Randomized dances
Not long after I started dancing Lindy Hop I was taking part in a modern choreography course, and I found out about dance notation . In our class, we specifically used Laban's concepts to create a randomized dance (where the length of time, and various movement qualities and spacial relationships were chosen by lottery). Since then I have had a pipe dream of generating randomized Lindy Hop dances.
My first obstacle back then was that I could not identify the fundaments of the dance. The system seemed way too complicated, especially since my approach was to treat each "move" as a separate movement choice. Unfortunately when I tried listing them all, the impossibility became pretty apparent. Now I was not so completely naive. However trying to list all the possible orientations of the lead and follow to each other seemed (and still feel) very daunting to tackle. Especially to then create a dance program that could stand a chance at creating balanced dances.
I eventually set the project aside, as the more I learned to dance, the more I realized I did not know how the dance functions. In that respect, now I feel that I have a much better sense of what goes into connected dancing. Creating a program that is innovative and takes account the entire vocabulary of Lindy Hop still is pretty far from what I think can be done easily. However, now I see a clear foothold.
My approach would be based off of various exercises and movement training that either I learned first hand, or from someone who attributed these ideas to him. This is the order in which I would add features to my dance choreographing program. These ideas are incidentally make up pretty much exactly how Anna Iakshina and I taught a class at the Lindy Hop Spring Festival in Poznań, Poland.
1) I would first give it the option of walking forwards with a groove walk and walking backwards with a groove walk in side by side. This would be built in with two possibilities of changing the direction with a double bounce or rock step. So we have two states, and from each state we have three choices, return to the same state via groove walk, or change directions via rock step or double bounce.
2) We add a second state which we will be a close position with the couple facing each other. This system is exactly the same as we described where we can walk forwards or backwards, and change directions in those two ways.
3) We add the possibility of moving between the side by side movement in step 1, and the front to front position in step 2. There are two fundamental ways of getting from side by side to front to front: lead goes in front or follow goes in front. And there are two ways of getting back to side by side: lead goes to the side, or follow goes to the side. This movement in our system will take two counts, some conventional wisdom informs us that it is four counts, but really in two counts the fundamental part of getting into the position is completed, and we are ready for more choices (so far, keep walking or change directions, or do another lead/follow to the front/side). Everything done forwards has an equivalent moving backwards, none of these movements in and of themselves reverse a direction.
I feel that at this point we have a robust system in terms of creating interesting dance shapes in Lindy Hop. So now it is time to let the computer choreograph some movement quality.
4) Now in addition to all these directional decisions we have provided let us now give the option of creating dynamic changes and rhythms. The simplest is that every groove walk can also be a kick-step, and every rock step can be a kick. Also a groove walk can become a triple step, and a rock step can become a kick-ball change.
So that is quite enough choices for a first prototype. This prototype in fact is much more robust than it seems. It certainly can generate a Lindy Circle for you using this method:
Lindy Circle instructions: Start in side by side: rock step, Leader in front with a triple step, rock step, follower to the side with triple step.
So let's break down how we can construct and easy to understand way of how the choices will work:
We have the following table of movement possibilities:
movement
choices
Walk
Groove, Kick-step, Triple-step
quick change direction
rock-step, kick, kick ball-change-
The choices for "Lead in Front," "Slow Change Direction," "Lead/Follow in Front" "Lead/Follow to Side" are the same as for "Walk."
Really they are a walk with a change of orientation or direction. So there are better way to organize. What clearly is the fudnamental approach, however, is to split the dances into two count chunks. At the end of each chunk we either end up on the same foot (take even number of steps) or on the other foot (take odd number of steps). Right now our "walk" involves changing from one foot to the other in two beats of music, and so the way of doing that can change energies (kick vs. groove walk) or change the number of steps (triple step is still odd, but has 3 rather than 1 step). The reasonable even choices will be 0, 2 and 4 steps, but all our rock- step variations involve just two steps in various rhythms (kick-ball change vs. rock-step) or number of steps or number of steps (kick with 0 steps vs. rock step with 2 steps).
So this is a mouth full. This is also pretty much an outline for a computer program which can make these choices. Once it is made it has to be tweaked to even out the probability of various choices to be made.
