New Module:Instruo - scion (Biometric feedback to CV module)

[ghandipants]

Hi Scion owner/user here, thanks for the linked paper on plant conductivity, certainly one to bookmark for serious study of the subject. From your explanation it seems you understand the instruo modules operation and have described it adequately.

At first glance it may seem that it's sort of "cheating" but in practice I have found the plant does actually play a valid roll in the circuit.

I have found using an aloe vera gel on the "tens" contacts greatly increases the conductivity/sensitivity of response of the scion and is tolerated well by most plants I've tried.

So although the plant is not directly generating the voltages or gates there are definitely observable phenomena when living material is plumbed in to the circuit, which can be further interacted with.

I'd certainly recommend the scion as an experimental tool and would welcome the development other eurorack interfaces into the living world.

For the ones wondering about how Scion "extract" values from plants (or any item), here's a review paper of plant conductivity.
Also, because of a lack of information from the scientific side of this module, I would be happy to have an open discussion about it.

https://nph.onlinelibrary.wiley.com/doi ... /nph.15395

In a nutshell, and from what I understood from the manual of Scion:
- Scion will constantly send a tiny electrical signal from the Plant Input. You can see this fluctuating signal when connecting the Plant Input to a oscilloscope (OTool here).
- Connecting a plant to the TENS cables or touching the tip & sleeve of a patch cable will close the circuit of the Plant Input.
- By closing the circuit, Scion will enable the sampling of random values generated by itself.
- The user define at which rate (A) and which level (B) the values will be sampled by adjusting the Sensitivity (A) and the Density (B) parameter on the module.
- The Sensitivity parameter will act as a threshold, enabled by the difference of the measured conductivity of the Plant Input and the Sensivity parameter.

So basically, you are not sampling values from a plant (subject), but you are using the conductivity of the subject to sample Scion's internal random generator. However, because conductivity can be influenced by many parameters (cf. review above in the case of plants), you still have a dynamic interaction of the subject and the module. When connecting a plant to Scion, the plant is not playing music (i.e., generating cv) but enabling the sampling of random values. The plant must therefore be seen as a trigger generator (e.g., artist deciding when to hit a note of a keyboard or not, the fluctuating conductivity being the nervous impulse moving the finger) and not as a cv generator (e.g. a VCO itself or a noise source from which you sample values).

Did I understood it properly?

[Cheumterling]

Ciao, thanks for the feedback, much appreciated !
Indeed, at first glance the module seems a bit "meh", especially when reading people playing the esoteric card of "plant energy" and whatnot...an expensive black box...

But by tweeking the sensivity and selecting which of the 4 outs to use (because there is a gradient in sensitivity from left to right) there is some interesting sweet spots! Even better when you try to understand the biology underneath. Would be nice to run the same patch with different species/ conditions following a strict protocol.

On which type of sensors are you applying the Aloe vera gel?
My unit came with a pair of sticky pads but I also saw some users with clamps. For the later I assume the gel will work well as the conductivity will be increased by the effective contact surface. The pads already are very sticky though.

Also, any though about plumbing a quantizer in between the Scion-CVout and the VCO-CVin ? I feel there would be some room for improvement on the octave-thingy related to the attenuator. To have more coherence among the outs. Well, that's why we are into modular I guess...plumbing and making noise

[djd_oz]

I’d like to know how it generates 4 CV and 4 Gates from one signal?

[Cheumterling]

I’d like to know how it generates 4 CV and 4 Gates from one signal?

By closing the "biofeedback" (seems like the right term is "Electrical Conductivity") circuit, the module will/won't generate Gates and sample CV based on the sensivity threshold defined by the user.
Independent Gates and CV will be generated for every channel (from left to right, channel 1 to 4).
As stated by the manual, channel 1 will be the most sensitive to the EC~Sensitivity setting and channel 4 the less sensitive, with everything in between for channel 2 and 3. This could be translated as "channel 1 is the most sensitive in EC fluctuation whereas channel 4 will only react at extreme EC fluctuation.

On a side note, I would be curious to know :
- What type of voltage goes through the TENS probes ?
- From what source are sampled the 4 CVs out (e.g., white noise, Perlin noise...)

[djd_oz]

Think there maybe something more than just sensitivity, otherwise, all 4 channels will have the same value at a given point in time.

Here is the source code for a DIY Midi Sprout.

https://github.com/electricityforprogre ... 26_kit.ino

[djd_oz]

This is the circuit with the 555 timer, the rest is done via. software.

Is someone able to explain what this circuit is doing?

[djd_oz]

I think what it is doing is it is getting 2 signals from the probes and then the voltages are compared using comparators which then triggers a flip flop which outputs a pulse signal for processing using s/w. This is essentially the raw signal on the Scion.

The s/w then somehow derives the Midi information, note, note length etc ... from this signal. No idea how this is done, can probably work it out if someone is able to interpret the code.

[Cheumterling]

Thanks for the code !
Unfortunately I can't read C++ code nor electronic schemes...just the point of view from a biologist

[electricityforprogress]

Hey all, back when I made my Biodata Sonification plans open source I didn't expect so many great companies to use the tech/code and make modules like the Scion. I'm happy to answer any questions you have about the biodata circuit (555 set as astable multivibrator measuring conductivity changes as pulse widths) or the code (simple average and standard deviation on a series of readings, thousands of times per second). I also have lots of info on my website electricityforprogress.com

cheers!

