| Music annotation is an indispensable and important part of the development and iteration of AI music.People use DAW and other audio tools to annotate music content such as melody,harmony,musical form,tonality,etc.,produce and generate a large amount of annotated music content,and then use these data to feed the model to achieve iteration.On one hand,scientific music date annotation is conductive to data integration and statistics,and on the other hand,it is also convenient for developers to learn the effective music information provided in music annotation based on their own different scripts,thereby improving the efficiency of development and iteration.Nowadays,those who make music annotations generally use mainstream DAW,such as Cubase,Logic Pro,or other audio tools,like Sonic Visualiser,NIAONiao,for data annotation and integration.The diversity of platforms,the low compatibility of export data formats,and the different operating systems based on the software itself all add complicated procedures and difficulties toe the work or music data labeling.Therefore,this thesis explores the possibility of using mainstream DAWs and online DAWs for music annotation,compares their functions horizontally and vertically,and finally proposes and idea of a fully functional and integrated music annotation tool. |
