New sights at Old Street

Visitors travelling to the Digital Shoreditch festival via London Underground may have noticed some subtle changes in their journey through Old Street station, as a pair of tutors from the Royal College of Art have been inserting interventions along Exits 1 and 2 to Shoreditch.

Karin von Ompteda and Kevin Walker have designed a series of posters, a special lighting track, and a bespoke LED installation – all intended to evoke the station’s past, present and future – and the future of the past. “The station, as it is, was built in the 1960′s,” said von Ompteda, “with this massive roundabout and pedestrians routed underground into tunnels and tubes. It was the Space Age, and we’re celebrating that – it was like a space station, with underground shopping and this crazy giant sculpture on top of the roundabout, for cars to see, really.”

Her posters, based on photos by photographer Carl Bigmore, depict particular features and vistas at the station – ventilation shafts, ramps and towers – that might be called iconic, if anyone ever noticed them. All are covered in acres of telltale, if drab, tiles; a close look reveals subtle digital manipulations in the photos – as well as cracks, missing tiles, bits of rubbish. The grittiness is celebrated along with the 60′s futurism.

Walker’s LED installation is on an exposed patch of those tiles inside the station, echoing a row of them in size and shape but giving them life with light. “It’s about time, and space,” he said. “Time is constant, linear; and the lights animate from left to right, pushing people outward and upward. But there’s some randomness – time is full of unexpected events. Similarly for space – it builds on what has come before.” Indeed, like the posters, a close look reveals a kind of narrative of past and future, here and there. And, he says, there’s a secret code embedded in the sequence.

Travellers might miss the new track of lights that leads from the LED installation down the tunnel to Exits 1 and 2. Unless, that is, they compare this tunnel with one of the others. A clean line of special fluorescent tubes was placed, with the help of lighting designer Tom Singleton. They give a blueish, and noticeably brighter, cast to the tunnel, in contrast with the fading, yellowish lights elsewhere in the station. “We were inspired by things like Kubrick’s film 2001: A Space Odyssey,” remarked von Ompteda. “That bright shiny vision of the future is still here – Digital Shoreditch represents that in a way – but it lives alongside the grit and the grime. You have to have both. That’s why we love this station!”

Dr Walker runs the new Information Experience Design programme at the RCA, where Ms von Ompteda is a tutor and PhD student. They were commissioned by the Sustainable Society Network and the Urban Prototyping festival. Transport for London and the Greater London Authority pitched in to realise the work in just a couple of weeks. “That’s par for the course for us,” said Walker. “We teach our students to prototype ideas and concepts quickly, but as part of a research process. And that’s what this is – just the first phase of ongoing research.”

###

Oh, and it’s Version 1. Wait till you see what’s included in v.2…

Find out about the boxes here.

Catbot was very well received. Overheard:

“He’s great. Brilliant.”
“Absolutely superb!”
“Beautiful, fantastic.”
“Sweet!”
“I’ve seen worse” [acting]
“Sort of the reverse of monkeys typing Shakespeare, isn’t it?”

Emma Henry, Artistic Director of Bristol Shakespeare Festival, was particularly excited to finally meet catbot, as was the rest of the Festival team.

More pics & video here. The entire event was great – particularly Jeremy Routledge’s production ‘I had a dream that I can’t die’ was incredible – so well written and acted, it deserves multiple awards. See full programme of events on the day here.

Technical details

Catbot hardware is adapted from the Dagu ‘Playful Puppy’ robot; the back legs and tail were not used here. IR sensor eyes were considered, to respond to passers-by, but we decided best for catbot to instead recite continuously in order to attract people over. Audio came from a Jambox (an excellent little speaker – highly recommended!). Catbot was tethered via USB to a Macbook, which read random sonnet lines using a text-to-speech Applescript.

