Rome Reborn

Available

 * https://www.flyoverzone.com/rome-reborn-flight-over-rome/

People

 * Project Director: Bernard Frischer, founder of Flyover Zone, Inc.
 * Assistant Project Director: Alberto Prieto
 * Director of 3D Modeling: Lasha Tskhondia
 * Director of Historical Art: Mohamed Abdelaziz
 * Director of Technical Art: Devin Good
 * Director of Technology: Jeremiah Stevens, CTO of Flyover Zone, Inc.
 * Director of Administration: Nathanael Tavares, CEO of Flyover Zone, Inc.
 * Scientific Advisors (active in 2023): Karolina Kaderka, Paolo Liverani, Russell Scott, Pier Luigi Tucci
 * Scientific Advisors (1996-2022): Carla Amici, Heinz Beste, Amanda Claridge, Diane Favro, Fulvio Cairoli Giuliani, Lynn Lancaster, C. Brian Rose, Barry Strauss, Mark Wilson Jones

Description
In September of 1974, Bernard Frischer was newly arrived as the Prix de Rome Fellow in Classical Studies at the American Academy in Rome. Along with other fellows and residents, he was taken on a field trip to the Museum of Roman Civilization in EUR/Rome. There he saw the physical model of ancient Rome created at a 1:250 scale under the direction of Italo Gismondi from the 1930s to the 1970s. Frischer immediately had the idea of finding a technological solution that could use the Gismondi model to create the illusion of walking down the streets of the ancient city. In the period 1974 to 1996 he explored various possible solutions. The first was proposed during the viewing of Gismondi’s model: UC Berkeley Professor Donald Appleyard was a Resident at the Academy and participated on the trip (https://en.wikipedia.org/wiki/Donald_Appleyard). He told Frischer about work at the UC Berkeley Environmental Simulation Laboratory to composite physical architectural models with videos of existing urban spaces (Bosselmann 1984). This solution might have worked, but there were several practical impediments, including the massive size of the Gismondi model and the fact that it could in all likelihood never be transported to Berkeley, nor could the equipment at Berkeley be shipped to Rome. So, Frischer next considered producing a videodisc on the model of the Aspen Moviemap (see Naimark, no date).

In 1986, he was invited to give a paper a small conference at Apple Computer on the future of computing in various fields. Since it was clear to Frischer that the Rome project would require many years, Frischer framed his vision for Classical Studies within the context so-called “Fifth-Generation Computer Systems,” the main features of which included Artificial Intelligence and a natural-language user interface (Moto-Oka 1982; https://en.wikipedia.org/wiki/Fifth_Generation_Computer_Systems). Frischer’s contribution to the conference was published two years later (Frischer 1988). In it, he described what he at that time called "Project Cicero," which several years later was renamed Rome Reborn. The new name was chosen as an expression of homage to the founding text of the field of Roman topography, Flavio Biondo's De Roma Instaurata (1444-1448; Raffarin-Dupuis 2005, 2012; https://en.wikipedia.org/wiki/Flavio_Biondo). Key elements of Project Cicero included a digital recreation of ancient Rome derived from the physical model of Gismondi and the use of AI-infused avatars playing the role of Romans with whom users could speak and who could illustrate and explain to users various aspects of Roman life. The idea was to create an educational application that could make the study of ancient Rome and its culture more readily comprehensible and engaging to young students.

After the conference, Apple asked Apple Fellow Alan Kay (see https://en.wikipedia.org/wiki/Alan_Kay) to collaborate with Frischer on a feasibility study of Project Cicero. The two determined that the Gismondi model did not lend itself to the proposed solution since the representations of buildings and monuments in the physical model of ancient Rome lacked a sufficient degree of detail to sustain a close look at a 1:1 scale. For the moment, the idea of digitizing the Gismondi model was put on hold, and Frischer pursued the solution of using CAD software which he saw in 1990 implemented on a small scale by IBM at an exhibition on reconstructing Pompey (see anonymous 1990).