This, and perhaps a video of a randomized dance in the near future. If I do have more quality stuff to say, I may reboot my dance blog.
My first obstacle back then was that I could not identify the fundaments of the dance. The system seemed way too complicated, especially since my approach was to treat each "move" as a separate movement choice. Unfortunately when I tried listing them all, the impossibility became pretty apparent. Now I was not so completely naive. However trying to list all the possible orientations of the lead and follow to each other seemed (and still feel) very daunting to tackle. Especially to then create a dance program that could stand a chance at creating balanced dances.
I eventually set the project aside, as the more I learned to dance, the more I realized I did not know how the dance functions. In that respect, now I feel that I have a much better sense of what goes into connected dancing. Creating a program that is innovative and takes account the entire vocabulary of Lindy Hop still is pretty far from what I think can be done easily. However, now I see a clear foothold.
My approach would be based off of various exercises and movement training that either I learned first hand, or from someone who attributed these ideas to him. This is the order in which I would add features to my dance choreographing program. These ideas are incidentally make up pretty much exactly how Anna Iakshina and I taught a class at the Lindy Hop Spring Festival in Poznań, Poland.
1) I would first give it the option of walking forwards with a groove walk and walking backwards with a groove walk in side by side. This would be built in with two possibilities of changing the direction with a double bounce or rock step. So we have two states, and from each state we have three choices, return to the same state via groove walk, or change directions via rock step or double bounce.
2) We add a second state which we will be a close position with the couple facing each other. This system is exactly the same as we described where we can walk forwards or backwards, and change directions in those two ways.
3) We add the possibility of moving between the side by side movement in step 1, and the front to front position in step 2. There are two fundamental ways of getting from side by side to front to front: lead goes in front or follow goes in front. And there are two ways of getting back to side by side: lead goes to the side, or follow goes to the side. This movement in our system will take two counts, some conventional wisdom informs us that it is four counts, but really in two counts the fundamental part of getting into the position is completed, and we are ready for more choices (so far, keep walking or change directions, or do another lead/follow to the front/side). Everything done forwards has an equivalent moving backwards, none of these movements in and of themselves reverse a direction.
I feel that at this point we have a robust system in terms of creating interesting dance shapes in Lindy Hop. So now it is time to let the computer choreograph some movement quality.
4) Now in addition to all these directional decisions we have provided let us now give the option of creating dynamic changes and rhythms. The simplest is that every groove walk can also be a kick-step, and every rock step can be a kick. Also a groove walk can become a triple step, and a rock step can become a kick-ball change.
So that is quite enough choices for a first prototype. This prototype in fact is much more robust than it seems. It certainly can generate a Lindy Circle for you using this method:
Lindy Circle instructions: Start in side by side: rock step, Leader in front with a triple step, rock step, follower to the side with triple step.
So let's break down how we can construct and easy to understand way of how the choices will work:
We have the following table of movement possibilities:
movement
choices
Walk
Groove, Kick-step, Triple-step
quick change direction
rock-step, kick, kick ball-change-
The choices for "Lead in Front," "Slow Change Direction," "Lead/Follow in Front" "Lead/Follow to Side" are the same as for "Walk."
Really they are a walk with a change of orientation or direction. So there are better way to organize. What clearly is the fudnamental approach, however, is to split the dances into two count chunks. At the end of each chunk we either end up on the same foot (take even number of steps) or on the other foot (take odd number of steps). Right now our "walk" involves changing from one foot to the other in two beats of music, and so the way of doing that can change energies (kick vs. groove walk) or change the number of steps (triple step is still odd, but has 3 rather than 1 step). The reasonable even choices will be 0, 2 and 4 steps, but all our rock- step variations involve just two steps in various rhythms (kick-ball change vs. rock-step) or number of steps or number of steps (kick with 0 steps vs. rock step with 2 steps).
So this is a mouth full. This is also pretty much an outline for a computer program which can make these choices. Once it is made it has to be tweaked to even out the probability of various choices to be made.
This, and perhaps a video of a randomized dance in the near future. If I do have more quality stuff to say, I may reboot my dance blog.