imerge Beta Tester, HitFilm Beta Tester Posts: 5,242 Staff
    I believe you are correct with regard to the lighting. I would recommend selecting the lights within the embedded comp, Copy them, then Paste them into the master comp, to see if that gives the result you are after. It is definitely good to do some planning beforehand as to how you are going to put the scene together, so you can plan for your comps accordingly, although once you get started there's always the chance that your methods will change during the process. I know mine often do. I need to do some more testing with 3D unrolled comps when I get a chance, and maybe put together another tutorial focusing on them specifically and explaining them more.
    I think the motion blur has to be turned on for the layer. If you turn on motion blur for the embedded comp, that is only going to calculate based on any animation you create for the actual comp layer; it won't factor in the movement of individual layers within the comp. But clearly, unrolled 3D comps aren't really among my areas of expertise.
  • NullUnitNullUnit Website User Posts: 791 Just Starting Out

    I think the motion blur has to be turned on for the layer. If you turn on motion blur for the embedded comp, that is only going to calculate based on any animation you create for the actual comp layer; it won't factor in the movement of individual layers within the comp.
    Ah, yes. That makes sense!
    Thanks for all the info!
    A "part 2" tutorial for embedded comps would be great. I learn a lot from the video tutorials since I'm a visual/audio oriented person.
  • SimonKJonesSimonKJones Moderator Website User, HitFilm Beta Tester Posts: 4,450 Enthusiast
    Yeah, it does seem that an 'Advanced comps' style tutorial would be useful, covering some of the more advanced and niche cases that have come up in this topic.
    By the way, it's great to see users like yourself really starting to get to grips with the software, Darth. With any new software there's a learning curve, which is why we've been pumping out these tutorials. Every time somebody makes something impressive in HitFilm it helps us because it shows off what the software can do, so we're really keen for people to keep learning and pushing the envelope.
Sign In or Register to comment.