Matt Gough Blog [closed]: dance first
Thursday, Nov 25 2004, 01:22
due to my general resistance of offering much (any) working code to CMP i'm being asked 'where's the computer science?'. i have a solid reason for this, most of the work i'm reading that's claims to had an arts edge falls over when you try hold it up against arts practice and theory. indeed often work presented at a Phd or Msc level contains concepts of art that would fail at undergraduate level arts studies.
there are many good and interesting tools out there for human movement simulation, well simulating body mechanics, but they are mostly put in the hands of people who don't really know how to use them and developed by people with a minimal understanding of movement practice. i'm not claiming that i'm going to come up with a total solution but i would like to help create a paradigm shift.
a blurry line between animation and simulation only clouds the issue further. i however, argue that if the goal is to enable interactivity, then simulation is the only way forward. yet animation perspectives have a stranglehold on simulation work, this leads to the odd assertion that human motion is distal in nature (ik) where as it is usually proximal (fk).
for example in the current paradigm a punch is guided by the fist, with the shoulder and elbow etc following paramaterised constraints. in a dynamic environment this will create a 'weak' punch. Punches are driven from the back and shoulder (depending on technique) with a specified alignment driving the elbow to fist chain. as humans we use a form of ik solving to target the punch but it is driven by fk. consider the fact that when we use stick based objects (bats, poles, pens etc) the far end of the object is under our control by manipulation of the near end.
there reason for this inversion is simple, it's the simplest way to animate existing character rigs. i've already noted here that most rigs (hierarchical skeletons) are under specified for not only dance but general human motion (simulation not animation). my position is not only that more advanced rigs are needed but also a different method of driving' such rigs. this is where a comprehensive method of computer based notation comes in. but developing the notation is not enough
danceforms (credo interactive) is a good example here, it's often critisised, in the past by myself too, but it's not the software at fault when you see those naff dance animations. it's the animator / choreographers. i digress. my point is that Rhonda Ryman's Ballet Moves is the bench mark for what should be supplied with dance animation software, a well developed corpus of movement notation.
whilst modern and contemporary dance is more challenging to create with danceforms (i'm exculding merce cunningham's work for obvious reasons) ballet moves allows the ballet choreorgapherto make work, rather than animate, for example
unexpected? it really shows what can be done with good software and related resources. whatever system i develop i intend to make it simple to create notation libraries. then it would be feasable to encode many dance techniques and have a swiss army knife of a choreography tool.
Filed by Article19 at 1:22 PM | 
Stumble It!