Two Processing sketches were running simultaneously – one read sonnets and sent characters out to the robot to make it move, and the other displayed text in a speech bubble on-screen. Why two? I timed everything in the first sketch using delay() statements, but these buggered the drawing of graphics. So… the first sketch reads a text file containing all the sonnets bunched together (this makes a nice rich dataset), pulls one line at random and writes it to a second text file. The Applescript reads this file, and the second Processing sketch reads this file to display the text in a speech bubble.

Here’s the Arduino sketch that lives on the ‘Magician’ Arduino-clone board on the robot:

#include  // Import the servo library

Servo servoSR;   // Make servo objects for all the servos - Shoulder Right, Elbow Right, etc
Servo servoER; 
Servo servoEL;  
Servo servoSL; 
Servo servoHead;  
Servo servoNeck;  

int pos = 0;    // variable to store servo position 

void setup() 
{ 
  servoSR.attach(2);  // Attach the servo objects
  servoER.attach(3);
  servoEL.attach(12);
  servoSL.attach(13);
  servoNeck.attach(5);
  servoHead.attach(6);
  Serial.begin(9600); // Open a serial connection with computer

  servoSR.write(160); // 160 to 0 // Starting positions, the numbers adjusted based on how they were mounted and for realistic movement.
  servoER.write(75); // 60 to 175
  servoSL.write(10); // 10 to 150
  servoEL.write(15); // 15 to 125
  servoNeck.write(80); // 0 to 180
  servoHead.write(170); // 100 to 180
} 

void loop() { 

  if (Serial.available() > 0) { // Check and see if there is a command coming from the computer
    int inByte = Serial.read(); // If so, read it into the variable called inByte.
    if (inByte == 'A') {        // If the incoming character is an A, run the function called strikeApose
      strikeApose();
    }
    if (inByte == 'B') {       // If the incoming character is a B, run the function called Rest.
      rest();
    }
  }

} 

void strikeApose() { // A function I created to pose the robot.

  int shoulderRstart = servoSR.read(); // Read the current positions of the servos into these variables.
  int elbowRstart = servoER.read();
  int shoulderLstart = servoSL.read();
  int elbowLstart = servoEL.read();
  int headStart = servoHead.read();
  int neckStart = servoNeck.read();

  int shoulderRend = random(160); // These variables hold the positions to end on.
  int elbowRend = random(115)+60;
  int shoulderLend = random(140)+10;
  int elbowLend = random(110)+15;
  int headEnd = random(80)+100;
  int neckEnd = random(180);

  // The following just move from start position to end position. The delay is to wait for the servo to catch up. Because hardware can'
  work as fast as software!
    // First move the head, then the neck, then choose a random limb.

  // HEAD
  if (headEnd < headStart) {     for(pos = headStart; pos >= headEnd; pos -= 3) {                                
      servoHead.write(pos);
      delay(15);
    } 
  }
  else {
    for(pos = headStart; pos < headEnd; pos += 3) {                                 
      servoHead.write(pos);
      delay(15);
    } 
  }

  // NECK
  if (neckEnd < neckStart) {     for(pos = neckStart; pos >= neckEnd; pos -= 3) {                                
      servoNeck.write(pos);
      delay(15);
    } 
  }
  else {
    for(pos = neckStart; pos < neckEnd; pos += 3) {                                 
      servoNeck.write(pos);
      delay(15);
    } 
  }

  int limb = random(4);

  if (limb == 1) { // RIGHT SHOULDER
    if (shoulderRend < shoulderRstart) {       for(pos = shoulderRstart; pos >= shoulderRend; pos -= 3) {                                
        servoSR.write(pos);
        delay(15);
      } 
    }
    else {
      for(pos = shoulderRstart; pos < shoulderRend; pos += 3) {                                 
        servoSR.write(pos);
        delay(15);
      } 
    }
  }

  if (limb == 2) { // RIGHT ELBOW
    if (elbowRend < elbowRstart) {       for(pos = elbowRstart; pos >= elbowRend; pos -= 3) {                                
        servoER.write(pos);
        delay(15);
      } 
    }
    else {
      for(pos = elbowRstart; pos < elbowRend; pos += 3) {                                 
        servoER.write(pos);
        delay(15);
      } 
    }
  }

  if (limb == 3) { // LEFT SHOULDER
    if (shoulderLend < shoulderLstart) {       for(pos = shoulderLstart; pos >= shoulderLend; pos -= 3) {                                
        servoSL.write(pos);
        delay(15);
      } 
    }
    else {
      for(pos = shoulderLstart; pos < shoulderLend; pos += 3) {                                 
        servoSL.write(pos);
        delay(15);
      } 
    }
  }

  if (limb == 4) { // LEFT ELBOW
    if (elbowLend < elbowLstart) {       for(pos = elbowLstart; pos >= elbowLend; pos -= 3) {                                
        servoEL.write(pos);
        delay(15);
      } 
    }
    else {
      for(pos = elbowLstart; pos < elbowLend; pos += 3) {                                 
        servoEL.write(pos);
        delay(15);
      } 
    }
  }

}

void rest() { // Function I created to make the robot go back to starting position
  servoSR.write(160); // 160 to 0
  servoER.write(75); // 60 to 175
  servoSL.write(10); // 10 to 150
  servoEL.write(15); // 15 to 125
  servoNeck.write(80); // 0 to 180
  servoHead.write(170); // 100 to 180
}

Here’s the Processing sketch that reads the sonnets and controls the robot:

import processing.serial.*; // We use serial library to talk to Arduino
Serial myPort;              // Create new serial port object
PrintWriter output;         // Create new object to write to external files

void setup() {
  String portName = Serial.list()[0];                // Port 0 is my USB port for Arduino
  myPort = new Serial(this, portName, 9600);
}

void draw() {                                        // Note - no visual component here, only robot and separate bubble.pde
  String lines[] = loadStrings("sonnets.txt");       // Read external plain text file which holds the lines of all Shakespeare's sonnets
  output = createWriter("line.txt");                 // Create new text file to hold one line
  String theline = lines[int(random(lines.length))]; // Choose random line from sonnets
  output.println(theline);                           // Write the line to text file
  output.flush(); 
  output.close(); 
  //delay(2000);
  open("/Users/kevin/Desktop/robot/speak.app");      // Applescript speaks text from the file
  delay(1000);                                       // Characters sent out via serial control robot movements, with some delays for timing.
  myPort.write('A');
  delay(1000);
  myPort.write('A');
  delay(1000);
  myPort.write('A');
  delay(2500);
  myPort.write('B');
  delay(2000);
}

Here’s the second Processing sketch that displays the speech bubble:

PImage bubble;        // image object for speech bubble graphic
PFont font;           // create new font object
int timer = millis(); // for timing display of bubble & text
String[] theline;     // variable for holding incoming line to display

void setup() {
  size(1280, 768);
  bubble = loadImage("bubble.png");
  bg = loadImage("bg.png");
  font = loadFont("Handwriting-Dakota-72.vlw"); 
  textFont(font);
  textLeading(80);
  fill(0);
  theline = loadStrings("/Users/kevin/Desktop/robot/robot2/line.txt");
}

void draw() {  
  if (loadStrings("/Users/kevin/Desktop/robot/robot2/line.txt") != null) {
    if (loadStrings("/Users/kevin/Desktop/robot/robot2/line.txt")[0].equals(theline[0])) {
      // if external text is same, do nothing and wait for new
    }
    else {
      theline = loadStrings("/Users/kevin/Desktop/robot/robot2/line.txt"); // load it up
      timer = millis(); // start the timer
    }
  }

  if (millis() - timer < 5500) {            // max display time     if (millis() - timer > 1000) {          // min display time
      image(bubble,0,0);                   // show the bubble graphic
      text(theline[0], 250, 200, 800, 600); // show the text
    }
  }
  else {
    background(0); // after max time, remove the bubble
  }
}

Finally, here’s the Applescript that reads the text:

open for access file "Macintosh HD:Users:kevin:Desktop:robot:robot2:line.txt"
set wordContents to (read file "Macintosh HD:Users:kevin:Desktop:robot:robot2:line.txt")
say wordContents
close access file "Macintosh HD:Users:kevin:Desktop:robot:robot2:line.txt"

Unsuspecting visitors thought they were on a regular tour. But the ‘tour guide’ was actually Invisible Circus actor Naomi Smyth. At one point, she receives a phone call, apparently from a father who thinks he’s talking to his daughter. All the visitors can hear the call; to accomplish this I placed a Jawbone Jambox (a small but beefy wireless speaker) in Naomi’s handbag. I synced it to my phone via Bluetooth, and at the right moment discretely played an audio file from my phone, which was broadcast from Naomi’s handbag – conveniently just below her actual phone, which was really just a prop in this instance.  You can just see the top of a ziploc plastic bag sticking out of Naomi’s bag – the Jambox is in there (it was quite showery the day of the performance).

At another stop they encounter the father/businessman himself – actor Ed Rapley – trying to make his way out of the cemetery by carelessly walking across graves. The tour guide berates him for this, and here many unsuspecting visitors joined in – one guy even tried to physically carry Ed off the grounds (Ed deemed this an acting success).

At this point mysterious voices are heard from all around. Jamboxes again, discreetly placed in the trees (also safely dry inside ziploc bags), and triggered by producer Jeremy Routledge huddled nearby under an umbrella, armed with an iPod and a tablet. I added some echo to the voices and Jeremy triggered two speakers at once, so the sound was suitably spooky. One voice – that of the businessman’s daughter – drives him into the nearby crypt.

The tour guide, and the visitors along with her, follow. Inside, they see the guy, calling for his daughter; she answers, and he comes to realise that actually he might never escape from the cemetery – he is in fact already dead. At this point, he vanishes before people’s eyes.

Now to accomplish this, we rigged up a projection screen covering the whole back part of the crypt, and rear-projected onto it this view from a live camera behind, where Ed was. I had also recorded a short clip of the space without him there, and at the appropriate moment, simply switched from the live camera to the recorded clip, triggering at the same time a ‘zzzt’ sound on yet another Jambox. (Video was done off of my Macbook in Quicktime Player, with one window directly atop the other, using keyboard shortcut to switch windows; audio played off of an iPod with my other hand.) I think it worked well; I couldn’t see visitors’ faces but there seemed to be just a stunned silence, before Naomi led them back out.

These photos by Dirk Dahmer. Thanks again to Invisible Circus for rigging up the screen – not to mention providing the excellent actors! And to Production People for renting me a short-throw projector on short notice. To Jeremy for loaning the camera gear, and for being enlisted at last minute to be a very damp stagehand. To ‘Dr Death’ John Troyer and Pervasive Media Studio for funding & organisation!

First version: I put together some modules consisting of Arduino, Wave shield (audio player), motion sensor (Maxbotix EZ1 ultrasonic rangefinder), and a nice portable but beefy speaker, the Jawbone Jambox. Everything was battery-powered or rechargeable, and encased in weatherproof housings. They were then attached to various trees in the forest at Arnos Vale cemetery, on a very showery 30 June 2012.

Second version: Battery life on the Arduino/Waveshield combo was limited to about 3 hours (with standard 9V). Plus we decided not to use motion sensors, reasoning that instead of sound being triggered by proximity (of person, animal wind-blown branch etc), playing the voices continuously (with 10 second silence between each sequence) might actually attract people over. And if they then stood still to listen, the audio would still continue to play.

So, I got these simple and inexpensive MP3 players, which I had used successfully in my research, knowing they have a very long battery life (my test clocked them at 35 hours) with a single AAA. The Jamboxes I got at least 24 hours of use out of, on a single charge. The MP3 players also have better sound quality than the Wave Shield, which is limited to 8kHz WAV files. And they’re pretty sturdy, easy to load and use.

There was a second performance of Voices on 8 Sept 2012.

[Main photo by Dirk Dahmer]