I had some problems upgrading my MM3 installation to MM4, so I ended up doing a clean (portable) install and just rescanning my media folders to rebuild the library. Then I installed autorate and let it scan. To my surprise, only a fraction of my library gets ratings. Turns out that my play history, play counts, and skip counts either aren't carried over or aren't picked up by the new version of autorate. My only rated songs are what I've played since recreating the library. That surprised me, as I thought such info was saved in tags. It's also a bummer, as I have years worth of history, and could generate high value playlists based on it. In both the old and new install, autorate is using Custom 5 and I can see there is info written there, even on a file picked at random from the "unknown" rating songs.
Does anyone have advice on exporting play info from my old install (which still functions, though it's become sluggish since the failed upgrade) and importing it into a MM4 library? Failing that, can I force autorate to base ratings on the info already in Custom 5, then update going forward? If no one has a better suggestion, my brute force workaround will be to manually add plays based on playlists form my old library: one play for everything with a count of at least 1, one more for everything played in the last 6 months, then 1 for everything rated 1 star, 2 for 2 stars, etc. I'm not very happy with that kludge but at least it would give shape to my library.
Unrelated question: I noticed that the new autorate beta includes the option to use play percentage instead of a binary count. I love that feature, as I have lots of well liked live music that includes dead air at the end of the track--previously I had to weigh sitting through it or skipping and throwing off my ratings. Am I correct to assume that play history synced from my iPod won't have percentage info, and will fall back to the playcount method? I understand the iPod is limited in how it counts plays/skips and what info it passes on sync. Related to this, is there a resource that describes the alternative rating algorithms in greater detail?
Thanks for any and all replies, Mike