How to generate a full game asset suite with Leonardo.Ai


In the dynamic realm of game development, an increasing number of professionals are turning to Leonardo.Ai as a comprehensive solution for game asset generation. Leonardo facilitates various stages of game development, offering tools for rapid asset ideation, multiple iterations, detailed finessing, prompt engineering, background removal, consistent style asset generation, and even 3D texturing.

In this case study, we delve into how an indie studio leveraged our platform to develop a complete game titled “BlueFor”. This tabletop board strategy features a map with four types of assignments, a solo campaign encompassing eight missions, 72 characters, and 180 pieces of equipment. Currently, the studio is also using Leonardo.Ai to produce several expansion sets for the game, potentially tripling the number of assets.

We will also examine the ideation process behind BlueFor’s game mechanics, the Design Matrix which ensures game consistency, and the three-step method for generating style-consistent game assets for any gaming project.


Crafting game mechanics: the hidden step before generating any game asset

Utilizing the Design Matrix for Consistency

The three-step process for asset generation

Step #1: Ideation and Conceptualization

How to create consistent style variations using Leonardo AI

Step #2: Refinement and Detailing

Step #3: Finalization and Integration

Final words

Game asset from BlueFor generated with Leonardo

(Game asset from BlueFor generated with Leonardo)

Crafting game mechanics: the hidden step before generating any game asset

Before venturing into game asset creation, it’s essential to first define the fundamental mechanics that will direct each stage of the game. In this instance, the story unfolds through eight distinct missions, each characterized by its own ambiance, storyline, objectives, challenges, and rewards. These elements form the core of the gameplay experience.

Leonardo’s user-friendly interface enables swift experimentation with diverse assets, each designed to enhance these mechanics. That’s why the game studio began with the basics of asset generation, starting with a straightforward level design using a simple Excel spreadsheet:

First step in level design

This initial sketch demonstrated the strategic placement of game elements, ensuring each mission presented a unique and engaging challenge. Here are a few examples:

  • In the game, one of the weapons is a shotgun. Due to its shorter range compared to a sniper rifle, thoughtful ammo distribution across levels was necessary. This meant that the placement of cover and other assets had to be thoughtfully configured for effectiveness. 
  • Additionally, the game starts with the character at a basic level, equipped with limited weaponry. Strategic placement of assets in this scenario was crucial to create a game that was both challenging and engaging.
  • Another critical aspect of asset development was enemy progression. Important factors included setting the right number of enemies per level to prevent overwhelming players, integrating secret passages and hidden areas for added intrigue, and ensuring the dynamic interaction of all elements to consistently challenge players with puzzles and obstacles.

Once these basic mechanics were established, Leonardo.Ai facilitated the quick removal of the grid from the map with the Image-to-Image tool, allowing for smooth continuation of the design process:

Map with removed grid

After producing a clean map, deeper considerations came into play before asset design. Some game studios overlook these aspects and consequently spend considerable time and resources on unused assets due to misaligned game mechanics. BlueFor’s developers avoided this mistake by following this process.

This highlights the importance of aligning art with game mechanics – without this synergy, the art loses its effectiveness and purpose. In this particular case, the developer meticulously crafted and carefully analyzed each level prior to embarking on asset creation. The outcome of these efforts is evident in the levels, as shown in the image below. We will revisit the final results later in this case study.

Final level design

After these crucial considerations, attention was turned to integrating these foundations with the design of the characters and equipment. Leonardo.Ai’s versatile toolkit enabled the game studio to ensure that each character and item was not only visually distinct but also functionally congruent with the game’s mechanics (more in this later).

Utilizing the Design Matrix for Consistency

With the game mechanics firmly established, key decisions about the actual game were necessary. This is where the Design Matrix played a crucial role. The Design Matrix serves as a blueprint, a comprehensive guide aiding in the translation of myriad ideas and concepts into a cohesive final product. Essentially, it ensures the coherence and unity of the game’s various elements, encompassing everything from visual style to gameplay mechanics.

Consider the following screenshot example. All elements here, ranging from the textures of the pavement and sky to the old houses and the gun, have been generated using Leonardo.Ai. Notice the consistency in their overall design language:

Screenshot with complete level details generated with Leonardo.Ai

A frequent challenge in game development is the temptation to incorporate a multitude of diverse ideas and concepts into a single game. Without a structured approach, this can result in a disjointed and confusing player experience. 

Imagine, for instance, if the image above included a character with a completely different skill set and design language. It would disrupt the game’s coherence. The next example, also created with Leonardo.Ai, contains all the right elements for quick production. However, merging it with the first example would be a significant misstep.

Complete character generated with Leonardo.Ai

To avoid such pitfalls, implementing a Design Matrix is crucial. The Design Matrix orchestrates the interaction between characters, environments, and game mechanics, ensuring that each element both complements and enhances the others. This process allows developers to thoughtfully select and integrate various ideas, facilitating seamless interaction within the game’s universe. The effectiveness lies not only in the ideas themselves but in their synthesis to form a cohesive whole.

Below is an example of the design matrix used in this particular game:

Design Matrix example

Notice how the Matrix reflects decisions made from the initial game design, such as the number of characters, weapons, and assets. It also outlines their functionalities compared to others. For instance, a police officer character would have specific attributes like cost, health, and personal traits such as speed. This character would then be strategically placed on the map, in alignment with the overall game design.

In a game with diverse character types and environments, the Design Matrix is instrumental in determining how these characters interact with each environment. It guides how the game mechanics will adapt to these interactions and ensures that all elements are in line with the game’s overall storyline and objectives.

In the case of BlueFor, the Design Matrix ensures that all shock weapons function consistently, maintaining game cohesion. Additionally, within the same game, one unit category is the Red team, which possesses a unique aesthetic art style characterized by specific prompts like “fire, explosions, red, dragon, and mercenary.” All these aspects are systematically organized within the matrix.

Leveraging this, the developer was able to swiftly create consistent style assets using Leonardo.Ai. Here are actual examples from “BlueFor,” showcasing assets from both the red and blue teams, that demonstrate this principle:

Red team character generated with Leonardo.Ai

(Red team)

Red team character generated with Leonardo.Ai

(Blue team)

The three-step process for asset generation

After laying the groundwork and defining the basic assets and their features, the next phase focused on crafting consistent style assets for the game. This procedure was divided into three distinct stages, each designed for expansive creative exploration and precise refinement as necessary. These stages were crucial in producing the comprehensive game asset suite, encompassing 8 missions, 72 characters, and 180 pieces of equipment.

Step #1: Ideation and Conceptualization

The creative journey began in this initial phase. Utilizing Leonardo.Ai, the game studio engaged in brainstorming sessions, conceptualizing various ideas for characters and equipment. The platform’s capability to rapidly generate a multitude of iterations provided a vast array of creative options, making it invaluable in this phase for establishing the game’s visual and thematic essence.

Here is an example of an image generated with Leonardo.Ai that was successfully incorporated into the final production:

Final image utilized in the game

The process involved sketching out initial ideas, drafting concepts for character designs and equipment, and developing the first visual representations. The aim was not only to create a diverse range of designs but also to ensure these designs were cohesive with the game’s world, storyline, and gameplay mechanics.

Sometimes, this process involves identifying ideas that ultimately won’t be used in the final production. In this project, numerous iterations of the game design were explored before settling on the final version. However, some of these experimental designs have become integral to the development of other game ideas, potentially emerging as finished products in future projects. 

Below is an example of wolf character designs that, despite their consistency, were not included in the final production:

Wolf example generated with Leonardo.Ai

(Wolf 1)

Second wolf example generated with Leonardo.Ai

(Wolf 2)

Third wolf example generated with Leonardo.Ai

(Wolf 3)

How to create consistent style variations using Leonardo AI

There are numerous methods to create consistent style assets using Leonardo.Ai. The platform provides a range of tools that facilitate prompt and image ideation, multiple ControlNet features, as well as various presets that can be adjusted to achieve a specific look and feel for any image, among other capabilities.

Below are just some examples that demonstrate how to achieve a variety of results, depending on your specific needs and creative process:

  • To quickly engineer and test your prompt ideas, utilize Leonardo’s Realtime Gen. It allows you to experience blazing fast text to image generations that evolve as you type. Reiterate, rephrase and refine while you create iterations on the fly.
  • Realtime Canvas is Leonardo’s implementation of the latest innovations in Image-to-Image transformation technology. This feature is designed to convert drawings into stunning artworks nearly instantaneously, revolutionizing how users interact with AI for image generation. We will continue to explore it below.
  • If you are looking to add a specific style to multiple images, then Elements is a great choice. Elements offer a highly efficient and resource-friendly approach to nuanced image generation. It allows precise and intricate control over the style and visual outcome of generated images with the simple addition of dynamic modifiers.
  • Sketch to image specializes in composition. It employs a sketch or scribble as a preliminary guide for the final image. It is particularly suited for projects where a specific arrangement or composition is essential, acting as a foundational blueprint to shape the outcome according to your artistic vision.
  • Pose-to-image focuses on the recreation of human poses. Pose scans a reference image, seeks out human (or similar) figures, and attempts to replicate their poses for the resultant image. It’s an invaluable tool for specific character positioning.

In this case, the game studio employed Realtime Canvas to rapidly create iterations and test design ideas before applying them on a larger scale. Realtime Canvas turns each brushstroke into detail in near real-time as you sketch. You can also interact with the Canvas through clicks, drags, and moves. It intelligently waits for your actions to complete before instantly transforming your sketches into detailed artwork.

Realtime Canvas

The developer’s use of Realtime Canvas was instrumental in quickly visualizing and refining design concepts. This tool’s ability to transform sketches into polished images allowed the game studio to see their ideas materialize almost instantly, significantly streamlining the creative process. By leveraging this feature, the team could experiment with various artistic directions without the time-consuming need for manual redrawing or extensive editing.

Another essential tool during production was Leonardo’s Realtime Gen. Realtime Gen generates images as you type, and instantly makes changes based on your prompt and settings.  This feature allows for greater focus on the artistic vision, reducing the need for additional settings or navigating complex interfaces. Its ability to select style presets is particularly beneficial, especially during the ideation phase of game asset development.

Realtime Gen

After generating multiple ideas, the developer efficiently scaled the images and embarked on another round of ideation. Let’s delve into how this process works.

Step #2: Refinement and Detailing

Once the initial concepts were established, the developer concentrated on refining these ideas. This process involved a meticulous examination of each element in the game’s design. This phase was not only about enhancing the visual appeal but also about ensuring that each asset fit seamlessly within the game’s environment and narrative context.

For instance, when working on character designs, the developer paid close attention to aspects like facial expressions, clothing textures, and accessories, ensuring they were consistent with the character’s backstory and role in the game. Similarly, when designing environmental elements, such as landscapes and buildings, they ensured that these components resonated with the overall mood and setting of the game, adding to the immersive experience.

One of the strategies used during this phase was using Leonardo’s Image-to-Image feature. Sometimes, the developer would employ stock images with commercial license to quickly generate ideas as starting points for the game. The following is an example of an initial image utilized for this purpose: 

Stock image utilized to generate further iterations with Leonardo.Ai

After some iterations with prompts and the background removal tool, the developer got the following image. Later, it became one of the cards of the game:

Tank with background removed

Also, during this phase, the game studio quickly generated between 50 and 100 iterations of each asset before arriving at the final result. This approach allowed for a wide array of options to choose from, facilitating a thorough and diverse selection process. Each iteration brought a new perspective or a subtle variation, enriching the creative pool from which the final designs were selected.

The process of generating such a large number of iterations was made feasible by Leonardo’s advanced AI capabilities. The speed and efficiency of the platform enabled the rapid production of variations, each offering a unique take on the initial concept. This was particularly beneficial in finding the perfect balance between creativity and coherence with the game’s thematic elements.

The studio utilized Leonardo’s Alchemy V2 Diffusion XL model during production, which enabled high-quality outputs with exceptional prompt adherence. Below is an example of a final asset that was successfully included in the final production:

Final asset that was successfully included in the final production

Step #3: Finalization and Integration

In the final stage, the assets underwent thorough polishing and were prepared for integration into the game. This process involved making final adjustments to guarantee that each asset was not only visually appealing but also seamlessly aligned with the game’s mechanics and narrative.

During this phase, significant emphasis was placed on how each asset functioned within the game’s environment. An important aspect of this stage was narrative integration. The game studio made sure that every asset significantly contributed to the storytelling. This involved ensuring character designs accurately reflected their roles and personalities, and environmental elements effectively enhanced the game’s setting and mood.

For example, one of the earlier images underwent several modifications before reaching the following result:

Tank that was successfully included in the final production

After meeting the required standards and passing all tests, the assets were methodically integrated into the game. This thorough and deliberate process guaranteed that the final product was a cohesive and immersive experience, with each asset playing a crucial role in bringing the game world to life. In this instance, the image was integrated as a card in BlueFor’s gameplay:

Tank in the final game card

The finalization and integration phase was also pivotal in refining the assets to ensure they functioned cohesively within the game’s ecosystem. It served as a critical step in transitioning from artistic conception to the playable reality of the game. For example, the following image underwent several revisions before achieving its final form:

Image of the soldier generated with Leonardo.Ai, before finessing

This result was attained using Leonardo’s Canvas Editor, which provided the capability to make detailed adjustments to the images, eliminate unwanted elements, or even create new content with advanced in-painting and out-painting techniques. It also allowed for cutting, resizing, and modifying various aspects of the image. The following is an example of the image that was finalized for production:

Soldier in the final game card

Final words

At Leonardo.Ai, we are incredibly excited to witness the creation of complete game asset suites using our tool. The journey of BlueFor, from initial concept ideation to final asset integration, showcases the immense potential and versatility that Leonardo offers to developers. With its comprehensive suite of tools, the platform facilitated the creation of a richly detailed and cohesive game environment, significantly enhancing both visual appeal and gameplay experience.

The successful completion of this project stands as a testament to the powerful synergy between technology and creativity. It highlights how AI can streamline the game development process while elevating artistic vision to unprecedented heights. As the gaming industry continues to evolve, we at Leonardo are committed to remaining at the forefront, aiding game studios in achieving remarkable results.

The following image is a testament to the multiple assets generated with Leonardo.Ai, seamlessly integrated into actual gameplay:

Final map asset

We believe this case study will serve as an inspiration for other game developers, whether they are working on indie projects or large-scale productions. The adaptability and efficiency of Leonardo make it an ideal choice for a wide range of creative needs, from rapid prototyping to detailed asset creation.

Our heartfelt thanks go to the BlueFor team for kindly sharing the assets used in this case study. If you’re interested in exploring more about the game, please follow these links:

BlueFor Free Digital Early Access on TTS, powered by STEAM:

BlueFor Discord Community to interact with the developers and find games online.

Source link


Lütfen yorumunuzu giriniz!
Lütfen isminizi buraya giriniz