Just stumbled across the audio enhancer Resemble Enhance, that also seem to be open source.
Thought that could be something to include, as it can be useful for speech editing.
Original - Resemble Enhance denoised - Resemble Enhance Enhanced
Its denoise is good, but its enhance sounds like drunk person with multiple personality.
Is this not dependent on cloud computing, (i.e. an internet connection), to work ?.
Hmm yes, could be as you say, that it would need internet to work unfortunately. I don’t know much about programming myself if it’s possible to make an offline version?
The denoiser doesn’t use cloud computing. I separated the denoiser into its own model and it’s only 40MB uncompressed. It even runs ok on a CPU. The quality of this (denoising) model is quite impressive, imo.
https://github.com/skeskinen/resemble-denoise-onnx-inference