Neural Networks in 2023: Pros and Cons

Contents:

• Artemy Kotov
• Asker Unaev
• Artem Makarov
• Ksenia Ponomareva
• Ekaterina Dmitrova
• Egor Polyakov
• Alexandra Kornilova
• Artem Gudkov
• Maksim Chukseev

Neural networks were previously perceived as an element of literary cyberpunk, rather than a real prospect for the future. Currently, they significantly simplify the work of developers and creators, allowing them to quickly create a variety of content, including text, concept art, backgrounds, videos, 3D models, and voiceovers with various voices. Even giants such as Marvel Studios are actively using content created by neural networks. A striking example is the opening title sequence for the series "Secret Invasion." Neural networks are becoming an integral part of the creative process, changing the approach to media creation and opening up new opportunities for artists and screenwriters.

The usefulness of neural networks continues to spark debate and discussion. There are a variety of opinions formed based on personal experience and individual views. For example, Chinese artists express concerns about losing their jobs due to automation. Meanwhile, Western creators accuse neural networks of unauthorized use of their creative assets. As a result, some artists have filed class-action lawsuits against technologies that are using their original works without permission. These issues raise important questions about intellectual property rights and the future of creativity in the digital age.

We interviewed 3D artists, level designers, composers, screenwriters, and voice actors to learn about their attitudes toward neural networks and their use in their work. The responses were varied, which makes this material particularly valuable. Each essay represents personal experience and unique insights that the developers are willing to share. Thus, we get the opportunity to better understand how neural networks influence creative professions and what prospects they open up for the future.

Artemy Kotov

The editor and narrative designer play a key role in Creating engaging content. The editor is responsible for the structure and quality of the text, ensuring it complies with the style and requirements of the project. The narrative designer, in turn, creates captivating stories and dialogues that hold the audience's attention and deepen interaction. It is important that both specialists work closely together, ensuring a harmonious blend of text and narrative. The combined efforts of the editor and narrative designer contribute to the creation of unique content that not only informs but also entertains.

Neural networks are a powerful tool for developers, but their application depends on the specific situation. These technologies allow us to solve a variety of problems, from data processing to the creation of intelligent systems. The effectiveness of neural networks depends on the context in which they are used and the goals of the developer. Given the constantly evolving technology, it is important to choose neural networks based on the specifics of the project and the required results.

Currently, there is an unhealthy atmosphere of hype surrounding new technologies. A similar situation was observed with NFTs, metaverses, and cryptocurrencies, and earlier with dotcoms. The narrative remains the same: this is the future, don't miss out on the opportunity to be part of it. Hype is whipped up to attract investors, while everyone else remains in the dark. It's important to maintain a critical eye and analyze the true value of the innovations being offered to avoid falling victim to temporary trends.

Neural networks are becoming an increasingly relevant topic, and this is due to several social factors. Living costs are rising, and finding decent work is becoming increasingly difficult. Amid constant news of mass layoffs and strikes, many are tempted to turn to alternative ideologies, such as Luddism, Buddhism, or communism. However, it's important to understand that technology, including neural networks, can both threaten traditional jobs and open up new opportunities. It's essential to adapt to change and find ways to integrate new technologies into everyday life to maximize their benefits.

Neural networks are increasingly finding useful applications in various fields. One striking example is generative fill in Photoshop, which significantly simplifies the process of image editing. Also worth noting is Adobe's Enhance Speech, which improves audio quality and makes speech clearer. DeepL machine translator demonstrates a high degree of accuracy in translating texts, making it an indispensable tool for working with foreign languages. Neural networks give developers the opportunity to integrate useful features into existing products, which not only simplifies tasks but also speeds up the workflow, freeing up time for more creative aspects of the work. Rewrite the text, preserving its core idea and avoiding unnecessary information. Optimize it for search engines and add useful content where necessary. Avoid using emojis or unnecessary symbols. Eliminate sections with numbers and symbols, providing only pure text. The use of the term "artificial intelligence" in the context of neural networks puzzles me. In my opinion, it would be more accurate to call it "artificial idiot." This term is certainly less commercially attractive, but it reflects reality. A neural network lacks intelligence: it is not aware of its actions and does not understand the goals assigned to it. Each of its outputs is a random combination of elements formed according to a specific algorithm, comparable to a collage. The creative process, by contrast, requires awareness and design. Humans do combine familiar images, but they do so with a clear understanding of their goals and intentions. It is humans who are capable of transforming the client's vague wishes into concrete, tangible results.

Stories of neural networks creating impressive drawings and coherent texts continue to attract attention and are widely shared. However, important aspects are often overlooked: how many attempts were required to achieve a satisfactory result and how well the final product meets the initial request. It's like putting a monkey at a typewriter, hoping it will create a masterpiece. This approach seems interesting, but in reality, it can lead to disappointment and inefficiency. It's important to keep in mind that successful content generation requires significant effort and time, and isn't just random.

