Adding note alignments computed with the `synctoolbox`

This PR adds an additional set of notes TSV files which correspond to unfolded/expanded versions of the 303 chorale settings that have a counterpart in and have been aligned with the recordings made by the Chamber Choir Of Europe under Nicol Matt in 1999 (e.g., CDs 122 through 127 (six CDs) of the release https://musicbrainz.org/release/5582b212-aea7-4355-8c4c-531ed438e5fc).
Unfolded/expanded, here, means that they correspond to a "playthrough" respecting repeat signs and first/second endings.
This representation is a prerequisite for aligning the notes with the corresponding sounding events in the recordings.

First, the MusicBrainz recording IDs contained in the column mb:recording of the file metadata.tsv were used to map the 303 scores (or rather, their unfolded notes TSVs) to the respective audio files (which are commercial and cannot be provided). Silence at the end of the audio files has been truncated beforehand using a 50 ms & -65 dB threshold.

Then, the latest version of the synctoolbox (1.3.2) and this Python script was used to compute start and end points of each note. The timecodes were added as two additional columns, start and end, to the unfolded notes TSVs.

Adding Musicbrainz work IDs and aligned recording IDs from the Chamber Choir Of Europe

The work IDs were easy to align with the 389 chorale settings thanks to the series available on Musicbrainz: https://musicbrainz.org/series/1b6aec18-db22-4f91-98fb-2a5770183374

In a second step, the recordings by the Chamber Choir Of Europe under Nicol Matt made in 1999 and released multiple times by the label Brilliant Classics (Leeuwarden, NL) were aligned according to CDs 122 through 127 (six CDs) of the release https://musicbrainz.org/release/5582b212-aea7-4355-8c4c-531ed438e5fc .

On Spotify, the same recordings are available, re-packaged into three "albums":

1. https://open.spotify.com/album/2b3Sia98rloJFFNlf3I49K
2. https://open.spotify.com/album/100T8SEVcNbGbJPxlWnR2Y
3. https://open.spotify.com/album/5GSiBNk22KHfUTBYUfKr6G

For their alignment with this score dataset, the tracks were added to a Spotify playlist in order to download their Spotify IDs and ISRC IDs using the free service https://www.chosic.com/spotify-playlist-exporter/

After integrating the aligned IDs in the two metadata.tsv files, they were written into the "Score properties" of the respective MuseScore files.

Adds alternative version with vocal parts only

This was achieved by

• copying the folder original_complete to vocal_parts_only
• removing parts
• removing instrumental staves
• removing cue notes

The score properties section in the README keeps note of which chorales were affected by these changes.

Furthermore,

• two of the three chorales that had four vocal staves have been reduced to two staves for homogeneity with the remaining ones;
• staff names have been homogenized.

Updates measure numbers

To make the measure numbers displayed in the scores comply with a wide-spread convention in music printing, all anacrustic measures have been systematically excluded from incrementing the counter. Likewise, second endings have been adapted so that their measure numbers comply with the a)-b) convention (i.e., in general, first and second have the same measure number).

A new column in the README table (Measure numbers) marks those pieces for which one or several measure numbers changed compared to the previous version of the corpus.

The original dataset

Scores as provided at https://gertim-alberda.com/chorales/ but converted to MuseScore 3.6.2 and fully extracted with ms3 2.5.1.