Rivendell – Miniature or CG ?

I’m sure most of you have seen both the Lord of the Rings Trilogy and the first Episode of The Hobbit : an unexpected journey. Some of the action happens in Rivendell, where the big boss of the Elves, Elrond, lives a happy and long life with his very old friends and family. Rivendell is supposed to be a very peaceful place, with nice elvish architectures, precious materials, a permanent golden hour light during the day, waterfalls, trees, and birds.
For the first movie, Weta Workshop decided to build Rivendell as a miniature (bigature). The work on this model was amazing. There was a lot of details everywhere, nice painting, and I must say, I’m a big fan of it. Here is how it looked like.

One big advantage of using a miniature like this one was for the lighting. You can have a very natural sense of light by reproducing the lighting from the set at the scale of the miniature and using real lighting sources. All the lights and shadows will be real, and the way the different materials and colours will react to these lights will assure the best realism you could hope. Of course, it might need a bit of additional painting work here and there to fix small issues, or add few more details. But it’s a pretty solid base to have, and I’m convinced that most of the time, it probably gives more realism to the shot than using a Full 3D approach.

Here are a couple of examples of DMP made by Dylan Cole for LOTR, based on the miniature. The lighting on the miniature drives the rest of the environment.

At the end, I personally think that Rivendell works very well in the shots. It feels natural to me. The most important on this kind of environment is to have the lighting correct on the building. For the rest of the DMP, mountains and trees, it’s easier to make it work. We don’t need to be “that” precise. That’s why we can keep a DMP approach.
Here are some examples of Rivendell views from Lord of the Ring.

Of course, on the other hand, you have loads of limitations using miniatures. First, and probably the most important, once it’s built, you can’t move anything around any more, and it term of flexibility it’s a nightmare. You want to be able to move this tower a bit on the right, or rotate this house of 15 degrees. On the top of that, you may have to break some part of it, or make them move, which is almost impossible to do right if we use a miniature. We still can use miniatures for explosions for example, like the White House shot in ID4, but it means you can do it just one time and it’s finished. We never would do it today, as we want to be able to control everything, and be able to tweak all the parameters, smoke, fire, debris, etc.

In the Hobbit, Rivendell has been build full CG, with a lot of different assets textured, lookdeved, and put together in the scene. Then, you can place your lighting and a TD can control everything, layout, lighting, textures, to render the most beautiful image and directly pass it to comp, or give it to a DMP/Env Artist who will do a bit of paint over.
It’s quite good as a method, except that this time, everything is based on the look dev and the lighting.

Let’s be clear, it still gives very good results, and in term of flexibility it’s perfect, as you can do pretty much everything you want. Let’s have a look at the full CG Rivendell of the Hobbit.

It still look very nice !

So the question is : is it better to use a miniature and real reaction of a real light onto a painted model, even if you don’t have as many controls as you’d like, or is it better to go for a full 3D build, and recreate the reality by tweaking shaders, textures, and lights ? It’s an open question I think. If I perfectly understand that production wise, it’s better to be able to change everything until the end, from an artistic point of view, it may be interesting to think about it.

The answer is probably to do both, which is what we often do now. If we 3D scan a miniature like Rivendell and use this scan as the base of the 3D model, we still can develop a 3D asset, fully textured and ldeved. Then for a shot, you can put some light around your miniature, and use it as a reference for the lighting TD to match as close as he can the real photography of the Environment. Of course it cost a bit more, but at the end, we may have the final result faster.

Feel free to share your thoughts about this!.

5 thoughts on “Rivendell – Miniature or CG ?

  1. I always preferred miniatures because they feel true. Indeed comp is more complex but in the Tolkien cinematic universe, they work much better. I really disliked most of The Hobbit in the narration, the visuals, and the 48 fps. I think that the hybrid solution you mention (especially if you LIDAR scan the prop) is the only way to bring the touchable feeling of hand-crafted material and then play with it in your scene.

  2. The problem is not necesserally the comp. When you use a miniature, if the layout doesn’t suit you, you can’t really do anything. We’re creatives, we want/need to be flexible. Using a miniature force you to use it as it is. Then you still can cheat it, but it’s more limited.
    The probleme with a Hybrid approach is that you double the cost. You go into build twice, one in a real build, and one in CG build… It’s not wallet-friendly…

  3. Using miniatures mean that you have to plan what you want to shoot. CG feels like it’s easier to change your mind. The reality of changing you mind may still cost as much as a reshooting a miniature, or more but since it’s out of sight and in the computer, it’s less obvious. Film costs skyrocketed after VFX went digital, not before.
    The cost of changing one’s mind is not cheap, either way.