A neural network and a monkey are both capable of ultimately producing something acceptable. The main question is how much time you're willing to invest and what your minimum quality bar is. In both cases, the result is achieved through significant effort and trial and error.

Solo developer Slava Gris decided to create a conceptual visual novel using Midjourney and DALLE-2 technologies. However, he recently announced the project's suspension and a switch to another game. The main reason was the inconsistent results produced by the neural network. Slava noted that the time spent experimenting with prompts and fixing artifacts exceeded the time required to create original art by hand. Therefore, he concluded that traditional artistic methods are more effective in this context.

In March, an interesting incident occurred when various public groups shared the work of artist Evgeny Zubkov, claiming that it was created by a neural network. The question is how much time and effort it will take for a neural network to create a work comparable to that of a real artist in terms of detail, cultural references, and witty associations. Currently, artificial intelligence demonstrates the best results in creating generic portraits of anime characters and other low-quality works. In this context, professional artists should be aware that neural networks may threaten their employment, especially in fields where quality and originality are not crucial. This highlights the importance of unique creativity and skill that cannot be replicated by artificial intelligence.

Neural network-generated graphics are increasingly being used in mobile games. The visual novel MeChat from PlayMe, for example, features backgrounds with noticeable artifacts. Despite this, the app is popular because users aren't looking for purely aesthetic appeal. Graphics quality in the mobile segment is generally lower than on other platforms. In this context, it's interesting to hear about Rayark, which allegedly reduced its artistic staff and began incorporating AI-generated art into its rhythm games. This confirms the relevance and development of neural network technologies in the gaming industry.

In some cases, developers may indeed lose jobs, especially where organizations are trying to cut costs. However, the quality of neural content is often unsatisfactory, and this is unlikely to change in the near future. I doubt that neural networks will be able to create visual works, write texts, or generate music at a significantly higher level than they currently do; this contradicts the very principles of their operation. Neural networks can help create mood boards, select references, or even generate placeholders for secondary content, but their capabilities remain limited. Therefore, developers still play a vital role in the production process, ensuring quality and creativity that automation cannot achieve.

Without the involvement of a creative professional, it is impossible to create high-quality content, and plagiarism is the only possible result. Neural networks are trained on copyrighted materials, and their use in projects can lead to lawsuits. This can also cause problems with product releases: for example, the Steam platform has already begun refusing to host games developed using neural network technologies by studios. Ultimately, cutting corners on professionals can result in significant losses.

Creating original content is a challenge faced by many artists and writers. Most will agree that developing an idea from scratch is much more effective than editing work created by artificial intelligence. For indie developers, originality is a key factor for success. In a competitive market, unique ideas and a creative approach help them stand out and attract an audience. Therefore, it is important to focus on creating original works, which can significantly increase the chances of success in the industry.

• Telegram channel "Kotemius"
• Artemy Kotov's blog on DTF

Asker Unaev

A screenwriter is a key figure in the creation of film and television projects, responsible for developing scripts that become the basis for visual storytelling. In projects such as "Kuzhlevka" and "Greed," the screenwriter shapes the plot lines, dialogue, and characters, ensuring the story's coherence and appeal to viewers. The ability to create engaging and original screenplays requires a deep understanding of the genre, audience, and current trends in the film industry. Each screenplay is not just a text but also an important element contributing to the successful development of a project and its commercial success.

Serious reasons to consider neural networks an "enemy" will arise only when their work becomes truly indistinguishable from human activity. In a short period of interaction with these technologies, most people have already learned to intuitively recognize generated images and text. Neural networks are becoming more sophisticated every day, raising concerns about their impact on society and the labor market. It is important to understand that the development of artificial intelligence can lead to significant changes, and that is why it is necessary to closely monitor these processes.

From the perspective of a developer working with video game scripts, it's worth paying attention to experiments integrating ChatGPT into role-playing games. These experiments demonstrate that text generation can add an extra five minutes of gameplay, allowing players to interact with NPCs in Skyrim and expect unexpected or even funny responses. However, it's important to remember that generative technologies do not create new storylines or unique elements in the fictional world, but rely solely on existing data provided by the user. This raises questions about the role of artificial intelligence in game development and its ability to enrich game scripts without compromising the integrity of the original concept.

If we task a neural network with creating its own fictional world, we would encounter the same phenomenon. A neural network operates on existing data and is incapable of generating truly new ideas. As a result, the universe it creates will be a combination of elements taken from known works and worlds created by humans. This approach limits creativity and originality, as the machine only uses existing concepts and images.

