Dear Antonio, your work and your answers comments deserve real attention, I'm sorry it took a long time. Here are my thoughts on the issues you raised: First of all let me repeat that for me there is a difference between movement and gesture that is essential. To b really short a movement is the change of spatial position of a joint in relation to a reference system (e.g. 3D coordinates). On the other side all the definitions of gesture are leading to the question of meaning. A gesture, for most of the dictionaries I've seen is a movement that has a particular meaning for an ethnic group. For example there are movements in a mechanical watch or in a locomotive, but there is no gesture. If all this is true what, in human body movement, is bringing meaning? For Reed, Godard, Berthoz etc. etc. the meaning is brought, from a physical point of view, by the tension that the premovement (Reed) or (I'm translating from french) Adaptive Postural Anticipation (the so called A.P.A.) actually operate by coloring, distorting, tuning and allowing the voluntary movement. I'll stop here in order not to go too del in medical and psychological stuff. I know that you know what i mean: we discussed it during lunch in Paris). >1. "Microdances". We introduced this term to refer to a movement >gesture (eg a dance step) of a duration usually of 20-30s to >1minute, therefore short enough to be analysed "scientifically" >(performed repeatedly) and long enough to embed significant >expressive content/intentions. If this hypothesis is true the length problem is, in my opinion, irrelevant in quantity. I try to explain: a relevant (meaningful or "significant", as you say) gesture can be very short (one sec or even less) or very long (three minutes in certains buthoh choreograpies for instance). Another way to say this is: we have to allocate machine "capture time" according to a meaning we have read in the movement. We have to divide the dance sequance in meningful slices and work on them. But I assume that your microdances division based on time is just a way of working with a relatively short sequence. Very pragmatical. And this is, of course, all right to me. > Segmentation assumes some meaning given a context. For example, we >could segment on peaks of velocity, on rests (with proper definition >of rest!) etc. according to the context. A our paper on these >aspects will be available next week, and on the EyesWeb release of >May 7 (2.5) we'll deliver some demo patches doing some of these >tasks on sample videos fragments. Yes, I agree with you, but choosing peaks of velocity can be (as you point out in an implicit way) irrelevant. Imagine a movement needing constant acceleration and deceleration like from rest to spinning to rest again. And the rest question (you're smart :-)) is of course a whole issue: there are ways of resting that are continuos variations of body tensions and meaning transmission. So there is no rest like there is no stillness..... Both concepts are, of course, relative to precedent or following movements. >2. "Expressiveness": I agree on the 6 domains (muscle, skeleton, >eeg, dyaphram...). But let us consider the perspective of a >spectator: a spectator of a dance performance does not have access >to several of such layers, but has only a visual input, usually at >a significant distance. For example, a spectator's perceptual >system is not able as all to measure motion capture data as those >measured by commercial systems (eg ascension, vicon etc or >bioengineering/biomechanic systems). Probably a spectator has access >to other different internal (eg mental model) and external >"domains". Nevertheless a spectator can be "emotionally engaged", be >"aroused" by such a visual stimulus. So, the perspective is not on >the arousal/expressiveness/emotional engagement in the dancer but on >what is communicated e.g. to a spectator. Well, I agree on the emotional engagement of the spectator not needing deep analysis. But, still the analysis is there in order the emotion to arise. In my opinion you're right when you say that a spectator doesn't have the richness and precision of mocap data (or eeg) but he doesn't need it. Perception is much richer than mocap +eeg + whatever thanks to what John Martins called Metakinesis in his book "Modern Dance" (New York, A. S. Barnes and company, inc., 1933). (By the way, for those not following medical research, Metakinesis has been confirmed by a scientist who discovered the so called "mirror neurons"...) So the spectator is making a POWERFUL even though mostly unconscious analysis. One of our main concerns in Paris 8 university is to understand these models in dance perception. >(I think this spectator-centered perspective might be also very >interesting from an aesthetic/artistic point of view) >In this direction, for example, there is a significant work by >psychologists in which there are some hypotheses on "cues" in >movement (e.g. energy/time dimensions) which are responsible of the >perception of emotional content (circumplex models etc). Yes, I agree, but the