Our new e-book follows on from previous work regarding algorithms for measuring the similarity of marks (focusing on the cases of colour and word marks which are amenable to exact definition and thereby lend themselves to objective quantitative comparison). This follow-up considers the case of sound marks (and other melodic sequences), which could also benefit from a more objective framework, to potentially augment or replace the existing scenario in which comparisons are primarily made on the basis of the subjective opinions of expert musicologists. A quantitative framework of this type has potential applications in the comparison of sound marks, and in the measurement of similarity between musical compositions, as is relevant to copyright disputes.

The proposed framework is directly applicable to melodies which are representable as sheet-music snippets, and comprises an ‘encoding’ of the melody as a string of text characters (representing both the relative pitches and relative lengths of the notes), together with the use of an algorithm previously proposed for use in word-mark comparisons (in this case, a library algorithm based on the concept of Levenshtein distance), to quantify the degree of similarity between the textual representations. The paper is illustrated using case studies, showing how the framework can be applied to produce quantitative measurements of the degree of similarity between two melodic lines.

A similar approach could potentially also be used for analysis of more complex musical elements, or in the comparison of chord progressions. Additionally, the specific configuration of the framework could be modified, based on the exact musical features requiring analysis. The option for applying ‘correction factors’, to take account of the commonness of use of musical sequences, to ‘offset’ the measured similarity, is also considered.

There is also potential for additional future development of the framework, potentially encompassing more complex melodic comparison concepts (such as the use of contour- or shape-similarity measurement ideas, or the analysis of N- grams – short combinations of notes comprising the basic ‘building blocks’ of more complex melodic lines, in some ways analogous to the concept of ‘tokens’ in word-mark analysis).

Going forward, as the representation of sound marks is likely to move more toward the use of digital files such as MP3s, it is likely that the use of other concepts, such as the use of file ‘hashes’, may also need to be explored.

You can download the e-book here.