True artificial intelligence has the potential to become a vital tool for the future of the entertainment industry. It can generate unique ideas and, crucially, have the ability to distinguish between good and less successful concepts. This will enable the creation of more engaging and innovative content that will meet modern audience demands and contribute to the development of the entire industry.

Modern shows powered by neural networks are already a reality. The recent success of the neural version of "SpongeBob SquarePants" sparked great interest among viewers. I myself watched these first steps into the world of generated content with great fascination. The pleasure of watching came not so much from the plot, but from the absurdity of the action and the curiosity of a researcher: "What else can neural networks offer?" This experience opens new horizons for entertainment and makes us think about the future of art and technology.

The road to neural networks truly mastering the art of storytelling seems to be quite long. Even with human-defined parameters, the ability of neural networks to create original stories remains a complex task. Developing algorithms that will allow them not only to follow templates but also to express creativity requires time and considerable effort. It's important to understand that achieving this goal requires not only improved technology, but also a deep understanding of narrative structure, emotion, and the human experience.

To date, neural networks have not replaced humans, especially in creative professions. The panic that gripped artists and other creative professionals proved premature. Instead of a threat, we have gained a powerful tool that can significantly facilitate the work of artists, programmers, translators, and, of course, screenwriters. Neural networks open up new horizons in the creative process, allowing professionals to focus on more important aspects of their work.

It is crucial for screenwriters to evaluate the originality of their idea before implementing it. If a neural network is able to assemble elements that create something similar to your idea, this indicates that something similar has already existed in one form or another. Understanding the uniqueness of an idea helps avoid repetition and helps create original and engaging content. There are various tools and methods for checking originality that can be useful to scriptwriters during the development stage.

• Official Callback Games community on VKontakte
• Kuzhlevka game page on Steam
• Kuzhlevka game page on VK Play

Artem Makarov

The head of Anoft, a studio known for its projects Loot List, Reflection, and Offline, is a key figure in the field of video game development. The Anoft team is actively working on creating unique gaming content that combines original ideas and high quality execution. Each project of the studio strives to offer players an engaging experience and innovative mechanics, which makes Anoft a notable player in the market. Interest in the studio's projects continues to grow, which is confirmed by positive reviews and a high level of user engagement.

A neural network is a useful tool in development. It can significantly reduce the costs of resources required to create products. Neural networks will evolve over time, but for now, in the gaming industry, they should be considered as a practical solution. When using neural networks, the key question is "Why?", followed by "What?" and "How?" This approach helps to optimize the development process and achieve better results.

A neural network is an effective tool for forming a database knowledge that provides users with accurate and up-to-date information. It is capable of analyzing large volumes of data, extracting useful knowledge and facts from them. The neural network is actively used in various fields, including education, science, and business, which makes it an indispensable resource for obtaining information. Thanks to its algorithms, the neural network ensures fast and high-quality knowledge extraction, which contributes to increased work efficiency and decision-making.

• Concise. Multi-page text can be converted into a few lines ("Briefly summarize this article, highlighting the key points").
• Targeted. You can quickly find out interesting things ("What is the most popular video game genre in the Philippines in 2023?").
• Formatted. For example, through lists ("Make 5 lists of objects associated with Sherlock Holmes, and in each list, the objects are combined according to random characteristics").

The neural network is capable of generating simple code. However, when working with more complex systems, it often provides inefficient or incomplete solutions. Fixing such errors can require more time and effort than if the tasks were solved manually. Furthermore, relying on a neural network for code optimization is usually impractical, as it does not always cope with this task at the required level.

Neural networks are excellent at providing hints across various programming languages ​​and software. For example, the query "How do I call one of many functions in Unreal Engine depending on a large number of bool variables?" is quite appropriate. Such questions help developers find efficient solutions and optimize their code, which is especially important in the context of modern programming.

Neural networks are able to take context into account, which allows them to provide relevant answers to questions such as "What if, in addition to bool, there is also a float?" However, it is worth noting that neural networks typically retain context only for 5-6 refinement iterations, after which the quality of their answers can significantly decrease. In such cases, it is recommended to restart the session to obtain more accurate information.

Neural networks demonstrate average results in solving creative problems. Their responses may appear coherent and meaningful at first glance, but upon closer analysis, they often reveal a lack of depth. However, such responses can be useful for those with lateral thinking skills and the ability to carefully select the necessary information. Using neural networks in creative processes is possible, but it requires the user to be able to critically evaluate and adapt the results.

A neural network effectively solves mathematical problems, including determining optimal tiling and calculating the parameters of the polygons that make up a given tiling. It can take given input data into account, making it a useful tool for solving mathematical problems.

It is important to realize that the results provided by neural networks do not always reflect reality and may be unreasonable. This should be taken into account, especially if you are working with information whose application goes beyond simple interest. A user of a neural network should be able to independently evaluate the obtained data or, if necessary, verify it using other sources. Experts, search engines, or even other neural networks can help with this.

