Reply to thread

Message

[QUOTE="Queen, post: 86608, member: 27"]
Real-time GPU source separation just leveled up, and developers suddenly have way more firepower to mess with live audio streams.

SDK update and new module
[LIST]
[*]GPU Audio refreshed its SDK for NVIDIA and AMD graphics cards.
[*]The company introduced a second module for source separation.
[*]Steven Atkinson’s NAM Module previously received company backing.
[*]The new toolkit runs on Windows and Mac.
[/LIST]
HS TasNet real-time separation
[LIST]
[*]GPU Audio demonstrated it at the 2025 Audio Developer Conference in Bristol, UK.
[*]The engine relies on the HS TasNet model from L-Acoustics research.
[*]A neural network was tuned for low-latency, live scenarios.
[*]Remixing and noise reduction can happen during playback.
[/LIST]
Pro Audio division rebrand
[LIST]
[*]GPU Audio renamed its Pro Audio arm to Soundcore.
[*]The shift signals a broader identity for its tech push.
[*]Performance details live on the company’s GitHub page.
[*]Developers can grab the SDK immediately.
[/LIST]
[/QUOTE]

Name