Learning from CAD / BIM and issues of modeling

(Polygon Pictures Inc. / Studio Phones)
We plan on brushing up the content of this article and making revisions as we hold future sessions.
This document is written with The Film Production Pipeline and Artist Techniques 6 in mind.
translated by PPI Translation Team

Discussions on this topic are continued in a special series held separately by Kobe Studio Seminar for Studies on Building Information Modeling.

Kobe Studio Seminar for Studies on Building Information Modeling
http://wwwmain.h.kobe-u.ac.jp/kobe_studio_seminar/talks_sb.html

■Overview


The practice of storing assets over a long period of time has finally begun to make appearances in film production. Until recently, studios have maintained quality by using trick shading and a variety of creative solutions created for each phase of production. However, as of late reusing assets has become more effective, and procedural generation has become viable on a large scale. Nowadays, procedures for large-scale generation can be archived as “templates,” and changes and additional functions can be added to them at any time.
However, if we do not carefully manage the thing that can be called a production’s blueprint, then the focus of production will shift too much towards preserving ideas, which will limit the amount that a studio can expand its capabilities.
Similarly to BIM (Building Information Modeling), we consider the versatility of each element used to construct assets in film production, abstract them, and give them a structure. We use this structure while working on implementing the long-term storage of assets, to construct procedures, and to combine these structures and other assets. Considering these applications, one gets the feeling that current film production is, even if only partially, approaching the developments and direction taken by software used in architecture.
There is no question that the ideas driving film production and those driving design or supervision in architecture are fundamentally different. Film production has developed using sets for filming to support fictional stories, and contains the pleasure of preserving momentary changes on film. In this document, I would like to write about long-term preservation and abstraction and their uses in production, about which I hope to hold discussions with participants in future sessions.
This document is written as a springboard for discussions in future sessions and as such its contents are rough and may wander, so I request readers’ attention and caution when reading it.


■Stored assets and procedure modeling


Until recently, it was most common for completed model asset data or data that is in a reusable state to be divided into props and stored. However, using procedural modeling presents a different set of challenges. For example, when using a round edge shader to render curves on edges, procedures have to be adjusted after-the-fact and managed to generate models that do and do not use this shader. Similarly, the way that assets are stored long-term differs greatly from storing assets themselves when one has to foresee after-the-fact changes, like adjusting multiple components’ placement or generating models.
At the same time, parts of stored assets that can be re-generated have to be saved, and procedures and structures used for assets also have to be stored. These are both factors that are different than the way assets have been stored up until now.


■Utilization for design process etc.


When discussing the design process with directors or within teams, it would be useful to organize and catalogue design-related know-how to make it easier to reuse. It is an approach that is taken relatively often by many studios that create in-house services for pre-production. This is useful because assets can be archived in order to reuse them for future procedural modeling. It can also be used to create several varieties of designs to be shown to the client so design work can be continued while receiving feedback. In this case, it functions as a way to generate starting points that facilitate smooth communication.
Documents created when preserving assets as production models may prove to be extremely important for production, in that they simultaneously archive know-how as described previously. They also allow one to consider reusing and improving the methods themselves.
Doing so should prove to be very effective at improving the average quality of assets produced by our studio.
I would like to talk discuss the expansion of pre-production services from these perspectives during future sessions.


■Replacement of assets and modification of design


Procedurally modeled assets are stored as abstract attributes, for example “door,” in a format that includes their design processes. After deciding only to which parts of the asset to apply the attribute “door” and determining the asset’s overall layout, it can be used in lighting design and production planning while its design work is conducted simultaneously. When it becomes time to finalize the production model, select the design method for the “door” attribute by selecting a single procedurally modeled asset, apply it, generate an asset as the production model, and render it. At this point, procedural model and production model assets are preserved in storage, which supports the pipeline, for use in all productions. Functions are added to make assets searchable, which expands the system for use in future productions. Old assets are taken out of storage and designs are changed in order to determine whether they can be used in other productions. In future sessions, I hope to discuss with participants whether procedurally modeled assets can be used in a manner similar to how the development environment’s repositories have been constructed until now.


■Long-term series production and the assets


In productions that turn into long-running series, large changes are often made to the designs of even main characters of animated series over time, and even live-action series are often continued after changing actors. While the same is true in part for the data for film sets and interiors that show up in a series, most series tend to maintain design features when creating sets. Many shows tend to pass down information like layouts or maps for the placement of buildings or the arrangements of streets.
When creating a new series, the design process in which these are created from scratch is very important. However, it is also important to consider updating existing designs in order to make production efficient and meet budgets and deadlines. Storing assets for procedural modeling can also be thought of as an important step, and it may be helpful to store all of the documents used during production and assets that were created during the design process as well.
An important point to recognize as a studio is that the know-how that artists accumulate is a tool they use for their own ends. Further, we should recognize that if we do not create an infrastructure that lets us reference the know-how used in previous productions, and lets us try our hand at creating new designs, we will face challenges in making production efficient. If we are able to incorporate generating procedural assets when optimizing these processes, we should expect an increase in our design skill as a studio at least to some extent.
In future sessions, I hope to call upon the experience of our participants from film productions studios and game development companies, and to hold in-depth discussions regarding archiving design processes and other things that studios are attempting to take on.