By a stroke of good fortune, Frischer knew architect William J. Mitchell, a colleague at UCLA and a pioneer in the development of CAD (Mitchell 1977; https://en.wikipedia.org/wiki/William_J._Mitchell). Through Mitchell and Franklin Israel, a friend of Frischer’s from the American Academy in Rome who was also a professor of Architecture at UCLA (https://www.getty.edu/research/collections/collection/113YHW), Frischer learned about the work of their colleague Diane Favro, who taught Architectural History at UCLA. She shared Frischer’s interest in applying CAD to the recreation of ancient buildings, and her area of specialization was Roman architectural history. With the help of Favro and other UCLA colleagues, Frischer started the Rome Reborn project on December 1, 1996 at a conference held at the American Academy in Rome. It was to be an international initiative whose goal is the creation of 3D digital models illustrating the urban development of ancient Rome from the first settlement in the late Bronze Age (ca. 1000 B.C.) to the depopulation of the city in the early Middle Ages (ca. A.D. 550). Following the advice of the project's scientific advisory committee, the project team decided that 320 CE was the best moment in time to begin the work of modeling. At that time, Rome had reached the peak of its population, and major Christian churches were just beginning to be built. After this date, few new civic buildings were built. Much of what survives of the ancient city dates to this period, making reconstruction less speculative than it must by necessity be for earlier phases.

From the beginning, Frischer divided the archaeological data to be modeled into two classes. Class I elements are those features of the city whose exact location, phasing, design, function, and name are well known and studied. Consisting of an estimated 250 features, examples include the Colosseum and Pantheon. Class II elements are all the other features that are known with a greater or lesser degree of certainty. Examples of the latter include the many apartment buildings of the city, very few of which have left behind any trace visible today in Rome. The best sources for Class II are the two late-antique Regionary Catalogues, which give the building stock of the city region by region. By following the indications of the Regionary Catalogues, an urban model of late-antique Rome can be quantitatively accurate, but--apart from the Class I elements--it cannot be accurate in detail.

The urban model (version 1.0) was first developed at Frischer's laboratory at UCLA, where he was a professor in the Department of Classics from 1976 to 2004. Version 1.0, owned entirely by the Regents of the University of California, was launched by Frischer and Rome's Mayor Walter Veltroni at a press conference in 2007 (see https://www.wired.com/2007/06/rome-reborn/). The Class II elements were 3D scan models derived from the great phyiscal model of ancient Rome created from the 1930s to the 1970s under the direction of Italo Gismondi. The complex scanning project was led by Prof. Gabriele Guidi (Guidi, Frischer, et al. 2005). The Class I elements were modeled by hand using Multigen Creator software. They included the Basilica of Maxentius, Colosseum, Forum of Julius Caesar, Ludus Magnus, Roman Forum, and Septizodium. Frischer raised all the funds used to create version 1.0 from industry sources (including Intel and Microsoft), philanthropists, and the Andrew W. Mellon Foundation. (Despite several attempts, Frischer never succeeded in obtaining support from the National Endowment for the Humanities.) All students and faculty who worked on the project through Frischer's Cultural Virtual Reality Laboratory at UCLA were paid for their work, and outside expert consultants were also compensated with honoraria.

It turned out that the idea of using scan models for the Class II elements was flawed: as Alan Kay and Frischer had already surmised in the 1980s, when the Gismondi model was scaled up to 1:1, as required by the project's aims, all its defects could be seen. Walls that should have been smooth were pocked with holes or disfigured with bumps. Lines that ought to have been straight were jagged. Not infrequently several sides of the physical model of a Roman building lacked any architectural detail. The time and money it would have taken to remedy all these flaws was so great that Frischer decided to discard all the scan data deriving from the Gismondi model and to find a different solution. He had learned about the work of doctoral student Pascal Mueller, who was writing a dissertation at the Federal Institute of Technology Zurich on the topic of procedural city modeling (Mueller et al. 2006). Frischer contacted Mueller, who quickly agreed to apply his software solution, the CityEngine, to the problem of Rome Reborn's Class II features. The result was presented in August 2008 at SIGGRAPH, where Rome Reborn was the featured project and given a booth 110' x 30' in size at the entrance to a hall of the Los Angeles Convention Center (see https://www.youtube.com/watch?v=zxvIlx7mHXs). By this time, Frischer had founded a company (Frischer Consulting, Inc.) that took over responsibility for developing and financing Rome Reborn. The new Class II features were owned by the company, which also licensed the Class I features from the Regents of the University of California. The decision to move the project to a private company was based on discussions on sustainability in the mid- to late 2000s initiated by the Andrew W. Mellon Foundation. "Sustainability" meant several things ranging from maintaining the long-term operability of new digital creations to ensuring ongoing research and development. These and related sustainability activities required a continuous flow of financial support.

