have been getting into making short reels for tiktok with my band, and I'm discovering that tiktok's speech detection algorithm is pretty lousy when it comes to detecting singing, if it even works at all.
Does anyone know if there is a plugin for fcpx that can detect speech? And if so, if it works halfway decently with music/singing? (probably worth a try anyway)
It's not a plugin but you can get the MacWhisper AI tool (which is a Mac app), export an SRT file, import that to FCP and it will produce subtitles for dialog or most song lyrics: goodsnooze.gumroad.com/l/macwhisper
The paid version is only 18 EU, and it is very powerful. There are various models which trade off conversion speed vs accuracy, but even the English "base" version is fast and fairly accurate. The "Large" model can produce transcripts from audio so poor I can barely understand it with my ears. Even the "Large" model is not that slow on a fast machine.
I tested it on a few English-language songs and it seemed to work OK. It can also do other languages, but it does not yet translate. That is planned for a future version.
+1 on recommending MacWhisper (Pro). It is a powerful app now, with a good roadmap for future enhancements. The developer, Jordi Bruin, is terrific, helpful, and has slate of other useful app/utilities.