Oktoechos Classification and Generation of Liturgical Music using Deep Learning Frameworks

An important feature of the music repertoire Syrian tradition is system classifying melodies into eight tunes,  called ’oktoe\={c}hos’.  In oktoe\={c}hos tradition, liturgical hymns are sung in modes or colours (known as ’niram’ Indian tradition). this paper, recurrent neural network (RNN) models are  used for  genre classification with help musical texture features (MTF) and i-vectors.The performance proposed approaches evaluated using a newly created corpus South language, Malayalam. Long short-term memory (LSTM)-based gated unit(GRU)-based experiments report average accuracy of  83.76\%  77.77\%, respectively, significant margin over i-vector-DNN framework.   The demonstrate potential RNN learning temporal information through MTF recognizing system. Furthermore, since Greek liturgy Gregorian chant also share similar traits musicological insights observed can potentially be applied to those traditions. Generation style has been discussed an encoder-decoder framework. The quality generated files a  perception test.