In 2007, Frischer received a grant from the National Science Foundation (no. 10104956) to study how best to achieve sustainability for projects of virtual heritage such as Rome Reborn. Known by the acronym "SAVE" (Serving and Archiving Virtual Environments), the study resulted in the recommendation to start a scientific journal for the publication and archiving of scientific 3D models of heritage objects at all scales (see Frischer 200;6 Koller, Frischer, and Humphreys 2009). Frischer found no interest among American university presses to support such a journal, and so he eventually turned to private publishers. In 2012, Elsevier agreed to Frischer's proposal to back a new journal called "Digital Applications in Archaeology and Cultural Heritage" (https://www.sciencedirect.com/journal/digital-applications-in-archaeology-and-cultural-heritage). The first issue appeared in 2014. This was in the period when the concept of Open Access was spreading through academic publishing. At first, Elsevier resisted instituting a policy of Open Access. Finally, it agreed, but the terms were onerous: the Article Preparation Charge (APC) was to be $3,000 per article across all of the company's publications. Realizing that a fee at this level would constitute a hardship for most of the intended cadre of authors whose needs the journal aimed to fulfill, Frischer and his co-editor-in-chief Gabriele Guidi argued for a far lower APC. The publisher refused to consider differential charges by field, so on September 30, 2016 Frischer and Guidi resigned from their editorial positions and founded a new, Open Access journal called "Studies in Digital Heritage" (www.studiesdh.org). Supported by the Library of Indiana University and by gifts raised by Frischer, since its founding the journal has been able to waive an APC.

For Rome Reborn version 2.0, launched in 2008, all this was in the future, but the problem of sustainability was one with which Frischer had to struggle on a daily basis just to keep his laboratory and the Rome Reborn project afloat. So, in 2008 Frischer founded Frischer Consulting, a company with the dual of rebuilding the Rome Reborn urban model using its own resources and finding commercial ways to monetize it. Hence, starting in 2009, the company set about replacing all the Regents' intellectual property with new 3D models of the same Class I monuments. It also started to add many more of the ca. 250 features belonging to Class I. The primary motivation in doing so was to move away from the FLT format of MultiGen Creator and to take advantage of all the affordances developed for OBJ. This meant that the Class I features could be presented with a higher degree of photorealism and could be more flexibly migrated from one scope of application to another. Frischer Consulting, therefore, did not renew its license agreement with the Regents and by 2018 owned all the intellectual property used in the Rome Reborn model. Version 3.0 was released that year. It was used for one of Frischer Consulting's first publications of virtual tourism called "Rome Reborn: Flight over Ancient Rome." This product was at first supported only on the virtual reality headsets of Oculus and HTC Vive. Soon, support for personal computers and mobile devices was added.

Version 3.0 of the city model left much to be desired in the degree of photorealism it supported. For a variety of technical reasons, the CityEngine models had to be displayed in their lowest level of detail (LOD). Meanwhile, the Class I elements were not presented using all the latest lighting features supported by software packages such as 3D Studio Max and Maya. So, Frischer decided to start work on Rome Reborn 4.0 through a successor company he founded in 2022 called Flyover Zone (www.flyoverzone.com). Version 4.0 was released in November 2023. It marks a significant step forward in terms of photorealism of the Class I and II elements (now displayed in their highest LOD), and it also represents a quantitative advance in the number of Class I elements of the city that are incorporated with detailed hand modeling. These now include the following:


