The transcriptions with Whisper are a game changing feature!! I would love to see this continue receiving support. Absolutely incredible! It'd be amazing to have these high quality audio captions live while audio plays too. I noticed the model sometimes skipping over small words or it places the marker a little bit before the audio actually plays but it works really well overall, it's mostly limited by Audacity's user interface design. How about adding AI text to speech and voice cloning? Would be incredible if we could generate voice lines or even singing in your own voice right in Audacity. OpenVINO performs incredibly well on Intel hardware. I get almost twice the inference speed than other runtimes do while maintaining the same quality. Perfect for getting the most out of your iGPU! It's unfortunately, as usual with AI, a rather large download, for a decent quality model you already need a couple gigabytes of disk space. Torch CPU uses 230MB on it's own. I'm not sure if all of that is necessary, OVMS doesn't have torch in it but works just as well? Maybe an optional toggle on setup can be included to leave out Torch and install OpenVINO only for people that have Intel hardware. Thanks a loot!!

NOTE: Only works with Audacity 3.7.4 and later. OpenVino is a powerful collection of AI tools, built specifically for Audacity! Includes high-quality noise suppression, music separation into individual stems, voice transcription, and audio super-resolution.
Previews
What’s New?
First release of OpenVINO for Audacity on MuseHub
Description
Elevate your Audacity workflow with our suite of powerful tools. Isolate instruments and vocals with Music Separation, transforming any track into individual stems. Clean up your recordings with Noise Suppression, perfect for crystal-clear spoken word. Spark your creativity with Music Generation, crafting unique snippets or extending existing melodies. Effortlessly convert speech to text using Whisper Transcription, and enhance audio quality with Audio Super Resolution