SONOTELLER is an experimental AI engine capable of 'listening' to music files, designed to simplify the way music enthusiasts understand and organize music.
By reading music files, it can provide lyrics, summaries, musical attributes, emotional atmosphere, as well as implement language recognition and flagging of inappropriate content.
Test 1
We conducted a test with an instrumental piece to see the results.
Test piece: Gould's performance of 'The Art of Fugue'
Identification results as shown in the picture:
The results include key words about the genre such as: Classical music, piano, baroque, orchestral, church music.
Sub-genres include:Fugue, polyphonic, keyboard, solo.
The emotions are calm, introspective, contemplative.
As can be seen, the identification results for this type of classical work are very accurate.
Test 2
Let's test with a piece of pop music next.
Test piece: 'You Belong With Me' by Taylor Swift
Identification results as shown in the picture:
It can be seen that the part about music analysis is quite accurate, but the accuracy of the lyrics analysis is not as good.
Perhaps this is because the song involves three characters, making it a bit complex for the AI.
Go ahead and try it with your favorite music.
