neumes-align is a tool which allows to collate multiple MEI encodings
of single sources into one MEI neumes file – with an critical apparatūs
in all those places, where the sources differ from each other. 

From a technical perspective this tool relies on a rather famous
algorithm from the bioinformatics, the Needleman-Wunsch algorithm.
This algorithm was originally built to align two DNA or protein
sequences. This algorithm basically assigns a score to every possible
alignment and then backtracks through this matrix to find an optimal
path.

This algorithm can be be applied to aligning two sequences
of neumes. I’d like to give you a little demonstration of that:

[slide]

Here you can see three sources of the chant “Domine Jesu Christe bone”.
Two versions the chant itself and an Old Roman version. First,
I will only align the first two sources, A and B. The encodings
were obtained from the cantus database using the Volpiano translator.

[demo – align two sources and show result]

When aligning multiple sequences, the tool tries either to find an
identical reading in an existing apparatus, to append a new reading to it
or to create a completely new apparatus. 

I’d like to demonstrate that with a third source added to the previous
example, which deviates from the first two sources in almost every place. 

[demo – align three sources and show result]

Now you can see at least three different readings for every neume. 

We can now display the newly created MEI file with the different readings 
in the MEI neumes viewer which Fabian Köninger presented earlier. 

[demo]

Thank you very much for attention!