@ghellquist, You ought to try one of the Musescore “proof reading” plugins. I’ve ported Dorico-written music to Musescore via XML to check for parallels. The plugin I use (“check for parallel 5ths/8ves”) puts a bit of text near each one, and if I recall, colors the notes that are involved. I don’t think the algorithm they use is overly complicated.
I am giving a try to the parallels plugins for Musescore in these days, through XML export.
It seems reasonably good at spotting both direct and hidden 5ths and 8ves, and even consecutive ones, which are harder to detect through sequential midi analysis.
I also believe, like @ghellquist and as I said before in this thread, that to really achieve a quantum leap in the check accuracy we will need a sort of heuristic AI program. Just like for humans, it is relatively straightforward to discover parallels in a mostly homophonic chorale writing or solo+accompaniment arrangement, but is much more challenging in a true concertato style, when layered counter-rhythms are in place.
In my long previous experience, the parallels checker of Sibelius is absolutely dreadful and non-dependable, to the point of being more a hindrance than a help in the revision phase.
There are, according to the MuseScore plug-in, “hidden” 5ths and 8ths in the works of every composer I’ve thrown at it, from the Renaissance to the Romantic; so I feel secure in ignoring them in my own.
Interestingly, Finale also had a plug-in for checking consecutives, which was also unreliable, and was removed in a recent cull of ineffective plug-ins.
1 Like
I totally agree with your approach @benwiggy: I too reckon, both by repertoire analysis and compositional experience, that every kind of music involving some kind of counterpoint or concertato literally lives on hidden parallels, without which in most cases it would be simply impossible to write, with respect to its main structural priority, that is thematic interweaving of repeated statements.