AUDIO EXAMPLES for “Remixing Stereo Music with Score-Informed Source Separation” 

Here are some example audio files showing the results of our source separation on an example 4-part Bach chorale played by clarinets.  All recordings are stereo 11k, 16bits.


descriptions:

'badBalanceMix.wav' - The "original" mixture that we are remixing. I tried to make the balance noticably poor, while still keeping all of the instruments loud enough to isolate them effectively.

'sop-ASE, alto-ASE...' - The individual parts, as separated by the combined algorithm (our ASE source separation method combined with an aligned score)

'remix-fixBalance.wav' - This is the separated tracks, stuck back together into the same places they were in the original mix, but with adjusted levels

'remix-hardpan.wav' - This is a second remix, mixed in digital performer using amplitude panning (so, a mixing board rather than my matlab mixing function) with some compression applied to the individual tracks

'remix-chorus.wav' - This is the same as the 'hardpan' one, but with really stupid sounding chorus applied to the soprano and bass parts.

'remix-reverb.wav' - The same as 'hardpan', except reverb is first applied to the soprano, then turned off, then applied to the tenor, then turned off.

'remix-loops.wav' - Just little loops of each of the parts, with a little verb on the whole thing. The bass part is pitched down an octave.