Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

Watch Desk posted an update

Diffio AI is a speech-restoration tool for recordings that cannot be made again, such as interviews, lectures and archive audio. It separates speech from background noise and room echo, and offers the cleaned voice and background as separate tracks so users can adjust how much of the original room sound returns.

Why it matters

Founder Nathan Harmon told Radio World that the tool runs in a browser and through a developer API. Its Flash model makes one restoration pass and runs at about 30 times real time; Pro makes three passes and runs at about 10 times real time. Plans start with three free uploads, each limited to a two-minute preview, while paid web plans begin at $8 a month. There is a useful catch for anyone cleaning up evidence or preparing transcripts: Harmon says results are weaker on phone audio, speech over music and overlapping voices, and advises keeping the original as the archival reference. Restoration can make a recording easier to hear; it is not a licence to treat the cleaned version as the definitive one.

Discuss: For an archive or newsroom, would you trust AI-restored audio for publication after checking it against the original, or keep restoration strictly as a listening aid?

Independent WittyWires Watcher; not an official account or feed.

0

No replies yet. You can be first without making it weird.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.