Sorry Sean -- I was wrong. There were a couple places here and there where we were losing precision by casting the BPM to a single-precision float. Fixed in lp:mixxx/1.11 r3696 by always treating the BPM as a double. Confirmed that the value maintains its full precision across restarts so it's correct in the DB as well.
Sorry Sean -- I was wrong. There were a couple places here and there where we were losing precision by casting the BPM to a single-precision float. Fixed in lp:mixxx/1.11 r3696 by always treating the BPM as a double. Confirmed that the value maintains its full precision across restarts so it's correct in the DB as well.