A selection of previous Max MSP patches can be found here. These show how the patch progressed.

Here the final Ableton Project, Max MSP project and associated samples and files can be found for our project.

As can be seen below, the Ableton patch is colour coded for conveniency and the loops are pre loaded. The bass, chord and melody tracks are grouped so that the effects chain is ONLY applied to the effects. This was done as rhythm or drums are usually the back bone to a song and effects like reverb would drown out the rhythm and thus the upbeat atmosphere of people using the installation.

The Max4Live instrument below is used on each of the tracks except the drums. Again, because these don’t need to be transposed. When a key change happens in a song, the drums won’t be key changed as these are a standard tone, they are un-melodic.

Below shows inside the Max4Live instrument. Our separate Max patch controls the number box at the top. This communication is possible through the use of the ‘udpreceive 7400’ object that recieves a signal over the network from the separate Max patch’s send. Changing the value of the number box changes the global transposition (key) of the 3 instrument tracks besides drums.

There appear to be two separate programming languages. One is simply Max MSP programming and the other is Max4Live programming, using API. These tutorials (here) helped us to understand the latter programming language.

The screenshot below shows all of the midi mapping that allows Max Msp to control Ableton Live’s features. As can be seen, there are 20 mappings for the 20 samples, a mapping for the tempo, 4 mappings for the random muter (volumes) and 1 mapping for the random effects.

So, to start, open up the max patch and follow the steps if you’re connecting sensors. Select loopMIDI from the drop down menu if you’re on Windows (You may need to download it here) or the IAC driver if you’re on Mac. We used 4 out of the 6 analog sliders shown, one of the group members used 2 other sensors for another project.

Then, if you come out of the presentation mode on the main window above you will see the path of code below. To set the port of all of the ports in the patch to the IAC Driver (or LoopMIDI) you will need to select from the drop down menu in ‘p port’.

If you go into ‘p comparison’ (by coming out of presentation mode) you will be able to find the inputs. ‘Speedlim’ and ‘change’ objects have been included but not used in the coding paths as these are optional. These objects can be used in the paths to slow down the signal. Our installed gate seems to deal with this anyway. This is currently set to 100, so that the sensors are sensitive, but not sensitive enough to trigger other sensors that are on a nearby Twister mat spot.

If you are ever having any problems with Max and Ableton communicating, it is worth checking that the Ableton Midi ports look like below.

The 2 screenshots below show inside ‘pspot1’ where most of the programming is held. All of the ctlouts should be set to the IAC Driver Bus 1 or LoopMIDI. Above the coloured panels is the other gate that stops the spots being triggered too mud, if for example, someone holds down on one of the spots. Below is the coding for each of the 4 spots at the top of the mat, red (tempo), blue (key), yellow (effects) and green (mute).

Below shows where all of the sends from ‘p comparison’ are being received. To the far left of the code is where each of the end ‘effect’ spots are, with the udpsend object sending the signal for the key randomiser, to the pre-made Max4Live instrument.

A nifty feature of the patch is that ‘p spot1’ all collapses down to the 24 buttons in presentation mode, represented as bangs. From here you can control Ableton and all of the top buttons, mute, effects, key and tempo. This is essentially the control panel.

 

 

Here is a link to the survey that we had users of the Mix Mat do after a go using the installation

Here are the results, and the slideshow below shows a pictorial representation of the results with graphs.

This slideshow requires JavaScript.

As you can see from the slideshow of results above, the responses were very positive. Unfortunately we only had 12 responses of which the majority were men and the entirety were aged between 18 and 24. Therefore, we do not have a very wide range of data due to a small and specific sample range.

As you can see from graphs and charts above, there are several findings:

• Most people heard about the Mix Mat through a friend along with some seeing it as they passed through the foyer.

• The majority thought it was better suited to younger participants.

• All participants agreed that it was best suited for having fun along with all participants rating it at least a 4 out of 5 for whether they enjoyed it.

• The majority agreed that the explanation of how the game worked was clear and that it was a professional product with the exception of one.

• Surprisingly, participants could not think of any other products that are similar to the mix mat which therefore may have caused them to state that they thought it was an original idea.

• Responses were very mixed regarding whether it gave participants a better understanding of music theory and understanding of genre.

• Responses were again, very mixed on whether or not the Mix Mat is a limited tool for creating music. The majority agreed that understanding any music theory was not necessary in order to use the Mix Mat.

• Again, the majority agreed that the Mix Mat could be used at other installation events and that the concept was understandable as well as it being easy to use and be creative.

• When asking what each participant liked about the installation the word that kept re-occurring was ‘fun’ which reinforces how enjoyable the Mix Mat can be.

• Most participants stated that there wasn’t anything they disliked about the Mix Mat in particular asides from sometimes the sensors didn’t work first time.

• A wide range of improvements were suggested which included; a larger mat, the mat being better secured to the ground, a larger choice of sound samples and getting all 24 sensors to work simultaneously.

 

All our owned equipment, ready to be set up!

It is also worth noting that during our group presentation we stated that the Mix Mat would only have one participant at a time. But on the day of the installation we decided to revert back to having two participants on the mat at the same time. This was done to keep the competitive nature of twister and therefore make the installation more fun for users.

Unfortunately, on the day of presenting the installation we had encountered more problems. Firstly, the macbook we were using for the installation had its hard drive corrupted during the final testing which meant we had to use our backup macbook. Secondly, we had been double booked for the TV space in the Eldon Foyer. This meant that we had to use another spot in the corner which we believed had a negative impact on the installation day. Being in the corner we were not within the view of as many passerby’s in Eldon as people would have had to come over to the mat in order to see that it was a twister installation. Not being in the CCi TV area meant we didn’t have access to the lights which wanted to use, which would have made the installation more aesthetically pleasing.

 

Here is an example of ‘The Mix Mat’ in action from three different points of view.

Progress was slow at first, as the foyer area was fairly quiet. However, to try and build interest, we decided to use the installation ourselves, and this generated sound gained the attention of passers by and those sitting in the foyer area, which slowly led to interest by actual participants. We decided to do some last minute advertising by contacting our peers, asking them to spare us some time and to try out the installation.

We also took part in ‘games’ against other users, which allowed us to actively explain the installation by demonstrating it. It also allowed us to take the place of a user by placing us alongside them in the use of the installation.

Once passers by saw users enjoying the use of our installation, many decided to come over and ask questions, and many decided to take part in the installation. We also shared favours with the neighbouring installation, using their installation in return for them using ours.

We received positive feedback from users, stating that they had enjoyed using our installation. They found that the fact that it was based on the popular game Twister aided the installation’s enjoyability factor, and even created a competitive atmosphere between users.

We were actually able to boost the installation’s volume quite considerably without causing disruption to the rest of the entrance area, which helped passers by notice us.

Although we had successfully placed posters around Eldon building and in the music studio we wished to create a poster to use on the day, this poster was used to attract attention to the installation while it was set up. These posters were placed around the Eldon foyer throughout the installation times to attract users passing by, they successfully achieved their purpose as we had a number of random users passing through Eldon.

This slideshow requires JavaScript.

 

Here are two video examples of the working ‘Mix Mat’. This is how it will be on the day.

As you can see, we have encountered many issues with getting al 24 sensors to work and have therefore had to settle for the just the four sensors working on the top row of parameters; the random muter, effects, key change and tempo change.

 

The two demonstrations (above) present the working ‘Mix Mat’ sensors. Each colour initiates a different parameter, as can be heard in both videos. Each of the effects are under a different coloured spot, these colours are as follows:

• The red spot initiates a change in tempo, causing the Ableton project to increase or decrease in tempo randomly. This is achieved by cycling through various set tempos between 100 and 175 beats per minute.
• The blue spot is responsible for changing the key of the piece randomly. Similar to the tempo parameter, this is achieved by cycling through set key changes, this key change is activated through Max4Live, as changing transpositions of tracks proved to be very hard through Max. We grouped the applicable tracks so that the drums wouldn’t change pitch.
• The yellow spot was used to apply a random effect to the piece, this worked by triggering certain chains via MIDI mapping. This effect fades out after 15000ms.
• Finally, the green spot was utilised to mute a track at random, this effect was used as a regular song does not have all instruments playing at all times. This effect, similar to the random effect parameter faded out after 5000ms.

As can be seen in the above videos the sensors are extremely efficient. No matter how many are pressed at the same time, they continue to work flawlessly. Although we were concerned about the sensitivity of the sensors through the Twister mat, it was clear that this was not a problem on the day, unless a participant missed the sensor altogether.

Previously we had problems with Max crashing, this was a problem that needed to be fixed before the final project as it was impossible to get 2 minutes into the installation without the program crashing. We managed to solve this by switching away from the Middle Street laptops (which run on an i5 processor), luckily we had access to a macbook pro with an i7 processor, we encountered no crashing problems from then on.

It is important to consider the reliability of the product compared to other similar products on the market. In a sense, the ‘market leader’ of musical mats could be said to be Dance Dance Revolution, often found in arcades. Our mat is nearing professional standard every time we fix a bug. With Dance Dance revolution you wouldn’t expect some of the squares not to work sometimes, and so this is why we’ve taken extra care and precautions to maintain that our installation is reliable.

• We originally thought our Arduino may have been a notorious chinese dud (as discussed in this tutorial). We looked through numerous forum posts (as seen below) online to no avail. We couldn’t find a fix for our Arduino.

  

  We then eventually got the Arduino working with that piezo we soldered. One of the lecturers simply suggested not pressing the first button on the ‘blink’ example sketch that you use on the Arduino software and just uploading it immediately. That seemed to work.

  Below is an example of the group conversation in which we are discussing the next step in regards to an error with the Arduino. We have been unable to contact Mark Sexton after many attempts so eventually Daniel took the Arduino into the studio, at this point Pere Villez found a method that successfully booted it. We are delighted that the Arduino works but disappointed by the lack of time left, in an attempt to use the Arduino we have decided to settle with using only four pressure sensors. To learn how to solder 24 piezo sensors we would need Mark’s help but we are too short on time now. We would need to know from him how to get more than 6 piezos working in MaxMSP, as only 6 appear to work. This could be an Arduino configuration issue.

  

  17:41

  Daniel Pearce
  Yknow ‘blink’ that arduino example sketch? Pere tried using that but he only pressed the upload button and didn’t press the one before it and it works..
  

  17:47
  Daniel Pearce

  Could still do this if I order another arduino and resistors with Amazon prime, breadboards and jump wires can be got from Maplin

  Assuming, like you said George, that we only need 1 lead going to the power on the arduino?
  

  17:52
  George Galway
  That is what I think… but you can’t trust my word on it!
  

  17:53
  Daniel Pearce
  Guess not

  

  18:44

  Daniel Pearce
  Anyway, we’ll stick to plan B of just triggering the loops from behind the laptop
  

  18:53

  Daniel Pearce
  Don’t even know how to get more than 6 sensors working anyway, for a start – this would all be very ambitious for the 29th April, let alone the 25th. So yea, it’s a shame but plan B will have to do
  

  18:59

  George Galway
  Alright if you think so, we need to concentrate on the blog and the paperwork
  

  19:01

  Daniel Pearce
  It’s a group decision, so tell me if you guys think we should still go ahead with Plan A
  

  19:10
  Daniel Pearce
  Plan A would require buying another arduino, a big breadboard, jump wires, resistors, soldering the rest of the piezos on, finding out how to get more than 6 sensors working through Max, finding out if we just need one wire going to the Arduino’s power
  

  19:22

  George Galway
  Yeah that is a lot of effort, here’s an alternate idea.. We wire up only the 4 randomised spots

  Therefore we include sensors and we’re not just pressing loops in time

  Well we are but there is an element of pressure sensors
  

  19:53
  Daniel Pearce
  That’s a good idea. Though will require our assumption to be correct (that you only need 1 wire for power). The arduino only has 2 GNDs

  …
  

  23:00
  Daniel Pearce
  Midday tomorrow to solder those 4 sensors then?

The video below shows George designing the poster in Photoshop.

Below are some examples of posters that were placed around Eldon building to advertise the installation event. They were created to attract attention to the ‘Mix Mat’ installation. The vibrant and recognisable colour scheme was used to draw familiarity to the Twister game mat and spark curiosity as the viewer passes the poster.

Using social media was also a fantastic way to attract attention to the installation, below is an example of a Facebook post we used to advertise our event.

The wires were cut back to the marked length by using a scalpel so that each core of wire could reach to its designated circle on the twister mat. The cores of wire were then stripped back to allow soldering to take place and for pairs of wires to branch off to their respective spots on the mat. We first soldered one piezo sensor which worked, so we then soldered 3 more.

 

In the two above pictures both ends of the wires can be seen, the end with the jump cables which go into the Arduino and the other end with the piezo sensors.

We still don’t know, however, how to solder these piezos onto the MegaShields that we have. We would solder more than 4 sensors but our Arduino doesn’t work, which makes us suspicious that it may be different because it isn’t a genuine Arduino from Arduino, the company.

 

 

In order to make the installation playable we need to secure the Twister mat to the ground. To do this we considered several types of tape, this process was important as if the mat didn’t stay in place then the piezo’s may be out of place and users could possible fall in a location that we were not expecting. Below are a couple of options we considered before settling for other options. B&Q had a wide variety of tapes available.

As you can see above, we looked into tape online with keetonsonline’s blog post ‘everything you ever wanted to know about tape and more‘, this helped us with our final decision.

 

• 

  15/04/2016 23:23
  Euan McLachlan

   

  maybe duck tape it down? thats what would general happen in a live setup? i cant think of any way to take away the bump

   

  15/04/2016 23:27
  Daniel Pearce

   

  Duct tape the mat to whatever surface we put it on? Any idea what the TV area surface is?

   

  15/04/2016 23:28

  Euan McLachlan

  its carpet isn’t it?

  like a hard floor carpet

  it would stop it sliding about
  

  15/04/2016 23:28
  Daniel Pearce

   

  I think it’s carpet yea
  

  15/04/2016 23:54
  Ben Newcombe

   

  Put holes in the foam for the sensor to come through?
  

  15/04/2016 23:58
  Daniel Pearce

   

  Just tried some parcel tape (it’s all I have) and that seems to hold it OK. So duct tape will hold it better. Wont be the most aesthetically pleasing (even if we get coloured tape) or transportable but does the job I guess
  

  15/04/2016 23:59
  Euan McLachlan

   

  fair enough

  

  16/04/2016 00:04
  Daniel Pearce

   

  Unless we think of a better option
  

  16/04/2016 00:15
  Daniel Pearce

   

  Dunno how easy it would be to get and cut foam anyway

  Eventually, we had decided to use carpet tape to secure the mat to the floor. This was much better at keeping the mat in a secure position compared with parcel tape. The tape was put onto each corner and then around the rest of the perimeter to keep it as scute as possible. The carpet tape was also used to tape down the wires to prevent tripping as part of the risk assessment. Again, we used tape to secure the sensors to the underside of the mat to ensure they remained in the centre of each spot.

   

Some of the university’s software either didn’t appear to have a licence any more or required an admin password to access. This temporarily halted any programming progress.

A post, here, told of how another problem, that we couldn’t open Max4Live devices could be solved.

Middle street only had demo version, which means had to log in with our purchased version of Max msp to authorise it, so then we could save.

Moving on, we made a final coding adjustment which would be crucial to users playing on The Mix Mat.

The small section of code above has helped the patch thoroughly throughout the creation process. It was created because pressure sensors were triggering many times per second due to the gating process. This would cause effects such as the muter to mute all the tracks at once, which of course would turn the sound off for a significant amount of time. The code functions by sending a ‘1’ into a number box, the ‘change’ object below recognises that the number has changed from the default, ‘0’, to ‘1’ and sends out a bang, this bang will only be triggered once, this means that the effects will not be pressed more than once per time the sensor is pressed. Each press of the sensor then triggers a delay, meaning that the delay is triggered every time a bang is sent through, once the delay is completed it will reset the number box with the ‘0’ message, that means the code is ready for the sensor to be pressed again.

The video below shows how we achieved power to multiple sensors using a breadboard. The power supply is sent into the breadboard and then carried to each sensor by a wire, the next sensor will then take the power supply from the previous wire, and so on.