Article19 V.8 Matt Gough Blog [closed]: beyond animation

"Hyperchoreography" Previous | Main | Next "Writing"

With robert (RoboLecturer, First Gypsy Demo), elizabeth (dimensions) and kema (via email) recently discussing motion capture it's worth expanding on the computer based notation i'm developing.

motion capture (mocap) records the visual impression of movement in 3D. the computer based movement notation (cbmn) records the 'motive drive' (dynamics, physics) of movement in 3D. where as mocap is fine for driving golems (avatars 1) in real time for animation, cbmn can be used for dynamic simulation.

two golems driven by (independent) mocap cannot engage in contact improvisation. not only is the data relatively 'fixed' but the lack of dynamic information means that the golems cannot interact physically. you could try to get round this by creating a large library of mocap clips that 'blend' together but the effect will still be visual rather than dynamic.

the 'realism' of mocap (even when just the markers) is due to our bias for perceiving biological motion [1,2] 2, we assume the dynamics involved or attempt to reconstruct them during our own simulation (learning) process 3. for this reason mocap is little better than multi angle video if we want to learn a movement sequence.

when we learn movement with a teacher they impart not only the visual effect /appearance of the movement, but the motive dynamics. cbmn describes both the visual and dynamic aspects of movement. as a dynamic movement notation cbmn has many advantages over mocap and video.

cbmn is a multi layered notation made up of a basic movement description, abstractions, and compound abstractions. even when using multiple compound abstractions the basic description can be used to 'tweak' the resulting movement. cbmn could be used by a wide range of people, there is no barrier apart from learning the how to combine abstractions (simple) or getting to grips with the basic description (harder).

it would be possible to implement existing notations with cbmn. this is a difficult process with motion capture, and would require massive amounts of motion data. with the cbmn approach it is much simpler to structure and manipulate the raw data to follow the rules of the donor notation. thus cbmn would simulate rather than animate the dance archive, recording the 'text' in 3d rather than the visual interpretation given by other digital methods.

motion capture was held up as the vital technology for virtual dance and motion in general. yet there are many motion capture rigs out there sitting idle, or being 'played' with. although motion capture seems like a great solution for digital dance, i'm not so sure. whilst it would help to speed up the notation process when encoding a technique a significant amount of extra information would have to be added to the system. it would be useful to add emg and eeg to mocap for more accurate translation to cbmn.

a few points,

it's important to remember that cbmn is movement notation, not dance notation. if you consider that the range of potential users is wide.
technology can speed up aspects of the creative process, for example acrylic paint can be applied faster than traditional oil paints yet the duration of the process is dependant on the artist . a painting may take days, months or years regrdless of the technology used.
digital dancers using cbmn for autonomous behavior may speed up the choreographic process. this is also not an issue of digital vs real dancers, i'm not suggesting that one is better than the other.. they offer different advantages.
key framing, is animation not motion simulation, digital dance needs to be more than animation - dance is dynamic (physics 4). a notation library only simplifies the process of writing, nothing else, but that's important so the end user can quickly get familiar with the system.
terminology is a big issue in my work, i am constantly finding conflicts between the wolds in which i work. being vague will not help the situation so i try to navigate my way through. i consider much of what i write targeted at dancers as i assume more dance knowledge than anything else. most of the linked terms are from other disciplines.
cbmn is a challenging project, but i believe in the long term it will prove to be a valuable undertaking.

one of the reasons i write this blog is that there is so little discussion or sharing of the nuts and bolts of dance technology theoreical practice. by sharing my work i hope to offer one way in which science and the arts are linked, and the relationship between theory and practice (practical theory).

1 i prefer the use of golem to avatar (other uses), the conceptual metaphor is more fitting.

2 models of biological motion manipulated by researchers may representation perception of stereotypes rather than inherent attributes see biomotionlab1.6 and the research pages.

3 see also in Calvo-Merino et al (2004) Action Observation and Acquired Motor Skills: An fMRI Study with Expert Dancers.

4 see also Physics and Dance

Filed by Article19 at 1:41 PM | Permanent Link | Share on Facebook |

Stumble It!