Poja Labs (CEO Heo Won-gil), a specialist in artificial intelligence (AI) composition, announced on the ninth that it has succeeded in securing data using the next-generation 3D audio standard ‘IAMF’ technology and can begin full-scale development of AI 3D audio generation technology based on this.
It was explained that this was a move to dominate the ‘3D audio market’ by applying IAMF technology developed by Samsung Research. We plan to offer an environment where creators can easily create 3D audio content using MIDI music.
IAMF is an open source-based 3D audio technology. It’s an audio technology standard first adopted by the Alliance for Open Media (AOM), which incorporates various global corporations including Samsung Electronics, Google, Apple, Netflix, Amazon, and Meta. Specifically, after YouTube announced its plan to introduce 3D audio services based on IAMF technology in 2025, it’s specializing in the opportunity of use in fields corresponding to virtual reality (VR), augmented reality (AR), streaming, games, and broadcasting.
Accordingly, Poja Labs plans to secure data using IAMF technology and construct an automatic 3D audio generation model.
3D audio technology can apply spatial audio information to dozens of tracks that make up a single song, providing an optimal sense of space regardless of what environment you take heed to. It was explained that with a view to implement this technology, a separate data set for every track have to be constructed, and at the identical time, it have to be possible to generate track-level sound sources.
Since its inception, Poja Labs has hired skilled composers to directly produce track-level composition data to resolve copyright issues. As well as, its track-based MIDI creation technology has been recognized by the world’s most prestigious academic societies corresponding to NeurIPS and AAAI.
This 3D audio data set can also be being built using Poja Labs’ skilled sound engineers and spatial audio studio. He said that the concept is to expand existing composition data by adding spatial audio information.
Meanwhile, global AI music creation services corresponding to Suno and Udio explained that since the whole sound source is generated without delay, it’s unimaginable to switch the sound source for every track or apply spatial sound technology.
Taehyun Kim, CSO of Poja Labs, said, “Because the IAMF technology has turn out to be open source, we expect an era wherein anyone can easily create and use 3D audio content.” He added, “We’ll support the spread of 3D audio content all over the world, and at the identical time, we’ll support global partners.” “We’ll work with others to construct a related ecosystem,” he said.
Reporter Jang Se-min semim99@aitimes.com
