1st cellF concert
On Sunday 4th October the 1st live performance of cellF was given. cellF is an analogue modular synthesizer controlled by live human neurons grown on an electrode array. More info HERE
This vid (shot with a cellphone) shows some of the setting upfor the show. Nathan is monitoring the CO2 levels and temperature in the incubator, btw the neurons were given a hit of dopamine before the show; it really pepped them up.
ABC Radio National interview with Guy - HERE
Pdf with good summary of the project and lots of pics - HERE
we-heart article - HERE
Regarding the electronics:
The neuron dish has 60 electrodes, each of these is hooked up to and amplifier circuit to boost the neuron signals by approx 100x and then a comparator. Neurons chatter away constantly but they put out bursts of triggers every now & then, called action potentials, and it is these triggers we want.
So the black panel contains 60 processing circuits to extract the action potentials (triggers) and then we use these to control the synth.
The panel also has 16 stimulation input to send signals into the neurons (reduced to a max of 1Vp-p and approx 7ms trigger). To prevent feedback loops there are 60 analogue switches to momentarily turn off the processing circuits when a stim signal is sent. All this took quite a bit of work to develop and get right, but it is now fkn great!
During the performance (I was not the performer, we had a real musician for that - Darren Moore), I had to check and adjust the comparators a couple of times as the neuronal activities changes over time. Otherwise the synth was untouched and just communicated with the Darren, his drumhits were fed to the stim inputs. During the performance, it was very clear the neurons were responding to stimulation. It was great to listen to and must have been an incredible experience for Darren. A pro vid & audio crew recorded the event, so I will add links when they become available.
The actual synth section contains only NLC modules:
6 tri-core VCOs
4 LFOs
4 LPGs
8 VCAs
2 VC spring reverbs
2 diff-rectifiers
2 jerkoff chaos
4 281 EGs
2 Sauce of Unces
2 7 stage clock dividers with staircase
2 quad logic with resistor ladder & VC slew
2 Quad oscillator / low pass filter
2 choppers
2 Delay no mores
2 Vactrol PiLLs
2 DP filters
2 FF chaos
2 Sloth chaos
2 offset mixers
4 voltage controlled matrix mixers (32 VCAs in each so 128 VCAs)
2 multiband distortions
12 envelope generators
8 Frigates (pitch2CV and pitch2gate modules)
Plus a bunch of mults....think thats it
This is the neural interface panel, contains 60 gain/comparator circuits to get signals from the neurons (the red LEDs indicate activity), 16 stimulation inputs and 12 envelope generators
back of the interface panel
back of the interface panel
tuning the comparators to catch the action potentials (triggers)
rockstars and geeks
Checking the neuro interface threshold levels
The PCBs are manufactured by Hackvana
this is an image of the actual neural network used for the 1st cellF gig. It has now been retired.
This vid (shot with a cellphone) shows some of the setting upfor the show. Nathan is monitoring the CO2 levels and temperature in the incubator, btw the neurons were given a hit of dopamine before the show; it really pepped them up.
ABC Radio National interview with Guy - HERE
Pdf with good summary of the project and lots of pics - HERE
we-heart article - HERE
Regarding the electronics:
The neuron dish has 60 electrodes, each of these is hooked up to and amplifier circuit to boost the neuron signals by approx 100x and then a comparator. Neurons chatter away constantly but they put out bursts of triggers every now & then, called action potentials, and it is these triggers we want.
So the black panel contains 60 processing circuits to extract the action potentials (triggers) and then we use these to control the synth.
The panel also has 16 stimulation input to send signals into the neurons (reduced to a max of 1Vp-p and approx 7ms trigger). To prevent feedback loops there are 60 analogue switches to momentarily turn off the processing circuits when a stim signal is sent. All this took quite a bit of work to develop and get right, but it is now fkn great!
During the performance (I was not the performer, we had a real musician for that - Darren Moore), I had to check and adjust the comparators a couple of times as the neuronal activities changes over time. Otherwise the synth was untouched and just communicated with the Darren, his drumhits were fed to the stim inputs. During the performance, it was very clear the neurons were responding to stimulation. It was great to listen to and must have been an incredible experience for Darren. A pro vid & audio crew recorded the event, so I will add links when they become available.
The actual synth section contains only NLC modules:
6 tri-core VCOs
4 LFOs
4 LPGs
8 VCAs
2 VC spring reverbs
2 diff-rectifiers
2 jerkoff chaos
4 281 EGs
2 Sauce of Unces
2 7 stage clock dividers with staircase
2 quad logic with resistor ladder & VC slew
2 Quad oscillator / low pass filter
2 choppers
2 Delay no mores
2 Vactrol PiLLs
2 DP filters
2 FF chaos
2 Sloth chaos
2 offset mixers
4 voltage controlled matrix mixers (32 VCAs in each so 128 VCAs)
2 multiband distortions
12 envelope generators
8 Frigates (pitch2CV and pitch2gate modules)
Plus a bunch of mults....think thats it