Based on the examples given, it can be argued that a neural network is a valuable source of information. The question of how exactly to use this information is a topic for a separate discussion, perhaps even with the participation of the neural network itself as the author.

• Official Anoft studio community on VK
• Anoft studio page on Steam

Ksenia Ponomareva

Composer and sound designer known for her work on projects such as "Vasilisa and Baba Yaga" and the music video for Sabaton's "The Night Witches." His work spans a wide range of styles and genres, allowing him to create unique soundscapes and atmospheric musical compositions.

My name is Ksenia Ponomareva, and I am a professional sound designer and composer for games. I specialize in creating sound design and music for indie projects. In this article, I want to discuss the impact of neural networks on creative work and their application in sound design and music production.

The GPT revolution has divided society into two camps: some are actively implementing neural networks into development processes, while others are wary and distrustful of their achievements. I belong to the first category, although I understand that neural networks may replace some professions in the future. One thing is certain: the world will never return to the way things were before. Neural networks open up new horizons for creativity and efficiency, and their impact on work processes and everyday life is becoming increasingly noticeable. We are on the threshold of a new era, where technology will play a key role in shaping the future.

Not only are well-known neural networks for generating visual content, but also a variety of tools for working with audio. Sound engineers and sound producers actively use AI-based software that analyzes sound and significantly simplifies various tasks. For example, plugins from Izotope and Sonible help with audio processing, improving sound quality and speeding up workflows. In addition, easy-to-use online services such as Landr, which offer automated mastering of music tracks using AI, are available. These technologies significantly improve efficiency in the music industry and open up new creative possibilities.

Technical tasks associated with sound processing will likely be largely automated in the future using neural networks, allowing humans to focus on their role as operators. This will significantly simplify and speed up the mixing process for sound engineers and enable low-budget projects to create high-quality releases. The use of neural network technologies in audio production will open up new horizons for creative solutions and simplify access to professional sound.

In the field of interactive audio, neural networks significantly simplify the integration of sound effects and music into non-linear gameplay. Rapid analysis of multiple scenarios allows for improved sound quality, as well as expanded capabilities for speech synthesis and in-game dialogue generation. As a result, games fully voiced by artificial intelligence are becoming possible, opening up new horizons for developers and players. Innovative technologies in audio design create unique user experiences, allowing players to immerse themselves in the game atmosphere on a new level.

The recent boom in neural networks capable of generating the voice of a singer or actor based on a "timbre audio fingerprint" is opening up new horizons for the creative industry. These technologies offer unique opportunities for music production, voiceovers, and other audio projects. However, copyright issues must be carefully considered and the risks of fraud that may arise from the use of artificial intelligence must be minimized. Effective regulation and a legal framework will help ensure the fair use of these innovative solutions and protect the rights of original artists.

The recently released guitar amp replicator Neural Amp Modeler demonstrates a principle of operation similar to other successful solutions in this field. When it comes to creating content in the style of a specific artist or genre, neural networks deliver unrivaled results. Neural Amp Modeler effectively reproduces the characteristic sounds and characteristics of guitar amps, making it an indispensable tool for musicians and producers. It allows for easy experimentation with various sound palettes and achieve professional-quality sound, underscoring the importance of neural network technologies in modern music production.

In the context of creative tasks such as writing original music and creating sound design, neural networks cannot completely replace human work. Composers create music by following certain rules and using patterns that neural networks can learn. However, this is only the technical aspect of the process. Music requires not only theoretical knowledge but also emotional perception, artistic vision, and a unique style, which remain beyond the reach of artificial intelligence. Neural networks can generate melodies and sonic textures, but they are unable to convey the feelings and ideas that inspire people to create genuine musical works.

Currently, neural networks are unable to fully grasp context and the emotional-semantic component, which significantly limits their creative potential. For example, any person can easily create a rhymed quatrain, but not every one of them becomes a work worthy of literary heritage. The difficulty lies in the fact that contextual understanding and creative expression do not lend themselves to strict algorithms. Creative works are imbued with emotions and meanings that are difficult to formalize. This makes the process of training neural networks in the field of artistic creativity even more complex and multifaceted.

Neural networks do not create fundamentally new content, but merely compile information from existing data created and processed by humans. If human participation in updating this data set is excluded and neural networks are left to operate solely on their own, then over time their output will lose both form and content. This is reminiscent of Alvin Lucier's famous sound experiment, in which an audio recording gradually loses quality. Therefore, it is important to understand that neural networks depend on human input to maintain the relevance and completeness of the content they create.

