AI Engineering

Jan 14, 2026 · Artificial Intelligence

NovaSR: A 52KB Audio Super-Resolution Model that Upscales 16kHz Audio to Clear 48kHz

NovaSR is a 52KB open-source audio super-resolution model that can convert blurry 16kHz recordings into clearer 48kHz output, processing up to 3600 seconds of audio per second on a single GPU, and offering fast, lightweight enhancement for TTS, dataset cleaning, and edge devices.