Virtuos Games worked closely with Ubisoft to bring Assassin’s Creed 3: Liberation from its PS Vita origins and bring it to modern day consoles as part of Assassin’s Creed 3 Remastered. We spoke to game producer Paul Loumouamou, and technical director Xavier Rozé about the process.
So the project was to bring Liberation tomodern consoles – using the PC version as a basis?
Xavier Rozé: Yes but not only. We also wanted to refresh the visual quality of the game in order to reach the current standard. As PS4 and Xbox One hardware are more powerful than the original devices, we have the opportunity to push the capabilities of the engine. A first HD version was originally released on PC with improvements from the PS Vita version, and we wanted to push those further for this remaster.
Paul Loumouamou: This is also why we released a PC version as well. This allowed us to offer the visual improvements of this remaster to PC gamers.
What are the key considerations when budgeting for such a conversion?
XR: One key consideration is to adapt the game to the current player audience, which has significantly evolved since the original release. We wanted players of the original game to rediscover it by amazing them on aspects they did not expect, and at the same time, we wanted to offer them a title which fully matches their original memories. This is why we were focused first on the visual improvements, but also on gameplay and UI adaptations as well as fixing bugs from the original.
PL: Another key consideration is that there are cross adventures between Connor from Assassin’s Creed 3 and Aveline, the heroine of Assassin’s Creed Liberation. It was important to keep the entire story as a whole, so it made a lot of sense to release both 3 and Liberation together in a single bundle.
How big was the team, and were they all based in a single location?
XR: At Virtuos, three studios worked together: Paris, Chengdu and Shanghai. Each had a dedicated topic. Paris, who was the lead studio, was responsible for all the visual and gameplay improvements regarding development and technical art. Chengdu was in charge of the target platform development and UI improvement, whereas Shanghai worked on the asset improvements. I think that around 40 to 45 people worked on the project at its peak.
PL: It is one of our key strengths here at Virtuos, we have a very solid co-development culture, both internally and with our clients.
What were your ambitions for bringing what was a Vita title to machines many times more powerful?
XR: Quite a large amount of improvements have been made to AnvilNext, the engine that runs the Assassin’s Creed series. For us, this was a wonderful source of inspiration to demonstrate what could be achievable for this new version of the game. We wanted to improve the experience by offering a more advanced immersion for the player. This has been our chance to demonstrate what we are capable of, by implementing the most appropriate features in the game.
PL: Assassin’s Creed Liberation had the potential to integrate visual features that can only be implemented on current gen consoles or PC. Our ambition was to give this title a new life with key ingredients that can be found in recent games.
What areas did you concentrate on in order to improve the title?
PL: The graphics were something we spent a lot of time on. One of our main concerns was the lighting, on which we made a lot of changes. We tried to bring a more realistic side to the game’s atmospheres, such as in the bayou area, while still staying true to the initial artistic direction.
As for the UI/UX, our main focus was to modernise it. We implemented UI features such as, for example, the mini-map which is an essential feature in games nowadays. We also redesigned some of the UI, like achievements, to reach a more up-to-date UI overall.
Were there any gameplay changes?
PL: Yes, we upgraded some of the gameplay with a particular focus on two main features. We made a free aiming system which is more representative of the latest Assassin’s Creed games. That introduced a lot of new issues we had to fix one by one due to the fact that we were modifying one of the core features of the game. However, it was a very worthwhile implementation, since it added more depth to the shooting gameplay.
We also made improvements to the canoe controls, making them more accurate. In addition to the visual improvements we made on the water, and specifically on the canoe’s interaction with the water, it made the canoe much more realistic.
XR: We had long discussions with Ubisoft to determine what the best angle to improve the original game could be. We wanted to optimise our effort in the parts of the game that were the most visible to players. For instance, one of the key points of improvement is the water rendering in the Bayou, a swampy map of the game that the player visits frequently. The original rendering was limited by technical constraints, and we have been able to make significant improvements by implementing a water rendering system adapted from the ocean system of Assassin’s Creed: Black Flag. Thanks to this new implementation, the waves are far better rendered on the water surface, and accurately reflect the moves of the canoe driven by Aveline.
Another example is the parallax occlusion mapping that we have implemented to bump up the grounds of the levels, especially in New Orleans and New York maps. As the ground usually fills a third of the screen, this is the kind of asset improvement which requires a very specific work, but is largely applied and obvious throughout the game.
We have also focused on frequently displayed UI such as weapon inventory or the mini-map to refresh their ergonomics and make them closer to what the players are now used to.
Were the original assets up to the task, did you have to work on them at all?
PL: We kept the same meshes as the PC version, but we upgraded the textures to a higher resolution. We also added new objects in the environment to increase the immersion of the player into the game.
XR: The most efficient task for us when it came to assets was to improve resolution of textures and details of some materials for the ones which actually needed to be refreshed.
The game is built on the Anvilnext engine, for which there are already PS4/XB1 releases, did that make the process easier?
XR: Actually, using existing PS4/XB1 implementation of other AnvilNext-based games has been quite useful to design our low-level architecture and quickly get a first version running on these consoles, even if each game requires a dedicated implementation to fit its own source code anyway. This has helped us to faster implement some multithreading optimisations as well, to benefit from new hardware capacities. Moreover, this has offered us more time to polish the improvements of the other game parts.
What was your relationship with members of the original development team and how did this help you achieve the final result?
XR: Numerous people have worked on various parts of the game and the engine! Fortunately, we have built a strong partnership with Ubisoft, who has provided us with support to explain some code parts or process to comply with in order for us to avoid wasting time.
PL: We had the opportunity to talk with the original developers of Assassin’s Creed Liberation for our specific questions. We also had weekly calls and a close relationship with Ubisoft Barcelona who was developing Assassin’s Creed 3 Remastered at the same time.
Did the project go as you were expecting? What areas were problematic? Where did you manage to do more than you expected?
XR: The project went as expected and even more! It is rare enough to be pointed out. This is the reason why we dug into R&D of PBR (physcially based rendering) implementation, whereas it was not expected at the beginning of the project. Earlier we wished to improve lighting on the materials to emphasise better specularity in the game, but the previous attempts were unsuccessful or not noticeable enough. So, we decided to first move to a rough implementation of physically-based rendering.
The jump was not obvious, as this is a shift of lighting paradigm and we did not want to break all the game rendering to rebuild everything from scratch after. So we first prototyped a hybrid implementation of no-PBR and PBR models, which was able to display untouched objects and PBR-ised objects in the same tiny scene to demonstrate that both kinds could cohabitate. This allowed us to validate that this strategy could be extended to the whole game in order to improve only the materials we wanted and keep therefore the production under control.
PL: If I had one thing that I would recommend to be wary of during a remaster production, it would be changes in the art direction.
It’s very easy to change the original intention of the art direction without intending to. We had weekly reviews with comparisons of both games (our remastered version and the original version) to be sure that we were still in harmony with the initial art direction. All our artists always had a second (or third) monitor with the original version.