Extracting Singing Voice from Music Recordings by Cascading Audio Decomposition Techniques
This is the accompanying website for the paper "Extracting Singing Voice from Music Recordings by Cascading Audio Decomposition Techniques" by Jonathan Driedger and Meinard Müller [pdf][bib].
The problem of extracting singing voice from music recordings has
received increasing research interest in recent years. Many proposed
decomposition techniques are based on one of the following two
strategies. The first approach is to directly decompose a given music
recording into one component for the singing voice and one for
the accompaniment by exploiting knowledge about specific characteristics
of singing voice. Procedures following the second approach
disassemble the recording into a large set of fine-grained components,
which are classified and reassembled afterwards to yield the
desired source estimates. In this paper, we propose a novel approach
that combines the strengths of both strategies. We first apply different
audio decomposition techniques in a cascaded fashion to disassemble
the music recording into a set of mid-level components.
This decomposition is fine enough to model various characteristics
of singing voice, but coarse enough to keep an explicit semantic
meaning of the components. These properties allow us to directly
reassemble the singing voice and the accompaniment from the components.
Our objective and subjective evaluations show that this
strategy can compete with state-of-the-art singing voice separation
algorithms and yields perceptually appealing results.
In the following we present audio examples to demonstrate the capabilities of our proposed procedure.
You can choose an example from the list and click on the stylized spectrograms to listen to the audio files.
SiSEC: Bearlin
Original
recording
►
Harmonic
component
►
Percussive
component
►
Residual
component
►
Harmonic portion
singing voice
►
Harmonic portion
accompaniment
►
Fricatives
singing voice
►
Instrument onsets
accompaniment
►
Vibrato & formants
singing voice
►
Diffuse instr. sounds
accompaniment
►
Estimate
singing voice
►
Estimate
accompaniment
►
SiSEC: Tamy
Original
recording
►
Harmonic
component
►
Percussive
component
►
Residual
component
►
Harmonic portion
singing voice
►
Harmonic portion
accompaniment
►
Fricatives
singing voice
►
Instrument onsets
accompaniment
►
Vibrato & formants
singing voice
►
Diffuse instr. sounds
accompaniment
►
Estimate
singing voice
►
Estimate
accompaniment
►
SiSEC: Another Dreamer
Original
recording
►
Harmonic
component
►
Percussive
component
►
Residual
component
►
Harmonic portion
singing voice
►
Harmonic portion
accompaniment
►
Fricatives
singing voice
►
Instrument onsets
accompaniment
►
Vibrato & formants
singing voice
►
Diffuse instr. sounds
accompaniment
►
Estimate
singing voice
►
Estimate
accompaniment
►
SiSEC: Fort Minor
Original
recording
►
Harmonic
component
►
Percussive
component
►
Residual
component
►
Harmonic portion
singing voice
►
Harmonic portion
accompaniment
►
Fricatives
singing voice
►
Instrument onsets
accompaniment
►
Vibrato & formants
singing voice
►
Diffuse instr. sounds
accompaniment
►
Estimate
singing voice
►
Estimate
accompaniment
►
SiSEC: Ultimate NZ Tour
Original
recording
►
Harmonic
component
►
Percussive
component
►
Residual
component
►
Harmonic portion
singing voice
►
Harmonic portion
accompaniment
►
Fricatives
singing voice
►
Instrument onsets
accompaniment
►
Vibrato & formants
singing voice
►
Diffuse instr. sounds
accompaniment
►
Estimate
singing voice
►
Estimate
accompaniment
►
Classical Opera: Weber - Der Freischütz
Original
recording
►
Harmonic
component
►
Percussive
component
►
Residual
component
►
Harmonic portion
singing voice
►
Harmonic portion
accompaniment
►
Fricatives
singing voice
►
Instrument onsets
accompaniment
►
Vibrato & formants
singing voice
►
Diffuse instr. sounds
accompaniment
►
Estimate
singing voice
►
Estimate
accompaniment
►
Romantic Piano Music: Schubert - Gute Nacht (Winterreise)
You can download all audio files which were used in the objective and subjective evaluation
here. Partially, these files were obtained from the
SiSEC 2013 website.