I provide examples of compositions created by various neural networks. These tracks sound effective as background music for linear and game audio. However, if a higher-quality result than simple background music is required, human editing is necessary. This allows for deeper sound and emotional richness, which is especially important in projects where music plays a key role.

• Exploration, Soundful neural network;
• A Lullaby for a Kid, Riffusion neural network;
• Mysterious Mood, Soundraw neural network;
• Folklore Fairy Tale, Mubert neural network;
• Cosmic Vibes, Boomy neural network.

Predicting the impact of neural networks on the professional market is difficult. Currently, they act as assistants and allies for creative professionals, expanding development capabilities and simplifying the process of creating soundtracks for games. This makes this field more accessible to a wider audience, opening new horizons for creative expression and professional growth. The changes brought by neural networks could significantly alter supply and demand in various industries, including the gaming industry.

Neural networks and artificial intelligence play a significant role in human professional development. By mastering the capabilities of these technologies, people learn to effectively apply them in their work, which contributes to expanding their horizons and intellectual growth. The introduction of AI into various fields not only increases productivity but also opens new horizons for self-improvement and professional development.

In conclusion, I would like to present ChatGPT's opinion on the potential risks associated with replacing the profession of game composer with neural networks. It is important to note that the introduction of artificial intelligence into the process of creating music for games could lead to significant changes in the industry. On the one hand, neural networks are capable of generating high-quality music quickly and efficiently, which could simplify the work of game developers. On the other hand, this raises questions about creative uniqueness and copyright. The profession of composer requires not only technical skills but also a deep understanding of the emotions that music should convey in the context of gameplay. Replacing human talent with a machine could lead to the loss of these nuances, negatively affecting the perception of games. Therefore, it is important to find a balance between the use of technology and the preservation of human creativity in game music.

Neural networks are useful tools for composers, facilitating the development of initial ideas and the generation of musical fragments. However, today they cannot completely replace human creative intuition and a deep understanding of music and the context of video games. Composers remain a vital part of the gaming industry, bringing uniqueness, emotional richness, and creative individuality to each musical composition. Their contribution to creating atmosphere and immersing players in the gaming experience remains indispensable, underscoring the importance of human creativity in the age of technology.

• Ksenia Ponomareva's Instagram page*
• Ksenia Ponomareva on VKontakte
• Follow the Compass music project
• Official website of the Attaque de panique choir

Ekaterina Dmitrova

Dubbing actress and co-head of the R.G. Mechanics VoiceOver studio, with extensive experience in dubbing various projects. Specializes in the creation of high-quality sound content, including animation, video games and commercials. Her work is distinguished by professionalism and attention to detail, which allows her to convey the emotions and personality of her characters. R.G. Mechanics VoiceOver is known for its innovative approach and high level of customer service, making it one of the leaders in the industry.

Hello, my name is Ekaterina Dmitrova. I am one of the leaders of the R.G. Mechanics VoiceOver team, as well as a professional voice actress and sound engineer. Our team provides high-quality voice-over services, ensuring a professional approach to each project. We specialize in creating unique audio content that meets all client requirements.

Neural networks are complex technologies, and their impact cannot be assessed as unambiguously positive or negative. Having personally used neural networks to remove and replace an actor's original voice in a soundtrack, I realized how useful these tools can be in the creative process. Such technologies open up new possibilities for content creators, allowing them to experiment with audio formats and improve sound quality. However, it's important to remember that using neural networks requires responsibility and ethical standards, especially when it comes to voice and identity replacement.

I was involved in the localization of the Burial at Sea expansion for BioShock Infinite, specifically helping the CGInfo team finalize the audio. The voiceover process began seven years ago, but audio processing stalled at one point. The main issue the team and I faced was the mono tracks in most of the cutscenes, including commercials, tutorials, and character death scenes. Extracting voices from music and background audio using standard software proved impossible.

Extracting the center channel is a complex task, especially in mono mode, where this channel is absent. The only option is to extract the center channel from a stereo recording, but this carries the risk of introducing a "kick drum" effect into the sound. So, what should you do in this situation? It is necessary to consider sound processing methods and the editing capabilities of stereo tracks to minimize distortion and achieve the desired result.

The most effective, but resource-intensive, solution is to create a backing track. This process involves carefully selecting music and sound effects, such as footsteps, background noise, wind, and other audio elements. The sound engineer manually selects all the sounds present in the original track, with the exception of the voice. The main task is to replace English speech with Russian, which requires a careful approach and highly qualified specialists.

There are simpler solutions - neural networks. I started searching for information on the internet, where I did not encounter any limitations. As a result, I found three neural networks, two of which are available for use online.

• Vocal Remover;
• X-minus pro;
• Ultimate Vocal Remover. This gadget needs to be installed on your computer and carefully studied. It has a ton of different settings and parameters. Trial and error can only cause a ton of errors and an inflammation of the buttocks of the nervous and impatient.

