Reddit asks: models for voice reconstruction and remaster of low-quality audio?

Ant_6431 · reddit · 2026-09-19

A Redditor with a large library of low-quality dialogue audio with severe frequency cutoffs asks for models that can upscale, reconstruct, or remaster speech in ComfyUI, and whether AI song-cover models like yue2 could work. Suggestions welcome.

Original post →

More from Multimodal

Multimodal channel →