Using Models as proxies for a scene

Hello!

So, working on learning Carrara, and thinking ahead to possible projects I'd like to work toward creating, and because of how the program handles models and scenes separately, it occurred to me, that I could pretty much "block out" a scene in Assemble mode with generic shapes... cubes, cylinders, whatever. If I'm understanding it right, I could then go into the model editor for each of those proxies, delete those objects,and create the actual final object to go in their place - or perhaps use the proxies as a size reference to be removed later.

Is that an ideal workflow, or am I not quite grasping the Model <> Assemble relationship just yet?

Comments

  • JonstarkJonstark Posts: 2,738

    It's an interesting idea, but I'm not sure of the value.  Carrara can handle a lot without slowing down down, and for most scenes it's probably easier/quicker to just use the actual models you want to use themselves, rather than using a group of primitive cylinders/cubes as stand ins.  If your assembly room scene starts to get sluggish as you're moving around in it, you can always make unnecessary things in your scene invisible (I do this sometimes with dynamic hair on characters, just to make sure my preview scenes don't get sluggish as I'm tooling around with the camera.  Or rather I used to do this, on my laptop, the new desktop is much more powerful and I haven't really seen any sluggishness anymore).

  • WsCGWsCG Posts: 391
    edited September 2016

    Well, it wasn't so much for performance, as it was to establish the overall setup/composition using the stand-ins, and then the items could be replaced with final objects. I was thinking in terms of something where you'd want to have things set up at a certain scale, or if having stand-ins would help give a sense of relation between different objects' locations, etc. 

    I got the idea from game design, and how a lot of times, they'll use stand-in objects in the game-engine, often with a texture saying "replace me" or something like that. Then, over time, the artists replace the stand-ins with their final models. It's a nice way to get things arranged as you want them, allowing for play-testing and pathing and such, without having to wait for the final objects to be complete.

    For me, it's a good way to plan out a scene and establish what's going where, scale, etc. before I even start creating the actual final objects.

    I more or less answered my own question after posting this, though. I think it should be do-able.

    Post edited by WsCG on
  • MarkIsSleepyMarkIsSleepy Posts: 1,496
    edited September 2016

    You can indeed do this - just make sure that the cubes/cylinders/whatever you are using as proxies are created using vertex models and use either control-D or replicators to place them.  Then altering one will indeed update all the others which are linked to the same master object. I've actually done this before and it works very well.

    For example, say you were going to fill a classroom with desks but you were not yet sure of the final layout or design of the desks.  You could insert a vertex object, then in the modeling room insert a cube about the size you think the desk might be.  Return to the assembly room and use control-d to duplicate your desk object (or a replicator to quickly place a whole bunch in neat rows).  Then later you can select any one of those desk-cubes and go into the modeling room. When you do, it will ask if you want to edit the master or create a new master.  If you choose edit the master then any changes you make, for example rescaling the cube and adding legs to it, will propagate to all the other copies.

    Replacing a proxy this way would be a little more involved if you wanted to replace all the desk-cubes with a completely different model (you'd have to open the different model in the modeling room, copy the whole thing, exit the modeling room, select your proxy, go back into the modeling room and paste your new object back in) and would not work if you were trying to replace them with a rigged model (I don't think anyway - I haven't actually tried it).

    My usual workflow is something more or less like:

    1. sketch it out on paper and make some notes on possible light/color/composition
    2. block in the scene with primitives or simple place-holder models
    3. setup my basic composition
    4. setup my basic lighting
    5. start replacing parts of my blocked in scene with real objects (this step usually includes a mix of looking through my library to see what I already have and modeling/UV mapping/texturing my own items) <- this is where I would use the technique above to start replacing my block proxies with real models
    6. futz around with lighting and composition
    7. add finishing details
    8. render,
    9. postwork
    Post edited by MarkIsSleepy on
  • WsCGWsCG Posts: 391
    MDO2010 said:

    You can indeed do this - just make sure that the cubes/cylinders/whatever you are using as proxies are created using vertex models and use either control-D or replicators to place them.  Then altering one will indeed update all the others which are linked to the same master object. I've actually done this before and it works very well.

    For example, say you were going to fill a classroom with desks but you were not yet sure of the final layout or design of the desks.  You could insert a vertex object, then in the modeling room insert a cube about the size you think the desk might be.  Return to the assembly room and use control-d to duplicate your desk object (or a replicator to quickly place a whole bunch in neat rows).  Then later you can select any one of those desk-cubes and go into the modeling room. When you do, it will ask if you want to edit the master or create a new master.  If you choose edit the master then any changes you make, for example rescaling the cube and adding legs to it, will propagate to all the other copies.

