*Dancers rehearsing with Michael Montanaro, Chicago Gray Center for Arts and Inquiry (2013)*

## "*Le mieux est l'ennemi du bien*"

**Sha Xin Wei: A Series of Temporal Texture and Rhythm Games**

*(*

**See also the etudes from Einstein's Dreams Time-conditioning Studio at the Topological Media Lab, Concordia University Montreal, 2013.**)One idea I’ve tried with Doug and Navid, inspired from John McCallum’s decryption of his percussion piece: Beam (radio) continuous sound or clicktracks synthesized from the sounds of footfalls (or other on-body sound). Ask for certain games: such as ask 6 or more people to invent halting, or false cadence, or shifting the leader games, or other anticipation / retrospection games. Contrast the experience of beaming click tracks from a single computer versus participants providing click tracks to one another . This is one form of relativity.

Second: Make the clicks NOT functions of ego-data, but of relations between ( parts of ) bodies (or objects in motion). Start with distance between bodies, NEVER the position or shape of a single body. More technically challenging: measuring proximity of parts of bodies (or a set of quotidian, sensor-tracked objects in iStage.)

The task these last days before the workshop is to come up with more exercises at the level of children’s games. (See Bancroft books on children’s games and movement games from the 1930’s. )

Ideally in week 1 we can do some games with no dependency on any computational media. (Week 1’s creative exercises do not exclude the computer, just ignore what the media programmers are doing :)

At the end of Week 1, I would like to plan some etudes in week 2 that do take advantage of what the media programmers and engineer researchers can hook together in the iStage. In week 3 I would like to repeat revisions of week 2 etudes that we find generate insight.

## Rhythm etudes

**GAMES**

Have N ≥ 3 people wear mics close to their feet, wear noisy shoes. Each person also gets audio via radio broadcast from a station with N->N mixer.

(Noisy shoe means some shoe that will make measurably loud sound as it is pressed into the floor. It is important that the shoe make such sound deterministically and reproducibly, and its behaviour is constant over a run of an experimental. So we don’t have the motion of the footwork continuously resetting the sonic response of the shoe. )

Beam the sound of each person’s shoes to a mixer. Re-broadcast the N audio streams in different ways.

A. Make a ring (idea simplified from Navid): person K’s sound beams to person K+1, etc.

B. Blend all sounds into single audio and broadcast same mix to all people.

C. Delay the sound by variable # msec from effectively 0 to many seconds, with msec fineness.

CONTRAST WITH:

Threshold to clicks (“footfalls”), using some discriminant that can be varied simply. (e.g. threshold so we’re listening to level-crossings)

Beam click tracks to other people according to A, B, C. (yielding games A’, B’, C’)

D’. Multiply click by kinetic energy of sender’s footfall. (eg. multiply by audio energy, but in this case we may as well revert back to previous game!!)

E’. Take uniform click but accelerate or decelerate clicks by moving window average of loudness of footwork.

F’. Take uniform click but accelerate or decelerate clicks by first or second differences of loudness of footwork.

## Navid Navab etudes

## Evan Montpellier video etudes

## Michael Montanaro movement etudes (week 3 or Fall 2015)

## References from TML research blogs

temporal textures 100804

http://textures.posthaven.com/temporal-textures-100804

Adrian: temporal textures, CNMAT

http://textures.posthaven.com/adrian-temporal-textures-cnmat

[Temporal Textures] Terry Tao: flows on Riemannian manifolds

http://textures.posthaven.com/temporal-textures-terry-tao-flows-on-riemanni

TML next temporal textures + phenomenology seminar: Friday Jan 26, 16h00 - 18h00

http://textures.posthaven.com/tml-next-temporal-textures-phenomenology-semi

Adrian Freed: barycenter, bunch

http://textures.posthaven.com/adrian-freed-barycenter-bunch

For TML agenda Wed 5-6: lighting experiments

http://textures.posthaven.com/for-tml-agenda-wed-5-6-lighting-experiments

rhythm / correlation research using xOsc devices from ASU

http://textures.posthaven.com/rhythm-slash-correlation-research-using-xosc-devices-from-asu

Rhythm Studies during TML residency at Synthesis workshop Feb 15 - March 9, 2013

rhythm / correlation research using xOsc devices from ASU

http://textures.posthaven.com/rhythm-studies-during-tml-residency-at-synthesis-workshop-feb-15-march-9-2013

• Everywhere we see rhythm, think also

Temporal texture is a cover word for my extension of Michel Serres’ “clamour of the sea? (Nausicaa) from Genese.

