From motion capture and video recordings of dance, we developed methods that allow the description of gestures in terms of shapes and spatiotemporal reference frames. Starting from commercial audio excerpts and field recordings, we demonstrated that samba is characterized by particular micro-timing deviations, as well as an unclear configuration of periodicities in the metrical structure of music. In this study, we investigate how this gesture is modeled through sound and movement and how the interdependence between these modalities has influenced samba music and samba dances. Together, these experiences convey a sort of unified gesture, or cross-modal gesture, which is acquired and performed through sound, movement and other modalities. Music and dance cannot be easily dissociated in the culture of Afro-Brazilian samba. Keywords: Music driven dance video segmentation, multimodal Indian classical dance data captured by Kinect, Onset detection on Indian music, music-to-dance video synchronization
Hence the system has 100% precision, but only about 56% recall. From over 13000 input frames of 15 Adavu's, 74 of the 131 key-frames actually present get detected. Beats are tracked by onset-detection to determine the instants in the video where the dancer assumes key-postures. We use Blind Source Separation to isolate the instrumental sound from the vocal. We segment the videos of Adavu's according to the percussion beats to determine the events for recognition of Adavu's later.
It is a combination of events that are either postures or small movements synchronized with rhythmic pattern of beats or Taals. An Adavu is accompanied by percussion instruments (Tatta Palahai (wooden stick)-Tatta Kozhi (wooden block), Mridangam, Nagaswaram, Flute, Violin, or Veena) and vocal music. Adavu's are basic choreographic units of a dance sequence in Bharatanatyam. We present an algorithm for audio-guided segmentation of the Kinect videos of Adavu's in Bharatanatyam dance.
0 kommentar(er)
