Over the past decade, one of the biggest problems in the game industry is the explosive growth in the production cost of AAA video games. In 2000, a dozen people could make a great PlayStation success. Today, a new iteration of any console or the PC game franchise requires hundreds of artists, designers and programmers working for several years. Even occasional games, such as hidden objects or puzzles, take several months to make the concept at the Apple App Store or Google Play.
One of the most important costs in game production is the creation of assets. The players want the content of their games not only entertaining but unique. The models, the textures, the skins, the exotic locations of the character, various power-ups at stake, must all be carefully personalized, refined, laborious by artists, which is incredibly expensive.
Another problem is scaling. Gaming studios must publish numerous updates and DLC asset packs to remain competitive. This is particularly relevant today, when monetization trends all go to integrated purchases, such as character skins, accessories, vehicles, emotes and cards. Talent is rare, and one of the solutions uses play outsourcing studios from low -wage sites, but it is not an easy task to find a reliable and create long -term collaboration.
This is why studios are always looking for technologies that could help reduce the cost of developing games. Recent progress in generation of neuronal image generation, such as Dall-E, Midjourney and StreeDiffusion, give a certain hope that the realization of this dream may not be so far.
Enter Get3D from Nvidia, the last entry into this field, announced in a recent post here
This AI model has been formed using 2D images and is able to generate 3D shapes with high fidelity textures and complex geometric details. The range of objects that can be created is quite wide: vehicles, characters, animals, people, buildings, various outdoor spaces that could be combined in whole cities with their inhabitants. Export formats are suitable for the most popular graphic software, allowing easy importation of 3D yields and motors, and can be smooth in existing artistic production workflows.
How much is we close to implementing this model and similar models in the development of real world games? “The technology is very impressive and we are sure to see many art assets generated in games. They will not replace a creative team of human artists, however, not in the predictable future, because AI always requires a well thought out contribution to produce good results, and a careful post-production work to make it coherent and polite “, according to an opinion of experts from the expert, Absolist Game Art Studio. It seems that, as in many other areas, in Game Dev, automatic learning applications seem to be very effective assistants for artists and designers, rather than a substitute for creativity and human skills.
