It would be useful to have a similarity check option, just like you can use [Artist] or [Genre], there should be one more forgiving than just adding duration to the mix. There should be a duration option with relative addition which would allow users to specify how many seconds different they have to be or they will be excluded.
If [Duration] is the current option (it may be a different word), then take this example:
-------------------------------------------------------------------------------------------------------
Artist:Coldplay----Song:Clocks----Album:A Rush Of Blood To The Head----Duration:5:07
Artist:Coldplay----Song:Clocks----Album:Artist Confidential (Volume 1)----Duration:5:09
Artist:Alanis Morissette----Song:You Learn----Album:Alanis Morissette Mtv Unplugges----Duration:4:21
Artist:Alanis Morissette----Song:You Learn----Album:Jagged Little Pill----Duration:3:59
Artist:Beach Boys----Song:409----Album:Beach Boys Greatest Hits (Volume 1)----Duration:2:00
Artist:Beach Boys----Song:409----Album:Greatest Car Songs----Duration:2:00
-------------------------------------------------------------------------------------------------------
Okay, now you'll notice that the first example with the song Clocks has two songs (both of which ARE different) but are very close in duration. The second example, You Learn, shows two different versions with GREATLY varied durations--that is good. Those two should undoubtedly be copied. The last example by the Beach Boys shows two identical songs with identical durations. Only one of those should be included.
As you can see, there are issues. Namely, when identical songs either are or are not the same duration. Unfortunately, there is no perfect solution. I guess it would be good to hear other people's ideas on the issue.
One idea would be to check to see if the album is a compilation. If it is a compilation, then it seems that there is a higher chance it is a live CD. Best yet would be a way for Media Center to examine the actual song contents and look for similarities. It can't just check average intensity or bitrate because with the song Small Town by John Mellencamp, the Scarecrow version shows a bitrate of 940 while The Best That I Could Do 1978-1988 shows a bitrate of 1008. The durations also vary by 2 seconds.
Any ideas on solving this issue? Copying duplicates is annoying because they waste space, increase the chance for listening to a certain song more, etc. I have a song or two with up to 5 copies where 3 or 4 of them are identical. I don't want to just rate the last 3 2 stars so they don't sync because I question that later and become confused.
Again, some form of audio fingerprinting would be quite nice but a lot of work.