Method for Developing Audio Design for Advertising using Generative Neural Networks

The paper discusses the application of generative neural network tools for sound accompaniment of advertising with a focus on Russian-language content. A comparative analysis of speech synthesis tools (YandexTTS, Play.ht) and music (Udio, Suno) is conducted, and their key characteristics, advantages, and limitations are identified. Special attention is paid to the challenges of integrating sound components with video content, including issues of synchronization, volume balance, and emotional matching. A step-by-step algorithm for creating sound accompaniment is proposed, from analyzing the video content and preparing text prompts to generating, mixing, and testing the final audio product. The study revealed that, despite the high potential of generative technologies, there are still problems with the unpredictability of the results, especially when creating a musical background. The proposed methodology confirms the possibility of automating sound design in advertising, with the need for a balance between technology and creative control. The results of the work have practical significance for specialists in sound design, marketing, and digital production, offering an effective way to automate the creation of audio content. The paper will be useful for researchers studying the application of artificial intelligence in creative industries.