Examples

We provide below some examples of the alternative approach proposed in this dissertation to singing voice detection. It involves the separation of harmonic sounds from the polyphonic audio mixture and their individual classification. As an outcome of this approach an automatic singing voice separation system is obtained. The separated sounds identified as singing voice can be subtracted from the polyphonic mixture providing a residual signal with the remaining sounds.

Two type of examples are included below, together with their corresponding audio files, namely singing voice detection and separation. The former compares the manual and automatic labels indicating time intervals in which singing voice is present. In addition, the pitch salience representation (F0gram) and the tracked pitch contours are shown. The latter, depicts spectrograms of the original, the vocal and the residual audio signals, in order to assess the separation.

Singing voice detection

Blues song excerpt from the testing dataset. It comprises singing voice, piano, bass and drums. Singing voice notes are correctly distinguished from piano and bass notes that also appear in the F0gram.
example bluesexample blues
Fragment of the song For no one by The Beatles. A singing voice in the beginning is followed by a French horn solo. There is a rather soft accompaniment which is almost not present in the F0gram.
example fornoone2example fornoone2
An excerpt of the file pop1.wav from the MIREX melody extraction test set. It consist of three simultaneous prominent singing voices in the first part followed by a single voice in the second part, and a rather soft accompaniment without percussion. The leading singing voice is correctly identified but the backing vocals are wrongly classified as instrumental sounds. Note that since close vocal regions are merged when producing the automatic labels the short pure instrumental interval is ignored.
example pop1example pop1
Jazz audio excerpt, with a leading saxophone, piano, bass and drums. The classification produces some false positives.
example jazz example jazz

Singing voice separation

The same Blues song excerpt introduced above, containing singing voice, piano, bass and drums.
example blues example blues
example blues example blues
example blues example blues
The same fragment of the song For no one by The Beatles. A singing voice in the beginning is followed by a French horn solo.
example fornoone2 example fornoone2
example fornoone2 example fornoone2
example fornoone2 example fornoone2

Additional singing voice separation examples

DescriptionOriginalVocalResidual
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio
example audio example audio example audio