Corpus Choice When Guiding Factor Oracle Improvisation With a Probabilistic Model

Here, we compare improvisations generated with a Factor Oracle combined with a probabilistic model trained on different corpora.

The probabilistic model is an optimised interpolation of smoothed sub-models learnt on a corpus. In the following examples, the probabilistic sub-models used are a bigram on the melody and a model representing the correlations between the melody and the harmony.

We generated improvisation using as a training corpus for the probabilistic model :

• Charlie Parker's Omnibook (database available here),
• a corpus consisting of about a thousand classical music tunes.

In all cases, the Factor Oracles are constructed on a single tune from Charlie Parker's Omnibook.