Gene Gajewski

Personal Blog

Close this search box.

Optical Confabulation


It is the Christmas Holidays once again, and of course, I am keeping busy with Christmas music-making as I always do. This year’s project is Tchaikovsky’s “The Nutcracker” music – the entire ballet performance – not just the 20 minute “Suite” we are all familiar with. The suite was an excerpt Tchaikovsky’s had taken from the ballet – the highlights that could be performed without the context of the ballet.

I have done a performance of The Nutcracker (“Suite”) in the past but using a midi file as a source. That was fun, but the ballet I am working on is much more complete and it is sourced from a sheet music print.

I am off to a good start, and I am halfway done and into the second act already. Great, but Man, I have had to deal with a lot of errors in the MuseScore source that I am converting over to Dorico. Since I cannot completely trust the MuseScore file, I went and sourced some scans of old prints at the International Music Score Library Project (IMSLP) / Petrucci Music Library. Bingo! They had two old prints, one in Russian and another in French – I downloaded them both.

Really, you need both for comparison purposes because they are from different publishers and printings and there are some differences between the two. The French version labels the instruments on a staff only on the first page and at system breaks that have a different number of staves, and there are no bar numbers anywhere. The Russian version puts instrument labels for every staff on every page, but also has no bar number numbers anywhere, yet it does put a rehearsal mark the top of every ten bars. These full score prints are condensed, meaning many of the string section divisi parts are on a single staff instead of two, and instruments with empty bars are not printed.

I am using the Russian version as a main reference because it is easier to follow along, the French version requires you to remember which staff is which instrument on a page, and if you’re confused, you need to go back to the last forced system-break.

This is that little bugger that you need to find to get the correct staff labels in the French version – the convention here is to put this mark and print staff labels only when the number of staves in a system changes. This is because the engraver would drop the staves for instruments with empty bars to conserve space. The Russian print does too, but at least they put staff labels on every page, so it obviates the need to remember which stave is what instrument.

So, while working with the score in Dorico, I managed to run into a few places where I just could not help but question the integrity of some of the bars – something just did not look right. When I get spooked like that, it is usually for a good reason. Why would some string staves have bars with notes with two slashes, and others staves of the same bar have notes with three? Three slashes are always an unmeasured tremolo whereas two slashes would mean to break note into four – the resulting sound (and rhythm) is very different between these two notations.

Looking back at the MuseScore source was not any help – Dorico had exactly what MuseScore did! I went back to look at the Russian scan, and although it was a quite grungy, it would seem to me to be, although indistinct, three slashes to the size of the blob of ink at that position. Still, I went to look at the French scan and then I could see where, yes, it might look like only two slashes for the notes in question. Then it dawned on me, The MuseScore digital source that I converted for Dorico, must have come from this exact scan of the French print!

It is highly likely the author, or shall I say copyist of the MuseScore Nutcracker file, used an optical music reader program to get the score into MuseScore using the old French scan. That is super cool and the thing to do with a score of this size, but you cannot trust those optical readers on crappy, grainy scans of complicated scores. They make bad guesses all the time and you must manually correct every one of those errors. The copyist missed this run of errors then.

Anyway, I now know I can trust my spidey-sense when it comes to old music derived from old scans. If it does not feel right, something is up to no good. Here are the reference scans I am using.

Pretty grungy, no? Also, notice the Russian scan is over a hundred pages longer than the French scan – which means the French version crammed more bars on their pages (which makes it that much harder to optically scan). Oh, and did I mention – no bar numbers? Yeesh!

Leave a Reply

Share the Post: