TechCrunch has a post on Pluggd, this is a searchable speech recognition service for audio.
What happens is that you can search an audio file for keywords, it then highlights where those keywords appear on a timeline (also highlights related terms)…this is great if you have a 1 hour audio file and you just want to hear a section you are interested in.
This is granular sections of an audio file, kind of like paragraphs in a story, and I believe you can even tag these sections.
Wow, social search for parts of audio files, search full-text of an audio files, and search not only the file by tag, but parts of the file by tag…this is as granular as it gets.
Hmmm…I mentioned something like this in my post, Granular tagging or index tagging (see “More” section).