I have found, many times, that Audiveris is inconsistent in recognizing multiple in-staff voices (in guitar music, in my case). Sometimes it works, sometimes it flags the measure with an error. Example:
In this case (common in guitar music) we have a middle voice appearing from time to time, representing a finger/string following its own rhythm. The OCR is absolutely correct, no errors. But the voice assignment is inconsistent. In fact, nearly identical measures are recognized correctly and not.
I don't know of a simple way to tell Audiveris "Yo, this measure has a middle voice." I can hack measures, of course, deleting and adding notes and rests, but that's a big chore on a big score on which this has happened regularly.
Note that Audiveris is nearly flawless on two-voice guitar music, presumably because the stems are in opposite directions. But additional (middle) voices have to share stem direction with the bass and soprano voices.
I have found, many times, that Audiveris is inconsistent in recognizing multiple in-staff voices (in guitar music, in my case). Sometimes it works, sometimes it flags the measure with an error. Example:
In this case (common in guitar music) we have a middle voice appearing from time to time, representing a finger/string following its own rhythm. The OCR is absolutely correct, no errors. But the voice assignment is inconsistent. In fact, nearly identical measures are recognized correctly and not.
I don't know of a simple way to tell Audiveris "Yo, this measure has a middle voice." I can hack measures, of course, deleting and adding notes and rests, but that's a big chore on a big score on which this has happened regularly.
Note that Audiveris is nearly flawless on two-voice guitar music, presumably because the stems are in opposite directions. But additional (middle) voices have to share stem direction with the bass and soprano voices.