I decided to choose the first neural network. It has a simple interface and is easy to use, but free access to its functionality is limited by the number of requests per day. If you need more features, you will have to pay. In my case, I initially did not need many requests, plus I have friends with different computers and IP addresses, which allows me to use additional resources.

Here is one successful example of using this neural network:

The neural network successfully extracted the voice from a video with effects simulating old television and provided a separate track with background sound. I simply overdubbed the actress's Russian voice onto this track without making any additional edits.

Neural networks, despite their advanced technology, aren't always successful in their tasks. For example, issues can arise when working with audio tracks. This illustrates the case where the Vocal Remover neural network failed to correctly distinguish between the English track and the background. This highlights the importance of understanding the limitations of technology and the need for further development.

This track contains numerous audio artifacts, such as whistling and a kick drum effect, which create a "beaten" sound. Even with overdubbed voices, the original speech will be audible alongside these unwanted effects. To improve audio quality, it's necessary to use sound libraries to download the necessary elements, such as microphone noise and background noise caused by equipment. After that, we can mix these sounds, cut out a fragment of the melody from the original track, for example, a notification about the start of a commercial, and achieve a cleaner and more professional sound.

Vocal Remover deserves positive reviews. In my opinion, this program demonstrates better results compared to X-minus pro. Ultimate Vocal Remover, in turn, is suitable for more experienced users, as it requires careful study and skill. However, it also provides high-quality results, allowing you to separate audio tracks into vocals, bass, drums, and other elements. To achieve optimal results, it is important not only to experiment, but also to carefully study the program's functionality. I recommend referring to English-language manuals for a deeper understanding of Vocal Remover's capabilities.

The topic of neural networks and voice actors is a source of much discussion online. Actors are concerned and dissatisfied with this technology, as it may threaten their profession. Extensive databases of the voices of famous actors, singers, and politicians exist, fueling concerns in the industry. Neural networks are capable of imitating unique timbres and intonations, calling into question the need for live voiceovers. It's important to consider that the use of such technologies requires an ethical approach and respect for the rights of artists to maintain a balance between innovation and the preservation of the profession.

You can upload a basic voice sample you need and then voice any line yourself. The neural network transforms your voice into the voice of celebrities such as Britney Spears or Clint Eastwood. This is a clear example of how one voice can be replaced by another using a base, a recording of your speech, and RVC technology.

A man instantly performed a song with the voice of a young Japanese girl. This technology is amazing and arouses great interest, but at the same time, it can also be dangerous.

Voiceover for films and games is possible using a simple voice database, a high-quality home microphone, and the skills of a confident PC user. This will require a significant amount of time. Many detailed guides are available online, with resources in English being especially helpful. The voice database is available on the Discord server, where samples of other users' work are also displayed. This is a great opportunity to study and share voice acting experiences.

I believe that neural networks will not be able to replace actors. This is because a human will always be needed to serve as a role model and example. Acting manner, style, phonetics, as well as emotions such as loss, happiness, and anger, require human interaction. A neural network can certainly reach a certain level, but at critical moments, it will not be able to convey the nuances that a human can express. Therefore, in my opinion, voice actors will continue to be in demand at both the professional and amateur levels.

Artificial intelligence is ideal for those who value speed and efficiency in the content creation process. Viewers always watch actors with interest, whether it is a successful take performed on the first try or an unsuccessful but funny moment. This interaction adds an element of liveliness and authenticity, which makes viewing more engaging. Using technology helps simplify the process while maintaining audience engagement in the creative process.

I plan to use neural networks exclusively for sound processing, not for integrating the voices of famous actors or actresses into games. My goal is to create original sound solutions that will improve the quality of the gameplay and immersion of players.

• Official Mechanics VoiceOver community on VKontakte
• Mechanics VoiceOver Telegram channel
• Mechanics VoiceOver YouTube channel

Egor Polyakov

Matt Painter is a talented concept artist known for his contributions to the visual art of popular projects such as Call of Duty: Warzone, as well as the series The Witcher and Shadow and Bone. His work helps shape the unique atmosphere and style of these works, impressing viewers with his skill and creativity. Painter is actively involved in the development of characters and the world, which makes him an important figure in the entertainment industry.

Today, neural networks have become an integral part of the artistic profession in the gaming and film industries. The question of whether they are an assistant or a threat to developers remains open and depends on individual perception. However, it is impossible to ignore the fact that neural networks are a powerful tool that can significantly facilitate the creative process and increase work efficiency. It is important to use them wisely to avoid potential negative consequences and maximize their potential for creating unique content.

