Robotics researchers in Japan have shown off a new humanoid model that can belt out a tune with both natural-sounding vocals and the facial expressions to make it seem much more realistic.
The robot learns by mimicking a real human singer. The HRP-4C bot listens to the tune and watches the singer’s lips as she performs. Then the humanoid re-creates its own rendition, using software named VocalListener to synthesize its own vocal sound.
That’s run in unison with VocaWatcher, which analyzes a video of a human singer and picks out facial expressions and movements that go along with each note. Those are emulated by the robot’s synthetic face to make the robotic rendition far more natural.
Classical Music | Classical Music Video | Orchestra | Classical Musician | Concert Hall | Violin | Cello | Piano | Symphony | Viola | Concerto | Philharmonic | Opera Singers | Music Lesson | Choir | Concert Band | Music Teacher Directory