 * Aqua Claudia
 * Ara Pacis
 * Arch of Constantine
 * Arch of Septimius Severus
 * Arch of Titus
 * Atrium Vestae
 * Aurelian Wall and gates
 * Basilica of Maxentius and Constantine
 * Basilica Neptuni
 * Baths of Agrippa (exterior only)
 * Baths of Caracalla (including interiors and sculptural decoration)
 * Baths of Constantine (exterior only)
 * Baths of Decius (exterior only)
 * Baths of Diocletian (exterior only)
 * Baths of Nero (exterior only)
 * Baths of Titus (exterior only)
 * Baths of Trajan (exterior only)
 * Capitoline Hill (complete)
 * Circus Flaminius
 * Circus Maximus
 * Flavian Amphitheater (Colosseum)
 * Imperial fora
 * Imperial palaces
 * Ludus Magnus
 * Mausoleum of Augustus
 * Mausoleum of Hadrian
 * Mons Testaceus
 * Montecitorio Obelisk
 * Naumachia of Trajan
 * Nymphaeum Alexandri
 * Odeon of Domitian
 * Pantheon (including interior and sculptural decoration)
 * Porta Praenestina
 * Porticus Liviae
 * Pyramid of Cestius
 * Roman Forum (including the interiors of the Basilica Aemilia, Basilica Julia, and Curia Julia)
 * Rotunda of St. Andrew in the area of the former Circus of Gaius and Nero
 * Septizodium
 * Stadium of Domitian
 * Temple of Diana on the Aventine
 * Temple of Claudius on the Caelian
 * Temple of Hadrian in the Campus Martius
 * Temple of Matidia in the Campus Martius
 * Temple of Minerva on the Aventine
 * Temple of Serapis on the Quirinal
 * Temple of Spes in the Campus Agrippae
 * Temple of the Sun in the Campus Agrippae
 * Temple of Venus and Rome
 * Templum Divorum
 * Theater of Balbus
 * Theater of Marcellus
 * Tomb of Marcus Vergilius Eurysaces

With the publication of version 4.0, Frischer stated that the project had reached its initial goal of making the Rome Reborn model ready for deployment in teaching. From the original vision of 1986, still missing were periods other than late antiquity and avatars infused with AI. Regarding the latter, Flyover Zone has developed a prototype solution with a natural language interface, but it has not yet implemented it across all its virtual tours. The company currently has no plans to offer coverage of other periods of Roman urban history. However, Flyover Zone has already published educational virtual tours of the Basilica of Maxentius, Baths of Caracalla, Pantheon, Roman Forum, and a Flight over Ancient Rome. More tours of individual sites as they appeared in 320 CE are planned for release in 2024 and beyond.

Dissemination of the model occurs through Yorescape, a streaming multimedia platform built on Unity and developed by Flyover Zone specifically to serve the needs of virtual tourism. Yorescape supports mobile devices (iOS/Android), PCs (Mac/Windows), and virtual reality headsets (Oculus and HTC Vive). Its features include the integration of text, images, maps, video, panoramic photographs, and free-roamable 3D models. Yorescape is available with a free one-week trial subscription to individuals and enterprises (https://www.flyoverzone.com/rome-reborn-flight-over-rome/). Those wishing to extend their use are asked to pay a modest annual fee to help sustain projects on Yorescape such as Rome Reborn (the platform also hosts Athens Reborn, Baalbek Reborn, Egypt Reborn, Hadrian's Villa Reborn, and Mesoamerican Reborn).

Meanwhile, as a public service, Frischer's companies have made free versions of the model available on Vimeo and Youtube, and they have also cooperated with Smarthistory (https://smarthistory.org/ancient-rome/) to create videos with commentary explaining the features seen. As of November 19, 2023, these videos have been seen by over 3.8 million people.

As illustrated by several of Frischer's own publications (Frischer, Abernathy, et al. 2006; Frischer 2008; Frischer, Pollini, et al. 2017; Frischer 2018; Frischer and Massey 2022), the Rome Reborn urban visualization has the potential to support new empirical (or what Frischer calls "simpirical") research in which the Rome Reborn model makes it possible for scholars to make observations and run experiments that would otherwise be impossible short of true time travel.