Randrum – a weighted random drum sequencer

Randrum is a drum sequencer that uses weighted random values to trigger audio samples. The idea came from programming drum patterns using Pbind in SuperCollider (SC). In SC there is a pattern object called Pwrand (SC description: “Returns one item from the list at random for each repeat, the probability for each item is determined by a list of weights which should sum to 1.0.”) which I really liked and wanted to see how that would work in Max. Randrum is the result of this experimentation.

It is still very much a work in progress, so I will share my code in a later stage I guess.



Control stuff with M4L – update [26-05-2015]

I have made some stuff in MaxForLive to integrate SuperCollider and the Roland JP8000 synthesizer more into my workflow in Ableton Live.

– SuperCollider SynthDef control: Send values from Ableton Live to a SynthDef in SuperCollider. Map MaxForLive device dials to SynthDef arguments using the openObject Quark. M4L dials can be automated in Live, so you can automate your SynthDef parameters in your composition.


The SuperCollider_SynthDefCtrl M4L device has at its top left corner a text field where the  publish name of the SynthDef should be entered. To map a dial to a SynthDef argument simply change param_name into the SynthDef’s arg name. Each dial has its own lower and upper boundaries (default set to 0. to 1.0) and a lag value.

// to install OpenObject Quark:
// Find and check the OpenObject Quark and hit Apply

// to start OpenObject Quark in SuperCollider:

// I always use a gate arg, it's not mandatory though.
// You can change the gate argument in the M4L device
// to something else.

SynthDef(\synthie, {| gate=0, arg1, arg2, arg3... |

// some awesome SynthDef code


a = Synth(\synthie); // create a synth
a.publish(\synthie_or_something_else); // publish name


– Automate Roland JP8000 synth parameters in Ableton Live (not updated – 10-2014): Download





Max object: mxj pandorabot

Pandorabot mxj object for Max, to chat with pandora bots.

Pandorabot java source and class file on github.

Here is a Max patch where two bots talk to each other: Pandorabot.zip This max patch also uses the external [flite~] (to do text-to-speech, so it sounds like the two bots are having a conversation out loud), and my [mxj htmlsource]. Mxj htmlsource is used to get all the ID’s of the available bots you can chat with. Not all bots work correctly atm, because of the way they use/named html forms.

Here is a short video of one of their conversations: [vimeo https://vimeo.com/97649674]


Max Twitter Stream [update 26-09-2013]


I have just made my first two mxj objects for Max, using Twitter4j, a java library for the Twitter API. You can filter the live Twitter stream by using keywords, users or locations. I made this for a sonification project I am working on at the moment. Next to the live stream you can also search twitter data, view timelines, direct message friends and some other cool stuff. This is my first time using Java (next to some stuff made in Processing), most of the code is a slightly modified version of the Twitter4j examples. The mxj objects where made in Eclipse, you can find a short tutorial on how to set up Eclipse to start developing mxj objects here, and check the WritingMaxExternalsInJava.pdf located in the Max applications folder under java-doc. The idea of combining Twitter stream location data with a world map came from this project. If you want to use this code you have to get a Twitter Consumer key/secret and access token/secret, and put it in your java code at the ConfigurationBuilder stuff, highlighted in the code below. Maybe i will implement this into the code in the future to streamline this process, but for now you can find some info on how to get these keys on the Twitter developer forum.

Here is a short video:

[update 26-09-2013] :

  • Added Sentiment Analysis, with LingPipe, which is a toolkit for processing text using computational linguistics. Tweets are analyzed and categorized, the dot on the world map, representing the location of the Tweet, wil have a corresponding color (positive=blue/neutral=green/negative=red). Amount of pos/neu/neg Tweets are counted (so you can see the overall sentiment on topics or area’s). (Only works on Tweets written in English)
  • Added a daylight layer, which is a world map showing day/night around the globe and cloud coverage. You can change the opacity of the normal (grey scale) map, and the opacity of the daylight map.
  • Added search location by name rather than bounding box coordinates. Location names are converted to bounding box coordinates using the website: http://isithackday.com/geoplanet-explorer/ For this to work I created an extra MXJ object that imports a websites source into Max.

You can download the Max MXJ objects here: https://github.com/Haighton/mxj-objects

And the Max patches here.