The 1995 Smithsonian collection "Big Band Renaissance: the Evolution of the Jazz Orchestra" is a
great example of group-source metadata FUBAR. dBPowerAmp's CD ripper program allows use of multiple
metadata sources, and by default does some sort of amalgam of whatever sources you've told it to
check. The amalgam on this set is comical! So I manually checked metadata from each source. They are
all different, and only GD3 (whatever that is) is anywhere near accurate. I find this often happens
with compilations -- for instance freedB and/or AllMusic will have different top-level stuff like
titles and whether or not it's a compilation for different individual CDs in the same box set.
All of this could have been prevented if the industry embraced CDText from the get-go and agreed on
uniform naming standards for artists and song titles. I remember the arguments back in the 80's --
it's hard enough to enter PQ codes into these balky Sony editing systems, and no CD players have
displays for CDText, so why bother. Very short-sighted. The net-net today is that anyone who wants
uniform naming and accurate information in a digital library has to spend a lot of time editing the
crappy metadata that's out there in group-source land. And, copyright owners have ceded control of
their metadata to a group-source no-QC cluter-you-know-what.
-- Tom Fine
|