Polytempo is let’s say a lot of different sequences of accents over common time.

http://textures.posthaven.com/temporal-textures-100804

Adrian: temporal textures, CNMAT

http://textures.posthaven.com/adrian-temporal-textures-cnmat

[Temporal Textures] Terry Tao: flows on Riemannian manifolds

http://textures.posthaven.com/temporal-textures-terry-tao-flows-on-riemanni

TML next temporal textures + phenomenology seminar: Friday Jan 26, 16h00 - 18h00

http://textures.posthaven.com/tml-next-temporal-textures-phenomenology-semi

Adrian Freed: barycenter, bunch

http://textures.posthaven.com/adrian-freed-barycenter-bunch

For TML agenda Wed 5-6: lighting experiments

http://textures.posthaven.com/for-tml-agenda-wed-5-6-lighting-experiments

rhythm / correlation research using xOsc devices from ASU

http://textures.posthaven.com/rhythm-slash-correlation-research-using-xosc-devices-from-asu

Rhythm Studies during TML residency at Synthesis workshop Feb 15 - March 9, 2013

rhythm / correlation research using xOsc devices from ASU

http://textures.posthaven.com/rhythm-studies-during-tml-residency-at-synthesis-workshop-feb-15-march-9-2013

• Everywhere we see rhythm, think also

**temporal texture**.Temporal texture is a cover word for my extension of Michel Serres’ “clamour of the sea? (Nausicaa) from Genese.

Polytempo is let’s say a lot of different sequences of accents over common time.

What I’m wondering is more along the following lines

1. Implicitly our tech and our concepts of temporal phenomena are of this form:

f: R —> V

where

R is the real line (or some interval)

V is some target space, usually (a subdomain of ) R^n.

V can be a model for say bitmaps (the case of video), or 96 bit numbers (the case of audio), or something more structured (abstract) like colour {a,r,g,b}, or all the values of all the parameters in some software application.

Implicitly the assumption is that for a given value of t, f(t) gives you a value, which possibly sits in some high dimensional space.

I have several questions in increasing depth:

Look at the structure of the inverse images : f^-1[y] = { t in R : f(t) = y }, called pre image of y.

If necessary reduce to case where y is scalar.

Imagine replacing R by R^n, (one more radical and deliberately “wrong-headed” approach to polytempo. Instead of thinking of all these different rhythms as ultimately indexible on a single abstract parameter (like SMPTE or time-of-flight corrected universal network time etc.) or the Max clock, think of the phenomena (e.g the variation of air pressure in the room, AKA sound) as a function S of MULTIPLE independent parameters (instead of one labeled “t”). More than S = S(t, x,y,z) permit S : W —> V where both W and V may be modelled by some vector space. But not necessarily. This is only a resting spot.

Watch Lebesgue–Stieltjes Integration Video

http://ocw.mit.edu/courses/mathematics/18-125-measure-and-integration-fall-2003/lecture-notes/

Measure theory and Lebesgue integration — a very compact but minimally motivated summary

Now replace f: R —> V

by

a measurable function G : M —> R

(restrict vector space V to just R for simplicity)

where M is

(For formal definitions, see textbooks on real analysis, measure theory and Lebesgue integration.)

Let the inverse image G_v := { u in M | G(u) = v }.

G is a measurable function => the inverse image G_v is a measurable subset of M.

Define the Lebesgue integral of measurable functions by: Integral [ v * µ[G_v] dv ]

The Lebesgue integral can make sense of wilder functions G than the Riemann integral can.

1. Implicitly our tech and our concepts of temporal phenomena are of this form:

f: R —> V

where

R is the real line (or some interval)

V is some target space, usually (a subdomain of ) R^n.

V can be a model for say bitmaps (the case of video), or 96 bit numbers (the case of audio), or something more structured (abstract) like colour {a,r,g,b}, or all the values of all the parameters in some software application.

Implicitly the assumption is that for a given value of t, f(t) gives you a value, which possibly sits in some high dimensional space.

I have several questions in increasing depth:

**2. Inverse images of****C****2****differentiable functions**Look at the structure of the inverse images : f^-1[y] = { t in R : f(t) = y }, called pre image of y.

If necessary reduce to case where y is scalar.

Imagine replacing R by R^n, (one more radical and deliberately “wrong-headed” approach to polytempo. Instead of thinking of all these different rhythms as ultimately indexible on a single abstract parameter (like SMPTE or time-of-flight corrected universal network time etc.) or the Max clock, think of the phenomena (e.g the variation of air pressure in the room, AKA sound) as a function S of MULTIPLE independent parameters (instead of one labeled “t”). More than S = S(t, x,y,z) permit S : W —> V where both W and V may be modelled by some vector space. But not necessarily. This is only a resting spot.

**3. Inverse images of measurable functions****http://ocw.mit.edu/courses/mathematics/18-125-measure-and-integration-fall-2003/****MATH-501 Real Analysis - I****http://video.bilkent.edu.tr/course_videos.php?courseid=12**Watch Lebesgue–Stieltjes Integration Video

http://ocw.mit.edu/courses/mathematics/18-125-measure-and-integration-fall-2003/lecture-notes/

Measure theory and Lebesgue integration — a very compact but minimally motivated summary

Now replace f: R —> V

by

a measurable function G : M —> R

(restrict vector space V to just R for simplicity)

where M is

*measure*space with measure µ, and dv is the usual Lebesgue measure on R.(For formal definitions, see textbooks on real analysis, measure theory and Lebesgue integration.)

Let the inverse image G_v := { u in M | G(u) = v }.

G is a measurable function => the inverse image G_v is a measurable subset of M.

Define the Lebesgue integral of measurable functions by: Integral [ v * µ[G_v] dv ]

The Lebesgue integral can make sense of wilder functions G than the Riemann integral can.