@DSmu5ljACnw (Stem Isolated audio from meta) Video by DSmu5ljACnw

My Notes: Stem Isolated audio from meta

---

Meta might have changed audio editing forever.

Meta just introduced Segment Anything Model Audio, or SAM Audio, a research model that can isolate any sound from complex audio. You can control it with text, a click in a video, or a selected moment in time.

You can type what you want to extract, click where the sound comes from on screen, or highlight a short time span. One model handles speech, music, and everyday sounds, instead of switching between separate tools.

Meta is releasing the model, benchmarks, and an open evaluation dataset so others can build on top of it right away.

🔗 aidemos.meta.com/segment-anything/editor/segment-audio

#ai #audio #meta #machinelearning #research

Instagram by: DSmu5ljACnw (@DSmu5ljACnw)

Source: https://www.instagram.com/p/DSmu5ljACnw/?igsh=NzhpcHY0YnU2NmN3