    Replacing a proxy this way would be a little more involved if you wanted to replace all the desk-cubes with a completely different model (you'd have to open the different model in the modeling room, copy the whole thing, exit the modeling room, select your proxy, go back into the modeling room and paste your new object back in) and would not work if you were trying to replace them with a rigged model (I don't think anyway - I haven't actually tried it).

    My usual workflow is something more or less like:

    1. sketch it out on paper and make some notes on possible light/color/composition
    2. block in the scene with primitives or simple place-holder models
    3. setup my basic composition
    4. setup my basic lighting
    5. start replacing parts of my blocked in scene with real objects (this step usually includes a mix of looking through my library to see what I already have and modeling/UV mapping/texturing my own items) <- this is where I would use the technique above to start replacing my block proxies with real models
    6. futz around with lighting and composition
    7. add finishing details
    8. render,
    9. postwork

    Ah nice! So you're already using a similar workflow, then. Cool!

    Do you find it helps a lot with your overall scene creation working that way? Any drawbacks you've found to it?

  • MarkIsSleepyMarkIsSleepy Posts: 1,496
    edited September 2016

    I think it does help me because I can quickly set up and see my basic composition before deciding exactly what models I want to use - my example of the desks isn't something I've done specifically, but I've used it to switch out and test different stacks of books, different styles of pillar, different window frames and lots of other things.  If I am just noodling around without a clear direction it really doesn't help at all, but if I have a basic idea of how my final scene will look but haven't finalized the details in my mind it can really save a lot of time and work switching out models I decide I don't like in a particular scene.  

    It can be nice to clearly delineate your composition and think about it more abstractly before you get bogged down in the details of a scene.  Same for being able to set up your basic lighting - sometimes you can really see if it is setting the mood or telling the story you want with just basic shapes in the scene.

    As for downsides, mostly just that it does take a little planning ahead in that you need to have an idea of when to create a new master model proxy and when to use another copy of one you've already put in.  You also need to remember to clearly name them so you can remember which featureless block is supposed to be a table and which one is meant to be standing in for a rocking chair. wink It doesn't lend itself well to those times when you are in the mood to kind of freestyle your scene as you go.

    Post edited by MarkIsSleepy on
  • WsCGWsCG Posts: 391

    Indeed. 

    Well, everything you're saying is very much in line with my general thinking. You've just done a far better job of describing/explaining it than I did lol.

    Excellent. Well, I've still got the training wheels on, learning my way around the program, so I'm not quite ready to start building entire scenes, yet. But I'll definitely be keeping that workflow in mind when I get to the point where I'm ready to put together something larger, and more "specific", rather than freestyling, as you put it.

  • DartanbeckDartanbeck Posts: 21,326

    My pal, evilproducer, made this awesome freebie for Carrara, Fantasy Village Terrain (ShareCG requires a free membership to download)

    He originally built the scene using some really cool freebie medieval buildings, which he couldn't redistribute with his freebie, so he made up little Proxies just like you described.

    He has a key that shows which of the buildings goes where, if the use wishes to collect the same ones - might as well... they're excellent and they're free! LOL

    In my opinion, any kind of messing around with models is good for the flow. The more we mess around with stuff, the better we get at all of it.

  • Inagoni has a plug in to make it easier to do what you are trying to achieve. I own it, but I don't think I've ever used it. I may have to give it a try.

    http://www.inagoni.com/?page_id=449

    It's also part of the bundle avaialble on Daz

    http://www.daz3d.com/advance-pack

  • Mitovo,

    That is a good way to go! Exactly what I do... Though I do all my modelling in 3dcoat,I render in Carrara. I usually lay out my "shot" in Carrara  and lock the composition early .This way when I am sculpting in 3dcoat I dont waste time making stuff I will never see! (mainley the back of things! Or hands that are going to be out of the frame.) And because my computer has just celebrated its 8th birthday it struggles with 3dcoat. So occasionly I have to split scenes in 3dcoat and import them to Carrara as seperate bits. So I slowly assemble my scene ,replacing half finished pieces with newer pieces as I go .This way I can start the lighting process early as well. So its sort of non linear,sometimes I finish lighting before I have even finished the models.

Sign In or Register to comment.