Set-jetting in an Outsourced Country

Romania’s landscape and folklore have become significant in global pop culture and its industry, yet their origins remain obscured. In this interview, Simone C Niquille explains the thinking and process of developing the Landscape Mode scenographic installation in Turn Signals—Design is not a Dashboard, including her research into film location databases, 3D libraries, and AI-generators to create a proxy representation of an outsourced country.

How would you describe the essence of ‘Landscape Mode’?

Simone C Niquille: Online travel agency Expedia confirms that ‘set-jetting’, travel inspired by movies and TV shows, is the top source of vacation influence, outranking social media. With a similar ambition, the Tourist Board of Ireland has partnered with Ubisoft of the Assassin’s Creed game series to attract tourists. The game is famous for its historical settings, so much so that a ‘Discovery Tour’ mode has been added for players to take virtual tours of ancient Egypt, Greece and Ireland. 

As a popular film location, Romania has experienced its share of set-jetting. Following the release of Tim Burton’s hit Netflix series Wednesday, tourism to Romania experienced a boom, immersive bus tours to the various film locations fully booked. The series, although set in Vermont USA, is filmed across numerous locations and film sets in Romania. In Burton’s words: ‘Trying to make Romania look like Vermont was an interesting challenge, but we felt like we found lots of new locations.’ The series star Jenna Ortega commented on the location’s accuracy: ‘One of the places we went to was Sinaia and everything looked photoshopped.’

The Romanian landscape, cities and landmarks have acted as proxies for New Jersey, the Appalachian Mountains, Paris, Moscow, the American South and more in various films and television series. Sometimes the reasons are financial, as the country offers substantial incentives to support the local film industry. Other times the authenticity of Romanian film locations is superior to ‘the real thing’. In the movie Cold Mountain, set in 1860s North Carolina, the Carpathian Mountains offered certainty of snowfall and less modern-day infrastructure to remove in post-production, compared to the actual Cold Mountain in The US. The same ‘wilderness’ attracted Panasonic to shoot several camera commercials in the country. 

It’s a fine line between the actual life in a place, and projection of a cliché, a stereotype produced through media. Similarly, Romania is no stranger to pop culture references with werewolves, vampires, giants, and ogres deeply rooted in local culture and lore. Over time these characters have been absorbed by mainstream entertainment and inspired entire IP franchises, such as Shrek, Dracula, The Wolfman, the Twilight series, Buffy… an entire catalogue of gothic fantasy. 

What methodologies did you use to investigate this compelling insight?

SN: An overview of the proxy locations in Romania and the reasons for choosing a location were established by searching through film location databases and production company portfolios. A map was created of films, series, advertisements and games produced in Romania, and whether the locations served as proxy for another location (many of course were filmed in Romania as the story is based on its landscape). 

To create the tapestry, as well as some of the characters, a collection of digital artefacts that relate to either the Romanian landscape and lore or the proxy landscapes Romania stood in for, were sifted from the 3D asset libraries of Sketchfab and Polycam—the largest repositories of downloadable 3D scans online. Online libraries of scanned 3D data are a curiosity cabinet filled with reality fragments. Uploaded by geology research institutes, real estate agencies, tech enthusiasts, game designers, land survey companies, the scans cover a wide spectrum of intentions and applications. Surprisingly all geographies were represented, even if only a 3D-scanned tree stump from Appalachia, a stone from Vermont. Clear documentation of file sources and metadata were recorded while collecting them and are shared on the maps in the exhibition space.

Recently launched text-to-3D asset AI generators ShapE and LumaLab imagine 3D were also used as data sources. LumaLab has an online accessible archive of all user-generated 3D models. This archive was used to search for prompts such as ‘werewolf’, ‘vampire’, ‘ogre’, and ‘Romania’ to give a glimpse into the popular imagination (or better said, data) of these folkloric characters turned pop culture content. ShapE was used to generate characters, prompting the universal regurgitation of the formerly local and mythical.

How are the AI-generated images to be understood in relation to the proxy representation of Romanian landscape and culture?

SN: For a text-to-3D generated AI to function, massive amounts of training data need to be available. Prompting it to generate a ‘werewolf’ does not in fact produce The Representation of the folkloric character but instead is an amalgamation of all the training data that is associated with the label ‘werewolf’. To end up as training data, information needs to be easily and widely available. Training datasets are often scraped from online sources rather than created from scratch and offer a strange portrait of what a particular something is according to the data available online. The werewolf then is neither a ‘good’ or ‘bad’ representation, that binary becomes trivial. Rather it is a glimpse into the available data, which offers a popular version, a homogenised imagination. Prompting an ogre generates William Steig’s Shrek, popularised by the 2001 DreamWorks feature animation.

How did you incorporate all these elements into the photo background for visitors to experience the outsourced land at the Turn Signals exhibition?

SN: In the 3D software Blender these characters and landscape fragments were assembled into a tapestry that is installed in the Turn Signals exhibition. The installation is inspired by tourist souvenir snow globes that cram all the ‘must-see’ attractions into the tiny space of a globe not larger than the palm of one’s hand. It borrows from the aesthetic of a film set, with a backdrop to simulate the landscape while characters roam in the foreground. The characters are printed on re-board, a thick cardboard used to construct temporary expositions, and fair booths. Exhibition visitors are invited to pose and take photos with the characters: Welcome to ‘Landscape Mode’—a place where everything looks eerily familiar, a ghost of binges past, a faint memory of childhood fairy tales, a composite of content.