psychologists works you cite (I'd like to know to whom you refer to) can also be dangerous if they want to recreate a universal model (a "religious" perspective most of the time) like Bartenieff "Foundamentals" or other kind of archetypes. My opinion is very close to Daniel Sterne : we all have a personal "profile of activation". So, the models are essentially personal. But there are some families linled to cutural values. >As for the cultural aspect, let us consider the following case: the >same dancer performs the same "microdance" (or if you prefer the >same "dance") in two different conditions, one in which he/she is >not fully inspired and one in which he/she is. >So the cultural context is exactly the same. Cultural context is not only the century and the country. The context is extremely variable even through tonus modulation) If you ask the same dancer to perform a microdance and then you ask someone to scream loudly for a few seconds and then you ask him/her to perform again immediately. You won't have the same capture in the machine. The scream changed the context and the tonicity of the dancer. A gesture is modulating the space and the space modulation is influencing again the tonus modulation. A gesture and a context are in constant feedback. That's way the cultural context is not the same. Of course it is the same if you consider a larger period of time (like a week) but the context is like when you make pizza paste. Always on the move in the short term. Or, let me put it this way: from the gesture point of view (Hic et nunc) a simple act in a space is enough to transform it completely. And a gesture is performed in a space, so it "drinks" all the flavour of what happened. (Imagine to go capture a same microdance in your lab and the day after in a church) >If we measure that spectators are "aroused/engaged" only in the >second performance (when the dancer is inspired), it would be >interesting to try to understand/measure what is the difference. >In short, the problem of facing expressiveness/emotion is huge and >probably will not be possible to be solved, nevertheless I deem that >we can make some little steps in the direction of understanding >something more of current state of the art. I really believe that you can do something. But machine Movement analysis must be done in parallel, in my opinion, with some body movement analyst done by a human (we strike back!) that is able to tell you differences that are not readable by actual technology in terms of meaning in order to double check. >Of course we are aware that, if successful, in our project we'll be >only able to understand only a very small part of such a very >complex phenomenon, but we think it is worth to explore and we are >very interested in collaborations in this directions. Quando vuoi (sorry guys, Italians can sometimes be impolite) >As for concrete artistic applications, these have driven our work in >EyesWeb implementation. It is my convinction that an augmented >sensitivity to more subtle and focused "cues" in movement (and in >audio) can enable much more (artistically) interesting scenarios for >interactive systems, in terms of mapping deeper movement gesture >params on visual and acoustic outputs. Even though I'm on mac (Love that admits no betraying) I think that your project is great and the music connection is maybe a "Passage to India" that can bring more consciousness of the dancer movement for selfcontrol and better performance (think about Rockeby articles about sensation after use of VNS). Feedback is definitely my central keyword for the moment. >Further, the visualisation of movement cues seems to be relevant for >the dancers we worked with (eg in terms of understanding and >"seeing" some subtle movement cues, maybe for improving learning and >self-awareness in certain difficult movement patterns) as well as >with therapy and rehabilitation (we are working with medical teams >facing two different diseases: Parkinson and severely handicapped >children). > Yes. Again, think about Rockeby Or Susan Kozel experience with magnetic mocap showing her movement in real time with a body of cubes and balloons.... >3. Macro: I mentioned an upcoming extension of EyesWeb concerning >the inclusion of "macros". Just to clarify, this concept of "macro" >is at a software implementation/design level, so it has nothing to >do with the concept of "macro" mentioned by Armando. "Macro" in our >system is only a way to incapsulate a process which can be >genealised (e.g. the extraction of a dancer's body silhouette from >an input image from videocamera), in order to reuse in a user's >patch (application). In this way, an application can be built in >terms of assembly of macros, thus reducing the technical skills >required to build EyesWeb applications. Yes, of course. The possibility that you have in Eyesweb to build customized object (like for Mark's FANTASTIC Isadora) and macros is definitely a great feature. I'm sorry tomake a late and verbose answer Ciao -- ___________ Armando
This archive was generated by hypermail 2b30 : 04/09/02