This diploma explored the possibility of sound visualization using various software programs and artificial intelligence. The research aimed to examine how modern digital tools and artificial intelligence interpret sound, and whether this interpretation can also be visual and understandable to the end user.
The theoretical part presented the importance of visualization, the basics of sound, its physical properties, possibilities of visual representation of sound waves and spectrograms, sound capture, the history of visualization through cymatics, and the development of artificial intelligence and its usefulness in generating sound into images.
In the experimental part, four software tools for sound visualization were first presented: Audacity, Wondershare Filmora, Neural Frames and Synesthesia. The central part of the research focused on using the artificial intelligence tool ChatGPT, which was used to develop a system that enables the analysis of audio files, generation of textual descriptions, and creation of final images.
The results showed that artificial intelligence still relies primarily on textual input data, as even a tiny change in the name of an audio file can significantly affect the final visualization. This confirms that humans still play an important role as intermediaries since how the sound is described directly guides the system's operation. The research confirmed that with the right approach and human assistance, it is possible to successfully visualize sound.
|
