“In 5 to 10 years, a lot of the movies can have Gaudio Lab’s sound. At its core is sound-generating artificial intelligence.”

Gaudio Lab (CEO Oh Hyeon-oh), which makes a speciality of AI-based sound, announced its intention to expand its core business by proving its skills in the sphere of ‘Foley Sound’ at a worldwide event.

Jeon Sang-bae, chief science officer (CSO) of Gaudio Lab, said on the thirtieth, “The subsequent technology that the corporate, which has won many awards within the acoustic field, is specializing in is the generation AI sound, or poly sound.”

Foley Sound is known as after Jack Foley, who was answerable for sound in Hollywood within the Nineteen Forties and Fifties, and is used to create sounds comparable to people walking, horses hooves, dogs barking, and sword fights in movies.

Gaudio Lab participated within the poly sound synthesis field of the world’s first AI sound generation challenge held last month at DCASE, a world-class audio-related event hosted by the Institute of Electrical and Electronics Engineers (IEEE), as a number company.

He took part within the competition and took second place on the earth. More vital than deciding who’s first or second is the proven fact that this field was created by the proposal of Gaudio Lab, and the influence of this field is great.

Sound-generating AI, like image-generating AI, generates sounds from text prompts. The principle can also be a ‘diffusion’ approach to image-generating AI, which is a technique of refining a meaningful signal from white noise (noise).

CSO Jeon Sang-bae said, “We have now been researching and developing sound generation AI since 2021, long before ChatGPT got here out.” “After regular research, we succeeded in generating sound for the primary time in June 2022. We have now reached the stage where we will pay,” he said.

The top goal is to “develop a whole ‘multimodal’ sound generation AI that generates sound from video input.” In other words, just as an engineer picks out sounds while watching a movie, the goal is to turn out to be the “king of the top of sound-generating AI,” where AI analyzes images and creates sounds that fit the scene.

It’s explained that 30 to 80 percent of the particular film sound engineering process is spent on sound effects. It takes loads of time for a sound engineer to search out, hearken to, and fasten sounds to make use of while watching a video, but Gaudio Lab reduces that point.

As a consequence of OpenAI’s ‘ChatGPT’, generated AI content is becoming generalized. Pointless to say, image-generating AI comparable to ‘Midjourney’ or ‘Stable Diffusion’, and recently, generative AI in the sphere of voice has also been highlighted through the cloned music of singers Drake and The Weeken.

Sound production, alternatively, is comparatively unexplored. Gaudio Lab is a world leader in pioneering this field.

The technology developed by this company is ▲GSEP (sound source separation) ▲’GTS (Gaudio Text Sync)’ that generates lyrics or subtitles ▲generated sound AI (poly sound) ▲’Binaural Rendering’, which is crucial for an immersive metaverse ▲Spatial audio ▲BTRS (field sound reproduction system) ▲volume leveling ▲spatial upmix, etc.

You will have heard of the technology for the primary time, but it could actually be said that there isn’t any person in Korea who has not heard the sound of Gaudio Lab. Applied to OTT, theaters, smartphones, wireless earphones, in addition to streaming comparable to Naver Now, Bugs Music, FLO, and Weverse, the variety of users who experience Gaudio Lab technology reaches 20 million on daily basis.

CSO Jeon Sang-bae is the principal character who created this technology. He has a PhD in acoustic engineering from Seoul National University and is the perfect expert in Korea with 20 years of experience in related fields. He developed MPEG-H 3D Audio, a 3D audio standard, at Samsung Electronics DMC Research Center (now Samsung Research) and applied it to products comparable to smartphones, TVs, sound bars, and residential theaters. Founded in 2015, Gaudio Lab also has about 40 audio experts, including 9 PhDs in acoustic engineering, that are rare worldwide.

The explanation why CSO Jeon Sang-bae and Gaudio Lab are specializing in sound-generating AI is that it has great potential for growth.

“It can be essential to enhance productivity in mainstream media industries comparable to movies, OTT, and games, in addition to to recreate the virtual space of the metabus platform.”

Within the metaverse, it’s needed to continually create sounds to suit the motion, comparable to the sound of walking or picking up something, but that is something only a number of experts can do. Nonetheless, poly synthesis technology helps anyone create sound easily.

Regarding the servicing plan, he said that every one possibilities are open. “Normally, there are cases during which technology is developed based on market needs, and quite the opposite, there are cases during which technology development is completed first after which the market is found, but this technology is closer to the latter,” he said.

“Because the difficulty of the project itself could be very high, we began with the intention that we should always attempt to do something that nobody else can do with the intention of preoccupying the worldwide leading position without considering the market,” he confessed. As such, the background is the arrogance that it is just a matter of time before it penetrates the market.

As well as, he identified that a number of the work needs to be done abroad because the sound work is at a level that can not be handled in Korea attributable to the craze of K-content nowadays. So, he said, “The proven fact that Gaudio Lab’s poly sound synthesis technology improves their work productivity can also be an amazing added value.”

Lastly, he repeatedly emphasized that “the goal is to create a future where anyone can easily create and experience the sound they need by enabling AI to create the identical sound as in point of fact wherever sound is required, comparable to in movies and dramas.”

Reporter Juyoung Lee juyoung09@aitimes.com