The benefit of neural networks lies in their ability to speed up workflows and significantly simplify routine tasks that artists may not enjoy. Thanks to neural networks, certain operations can be automated, allowing you to focus on the creative aspects of your work. This makes them an indispensable tool in the arsenal of modern artists and designers, as they help optimize the time and effort required to achieve high-quality results.

Neural networks are a powerful tool that can significantly speed up and improve workflows, but they cannot completely replace humans. Even with access to neural networks, a person without the necessary skills and knowledge in a particular field will not be able to achieve the same results as an experienced specialist. It is important to remember that neural networks serve to support and improve work efficiency, but individual skills and creativity still play a key role.

Further, I will provide several common examples that clearly illustrate my point.

Let's consider the matte paint tool in the context of Photoshop updates. With the introduction of the generative fill feature in the latest version of the program, many experts predicted that it would lead to significant changes in the matte painting profession and possibly its demise. However, in practice, this tool has become useful primarily for generating ideas and creating compositions. The quality with which generative fill restores missing image elements or modifies them does not meet the high standards set in the film and game development industries. Matte painting still retains its value as a professional discipline, as it requires artistic flair and skills that cannot be fully replaced by an automated tool. When it comes to sketching, quickly placing objects, or testing different ideas to assess their appropriateness within a composition, it is a truly useful tool. It allows you to quickly visualize concepts and adapt them to the overall structure, significantly simplifying the creation process. Using such tools helps designers and artists effectively develop their ideas and find optimal solutions.

If a person If a user lacks fundamental skills, they won't be able to effectively refine the material they've received. Without a basic understanding, they won't be able to determine what exactly needs improvement and in what areas. This emphasizes the importance of foundational knowledge, which allows them not only to perceive information but also to critically analyze it and make necessary changes and improvements. Concept art also requires a careful approach, and the Midjourney neural network provides the ability to create a variety of image variants. For example, it can generate different styles of sniper rifles in the sci-fi genre. However, determining which variant is the most successful and which elements need to be refined requires knowledge and experience that aren't always available to ordinary users. Furthermore, many critics point out shortcomings in the functionality of designs created with Midjourney, which should also be taken into account when working with this technology.

I often apply neural networks to areas that don't particularly interest me. For example, I was curious to use neural networks in my personal projects. To create a high-quality portfolio, I need extensive background and backstory for objects, architecture, or characters. Having only a brief description can lead to a superficial design. Meanwhile, detailed information and answers to key questions allow me to deepen the project and make it more engaging. Using neural networks in such cases not only saves time but also improves the quality of the final result.

There's no need to waste time on in-depth research into the background of each work and writing a detailed description. I use Midjourney to generate queries that serve as the basis for the technical specifications. This helps me improve my design, composition, and lighting decisions. This approach simplifies the process and allows you to focus on the creative aspects of the work.

It is important to consider the ethical aspects associated with The use of artists' works in neural networks. Artists' works are often used for commercial purposes without their knowledge or consent. Generating images in a specific artist's recognizable style raises serious questions about copyright compliance and respect for creativity. This approach can be considered a violation of artists' rights and their intellectual property, highlighting the need to develop clear norms and guidelines for the use of their works in the digital world.

In conclusion, I want to emphasize that regardless of your opinion on neural networks, it's worth considering how they can improve your creative process. Using neural network technologies can free you from routine tasks that have long been irritating and unfulfilling. I also hope that in the near future, neural networks will allow us to effectively create UV mapping with minimal effort.

• Egor Polyakov's Instagram page*
• Egor Polyakov's ArtStation profile

Alexandra Kornilova

A Project Manager at Dixotomia is responsible for coordinating and managing projects, ensuring their successful completion on time and within budget. This role includes planning, organizing, and overseeing all project phases, as well as collaborating with the team and clients to achieve goals. A Project Manager must possess communication, organizational, and strategic thinking skills to effectively solve problems and optimize processes. At Dixotomia, we value initiative and a proactive approach, which allows us to create high-quality solutions for our clients and deliver outstanding results. Experience in this field is an important aspect that helps a project manager successfully cope with tasks and achieve the desired results.

Neural networks have become an integral part of our lives, and everyone, except perhaps the twins from Atomic Heart, is talking about it. Even internet-savvy grandmothers have repeatedly encountered generated Russian cities and expressed their admiration. It's impossible to ignore the influence of neural networks on various spheres, including art, science, and everyday life. For example, Sberbank also released its own neural network bot, confirming the growing popularity and necessity of such technologies. Neural networks are permeating our everyday lives, changing our approach to problem solving and improving our quality of life.

Despite its simplicity, the bot is capable of producing images close to the desired results when used correctly. We won't list all the existing generators, but will focus on Midjourney. This tool plays a key role in the development of the VR shooter Dixotomia, enabling the creation of unique visual elements for the game.

Artem Gudkov

3D artist with experience working on projects, With projects such as "Battles for Glory" and "Dixotomia," he specializes in creating high-quality 3D graphics and character modeling. His deep knowledge of texturing, lighting, and animation allows him to develop visually appealing and realistic game worlds. He is proficient in using modern 3D modeling programs such as Blender, Maya, and 3ds Max, ensuring a high degree of detail and aesthetics in each project. He strives to continually improve his skills and explore new technologies in 3D design to create unique and captivating visual effects for games.

Dixotomia is a unique VR shooter that immerses players in the world of vampires. One of the key features of virtual reality is the feeling of real presence in a digital space, which is achieved through surround sound and high-quality visuals. In VR headsets, the system tracks the position of the headset and two motion controllers held by players. To improve body recognition in the virtual environment and minimize desynchronization, only the hands are displayed, allowing the user to more easily navigate and interact with the game world.

Our first prototype was developed a long time ago, and it was time to update it. In creating the new "handles," we used a neural network, which made the process more efficient, albeit labor-intensive. MJ didn't always produce results that met our expectations. Rumors have surfaced online that the AI ​​has been updated, and now the hands and fingers are generated significantly better. However, sometimes the number of suitable options varied, which affected the final result.

Based on several generated images, we We created a moodboard, which became the basis for modeling the hands.

The result is hands that allow you to explore the world around you and have fun move your fingers.

We will release a standalone version for Oculus Quest 2 soon! This update will provide users with an improved experience and new interaction options. We strive to make your game more engaging and accessible by offering high-quality content and optimized features for this platform. Get ready for new adventures and unforgettable moments in virtual reality with Oculus Quest 2!

• The official Dixotomia community on VKontakte
• The Dixotomia page on Steam

Maxim Chukseyev

A level designer is a specialist responsible for creating game levels and locations in video games, such as Mot de Pass, Last Dissident, and Dixotomia. Their job is to develop engaging and balanced game environments that help immerse the player in the game world.

Level designers work on the placement of objects, enemies, and controls, creating unique scenarios and challenges for players. They take game mechanics into account to ensure interesting and dynamic gameplay. Professionals in this field must have good design skills, an understanding of gameplay elements, and the ability to work with various level development tools.

In games such as Mot de Pass, Last Dissident, and Dixotomia, level designers play a key role in creating atmosphere and mood, as well as shaping the overall user experience. Their work requires creativity, technical knowledge, and the ability to adapt to changes during game development.

Artificial intelligence has become an essential tool in level design. Using technologies like Midjourney and ChatGPT, designers can significantly reduce the time spent creating mood boards with references. These tools help designers better understand the architectural features of buildings, explore their historical and technical characteristics, and identify additional nuances that play a crucial role in the design of game locations. Using artificial intelligence not only optimizes the workflow but also creates higher-quality, more detailed levels.

Designers are sometimes required to recreate existing buildings based on their knowledge and references, both fictional and real. It's easy to miss important details during the process, as remembering all the features can be difficult. In such cases, the Midjourney neural network comes to the rescue. By formulating a structured and clear query, you can obtain visual reference that will bring you closer to the desired image. This approach can significantly improve the quality of work and reduce the time spent searching for inspiration.

If you are having difficulty creating a stylistically correct If you're looking for a location that takes into account the environment and details, don't despair. With a few queries to ChatGPT, you can find interesting nuances that will add uniqueness and depth to your level. Use the power of artificial intelligence to generate ideas that will make your location more appealing and harmonious. This will help create an atmosphere that will immerse users in the game.

When creating new locations for the Dixotomia project, I sometimes use neural networks. Before designing an industrial location, such as an iron ore processing plant, I initiated a dialogue with ChatGPT. During the process, I clarified the stages iron ore goes through in steel mills before being turned into steel ingots. I also inquired about the equipment in the workshops, how the cooling systems functioned, and how chimneys had changed over the past hundred years. This data helped me better understand the scale of the location, its content, the player's journey, and how the visual narrative would be implemented. The use of technology allows us to create more detailed and realistic gaming spaces, which enriches the player's experience and makes it more engaging.

When When I created the mood board, I already had a concept in mind, which allowed me to select suitable references from the internet and generate unique images. This process helps to visualize ideas and create a harmonious design that matches the chosen theme.

If you encounter difficulties in Don't despair when working with neural networks. Try again, and you're sure to succeed. Neural networks may seem complex, but with practice and patience, you can achieve success. Don't be afraid to experiment and explore new approaches, because each attempt brings you closer to your desired result.

• Official Dixotomia community on VKontakte
• Dixotomia page on Steam

A court has ruled to ban Meta Platforms Inc. from operating in Russia due to its involvement in the Facebook and Instagram social networks. The basis for this was the accusation of extremist activity.