  4. I think one of the forgotten words in film today is commitment. And digital in my opinion is the one to blame on the lack of commitment. When digital was not around and everything was shot on film, the room for a change was minimum. Exposure had to be right bang on, a bulb couldn’t explode during shooting otherwise it will ruin the exposure, focus had to be right. People were committed to make it work as there was no more film left to reshoot.
    Todays work is different with digital: one shot had the focus wrong? Don’t worry, we’ll do it again we can reshoot as much as we can.
    In my opinion that attitude led to one thing: Movies are not made in the shooting anymore, they are made in the visual effects houses, where the “let’s start all over again” has become the norm.
    Commitment for me means to know what you want. But that doesn’t mean you can’t be flexible. Flexibility to me means to allow room for a change, and it doesn’t mean throw away 3 months worth of work because I like to try another thing, which leads to money problems.
    If you shoot a miniature and the layout doesn’t suit you, think about it before building the miniature what is the layout you want (commitment). Once you committed to that and If later the layout doesn’t suit you say in 30 or 40%, you can change until it does (flexibility). But if the layout doesn’t suit you in more than 70%, you didn’t have commitment. Starting again=money problems, that’s when the wallet starts complaining.
    You are not more creative because you have the ability to change things until the very end. Looks like if you don’t turn thing upside down you are not creating. I don’t share that way of thinking.
    Creating a miniature is not only useful for the expensive task of rendering, we the matchmovers, could benefit from having that miniature scanned as we can use it as a LIDAR for photogrammetry to align the cameras and solve our matchmoves easier.

    I started with magic 20 years ago as a hobby, I am 27 now and I still continue doing it. Magic for me is the aesthetic exploration of a mystery, and my goal when i do magic is to make people experience that mystery. To do that we use a tool we call “illusion” which starts by presenting a fact that is real and ends turning that real fact inside out. The contrast between both we call it the “effect”, which is what the audience sees.
    We start from a fact that is real, and it conforms to the physics of the world they know (the miniature). Then we start to slowly add the lies so that reality mixes with other stuff that doesn’t necessarily need to be true (the 2D DMP); by doing this in an intelligent manner you led the audience to believe the unbelievable.

    In my opinion the best visual effects are the ones who are invisible. Yes, I am biased because I am a magician, and that’s our holy grail for us. For me the best way to create a film illusion is by combining stuff that is real with stuff that is not. The audience are the ones that blur the line between the two not letting them to see where one ends and the other starts, they have a very strong force attracting them, (they see stuff that meets the laws of nature they know).

    In conclussion, If I was to make Rivendell, in order to create it I want to believe that exists. And to show that, yes I would use miniatures and 3D and 2D DMP.

    Just my opinion.

  5. The hybrid solution of model scanning and rendering is what was done in the hobbit. I have been talking with concept designer John Howe and he explained that scanning highly detailed models gives a much more realistic approach now. The two main problems with the miniature .. one if the use of 3D . In stereo miniature falls apart because of the sheer size of the camera. Secondly the problem is the scale and the weight. As good as your miniature is , it can never account the same weight of the actual scale… For example The Black Gate of Mordor was an impressive piece of art , but it lacked the weight.. Also time is an important factor.. Jackson didn’t have enough time for the miniatures which would take 2-3 years of proper shooting. Also shots are limited. As far as i am concerned the hobbit rivendell is far better , more dynamic and a sense of the whole vista is there.. Also in Harry potter the same thing happened with Hogwarts where in the last few parts they replaced the miniature with the CG model. CG model’s are highly detailed , textured and lighting can be controlled properly.. gives more artistic freedom. Blends better with the environment with proper compositing.. As good as the lord of the rings miniatures (big atures) were i am quite into CG detailed models.. I think computers come most into use in these situations.. For example Laketown was a classic example.. The CG model was so much detailed that it felt almost like a chopper shot over a real location.. The shot before they leave laketown we see the whole town .. Its so perfect.. and mixed with fog and mist and layers of smoke … creates a great environment..

    In my honest opinion everything technique in movie is valid.. They use 48 frames and stereo there was no way a miniature was going to hold.. Also the scale factor is essential..and except rivendell and laketown most of the other locations were interior worlds compared to Lord of the rings. so no point building a miniature of goblin town or woodland realm when you are panning within the environment.. the flexibility of A 3D model is much more..

    And the thing is the CG work was neatly done , nicely done and it should not bother anyone who does not specifically want miniatures only . All techniques are suitable. I think the 3D modelling and rendering and compositing was brilliantly done in the hobbit.. And it gives a richer aesthetic to it , a bit more fantastical than LOTR which i think the hobbit demands.. There is a sense of whims and fantasy which i love. On top of that you have great sets like rivendell , bag end , rhosgobel , parts of goblin town , laketown … I think it was a very nice amalgamation

Leave a Reply

%d bloggers like this: