Over the years we’ve explored many ways of patching brain and body to sound and music using EEGsynth. The possible sonic permutations are infinite, and then there are the artistic, musical and performative aspects. For quite a while I’ve been wanting to organize our musical patching and visuosonic performances in a way that can be explained, and even reproduced (or rather: re-enacted). To develop this niche of neuroscience and music further, I believe we should have collaboration and dissemination in mind. For one, we’ve started to organize ourselves around quasi-schematic concepts, shown in our 1+1=3 website. Secondly, I will create a brainwave music ‘Sampler’. In this sampler I will record simplified patches live, at home. The hardware (modular synthesizer) patches will be explained in full, and the EEGsynth patches will be pushed to our GitHub repository, so that anyone can reproduce and adapt the patches. Some patches will be based on ideas we’ve worked with before, while others (most) will be new and fresh. By developing them at home, at peace (and not shortly before a performance), I hope to optimize prototyping and idea-creation and fine tuning.
Melodies
We’ve not yet been working much on melodies, and the question of how to integrate melody and brain wave music is not an easy one. Neither is the modulation of temporal progressions and rhythms. I don’t mean that we haven’t created rhythms and patterns, and modulated speed and change and chance an all that. But before we typically did this in the hardware, which needs a lot of modules and fiddling around to get it right. Enter: the “Drunk Robot Module”, or “Drunk Turing Machine”: an EEGsynth module which is a mix between the famous Turing machine module, a burst module, a clock (divider), and a random/trigger delay. The novelty is the ‘drunkenness’ setting, which determined how off-beat the triggers are. This gives a a more ‘human’ feeling, and at high settings invents nice and groovy rhythms. Furthermore, the randomness is based on statistics that create a ‘natural’ feel. The EEGsynth might not be the ideal place for precise timing, but it’s okay enough, roughly at around 1% deviation in my setup. It just can’t go too quick because of MIDI. In any case, all parameters can be brain controlled so that serendipity can be further exploited in a real-time setup. I think especially in a dreamy/ambient setup it can be very nice. The random note/CV outputs can of course be quantized (using the quantizer EEGsynth module or with hardware), and in any case, the result of the module is pretty much always a musical output. I’ve show a video so you can hear what I mean, below. In the meantime, I’ve added (as in the Turing machine) a probability of changing a note within the sequence (as well as the occurrence of a burst, the number of bursts and the drunk jitter). Before I push it to our repository, I’d like to add multiple output channels (on MIDI channels) and the ability to be ‘pinged’, rather than use it’s internal clock. To conclude, this EEGsynth module will bring us a lot further in exploring interesting concepts and musicality with melody and rhythm in a way.
Harmonics
The “Harmonics” patch is at its most basic level the mapping of different (harmonic) EEG frequency bands to different harmonics added on an oscillator. For this I use the Verbos Harmonic Oscillator (I bought it pretty much for this purpose). It works very well, as harmonics always sound good, and the vibe can be very ambient, relaxing, and with a dynamic that is subtle and pleasing to the ears. Especially if you add a pinch or more of reverb. However, more should be possible to create a more interesting and developing patch. Firstly, I wanted to use a programmable switch to move the EEG control values over different harmonics (from H0 to H7), to create further subtle complexities. However, the problem, as you can hear here, is that switching creates a clipping, when some harmonics are suddenly cut off, and others suddenly enter.
To deal with this, I thought of the following, which works pretty well, but not perfectly; Instead of only e.g. power in theta, alpha and beta, I also output (MIDI to CV) the maximum value of the three (using the postprocessing module). This control voltage is then compared with a reference voltage, so that a switch will occur – and only occur – when the former is lower than the latter. In other words, when there is silence due to relatively low power in the EEG, the relationship with the harmonics is switched into another configuration. Specifically: I also output the maximum value between them with the Endorphines Shuttle Control MIDI to CV module. This max CV is compared to a reference voltage created with the Doepfer A-185-2 precision adder, using the Disting Mk4, which puts a gate high when this manual threshold is crossed. I then use the Klavis Logica XT to convert the gate into a trigger for the SSSR labs Matrixarchate, which is a nice DIY kit I recently put together. Luckily the Disting comparator function (algorithm A7) also has some hysteresis (Z knob), which has to be used to prevent flip-flopping of the gate during the crossing of the reference voltage. The reference voltage and degree of hysteresis need to be manually calibrated, and it’s not perfect, but it does create a very interesting way to use the power of the EEG to create transitions, while at the same time reducing the problem of clicks. Below is a video I shot with some simulated EEG signal. The scope shows both the reference voltage and the max of the theta/alpha/beta. In addition, at every switch I also step through a progression of 4 different notes, further creating interesting interactions between the harmonics and the base note melody. For this I use the MakeNoise Renee V1, as well as the Klavis CalTrans. The latter is a great utility module that calibrates the pitch of any oscillator. In this way, even when the 1 volt/octave tracking is not very good on an (analogue) oscillator, this module deals with that. It is indispensable when you want to play in tune with more than one oscillator at a time. In fact, let’s see if we add another oscillator to the mix next time, and add the Drunk Robot Module to the mix. The MakeNoise Renee is also capable of much more than just stepping through a regular series of notes, so we could explore it’s generative capabilities as well at some point, but let’s see how far we get with using our own software and whether I can completely remove the clipping at switching. What I’d like to try is to use some volume ducking before every switch. I have an idea where to start here, but it’ll have to wait until tomorrow. More soon!