I want to identify the speakers in some recorded conversations. It’s a small group, about 10. I have large samples of their speech, so I could create reliable voiceprints for each.
Has anyone ideas for how to think about this?
I doubt any of the Audacity tools would be useful for this. The best we can do is the standard spectrum and frequency analysis tools and they wouldn’t get specific enough for what you want. You need software that “knows” what a voice is.
Plus, they were speaking in a quiet studio, right? Any background noise would pretty much doom the project. Audacity can’t split apart different voices, instruments or sounds.
Koz
The speakers were speaking in a quiet studio. I expect to identify only solo voices. I find a couple of open-source projects on-line, but they are very old. There are a few commercial products.
very old…
Do they work? We’re still using a sound processor whose developer has been dead and buried for a while now.
Koz