[Cheumterling]

Thanks Sam !

[electricityforprogress]

A couple more 'answers' from the thread:
1. The Biodata Sonification process produces multiple notes which can be treated with 'polyphony'. I usually use 5 note polyphony, due to memory constraints on the old MIDISprout design. I think the Scion uses four notes similarly, considering costs and complexity of four CV outputs. I'm building a Biodata Modular which will have 3 note CVs and Gate outputs (shameless advert). The number of notes really comes down to number of jacks and desired cacophony.

2. The 'tens' gel pads are great for use with large leaf, fleshy plants and they can wash/peel off easily without damaging the plant under study. Tens pads are dermal stimulation/beauty products which are perfect for working with wet organic systems, easy to find at the drugstore/'zon/aliexp. I always recommend Charismatic MacroFlora (large pretty tropical plants). It is easy to use the biodata system with 'green' plants, but when it comes to Trees things are much different. I mean, lots of people think that all plants are the same, but woody large structured and hard leaves don't work as well I also have been using Silver Wire electrodes inserted into mycelium for fungus, a bunch of my buddies are making cool mushroom music on the assorted media platforms. And very popular for electrodes are copper clips (alligator clips without teeth, cause that would be rude) which can be moved between different plants and for testing 'on the go'. Surprises with blueberry bushes, but most shrubbery isn't very talkative.

3. Most users want some sort of 'note scaling' in order to take the Biodata measurements and squeeze it into something more musical. My Biodata devices can switch between 5 preset scales, Minor pentatonic by default, and of course Chromatic for that 'raw biodata'/'toddler on piano' signal. Scaling for music can take the 'pitch' approach and constrain to certain notes, and the scion also performs some 'quantization' to make the raw biodata signal a bit more musical. Now the most amazing part of Biodata Sonification for me is the Timing of the notes, that's the magic, pitches are a reference to the 'electrical conductivity' level, but the timing of the notes should accurately represent changes occurring in the plant.

4. Schematic and 'Electrocuting Plants' - The biodata circuit is a simple version of the Forrest Mims III 'Lie Detector' circuit from the RadioShack 'Getting Started in Electronics Engineers Notebook'. If you aren't familiar with Forrest I highly recommend looking at their work. We are measuring the Electrical Conductance across the surface of the plant's leaf. There are two electrodes, one electrode provides a very small current source; in the MIDIsprout 5v through a 100k resistor, in my latest Biodata Bluetooth 3v through 3.9k. An LMC555 timer is set as an Astable Multivibrator which measures changes in conductivity between two probes and outputs a varying pulse width signal (PWM). These pulses are measured in microseconds and variations in the pulse widths are used to identify changes and bioactivity. The 'trick' to this process is to isolate the power supply and ground similar to a medical device, in order to prevent loading and grounding of the signal. Electrical isolation is the reason that this device used MIDI in the first place. Most clones and commercial biodata devices, and I think the Scion also, do not have proper power isolation from the Biodata sensor.

5. Many students and biologists are using this technology for studies, and that is ridiculously exciting to me. There was even a mushroom music article in the Guardian last week!

Cheers!

[Cheumterling]

[...] There was even a mushroom music article in the Guardian last week! [...]

https://www.theguardian.com/music/2022/ ... -mushrooms

[TYGTRFB]

Will this still generate random signals even if the sensor thingies aren't connected to anything and you're not touching the leaf? Or do you need to feed it an external noise source of some kind in that scenario?

[namon]

Yup

[TYGTRFB]

Yup

To which question?

[Grum]

Will this still generate random signals even if the sensor thingies aren't connected to anything and you're not touching the leaf? Or do you need to feed it an external noise source of some kind in that scenario?

You need to feed a signal to the Sensor input - it can also be audio or CV. Connecting the sensors to a bunch of bananas also works well if you don't have a nearby houseplant

[Cheumterling]

As well as a patch cable going from your mouth to the "plant" input.
I don't have the module anymore but I remember connecting a (white) noise to engage the sampling process.
ps. For what it is worth, the CV values generated by the module have nothing in common with the subject activity (i.e., CVs are not derived from the EC). At least from what I understood. Electricityforprogress can develop in detail I guess...

[lauprellim]

For the ones wondering about how Scion "extract" values from plants (or any item), here's a review paper of plant conductivity.
Also, because of a lack of information from the scientific side of this module, I would be happy to have an open discussion about it.

https://nph.onlinelibrary.wiley.com/doi ... /nph.15395

The link is to an article about the conductance of water, not electricity.

[Cheumterling]

For the ones wondering about how Scion "extract" values from plants (or any item), here's a review paper of plant conductivity.
Also, because of a lack of information from the scientific side of this module, I would be happy to have an open discussion about it.

https://nph.onlinelibrary.wiley.com/doi ... /nph.15395

The link is to an article about the conductance of water, not electricity.

The gel-sensors used by the Scion are here to measure Electro-Conductivity (EC). Therefore water conductivity of organic subjects. To add up, EC is strongly influenced by cuticule thickness and gaz exchanges via the stomatal interface, which is measured using electricity through water conductance of the subject.
Not to cite all cell content (sodium, calcium...) that are constantly fluctuating within and among cells of a plant and therefore, influencing its EC between the electrodes.
But I'm not a plant biologist so please correct me if I'm wrong.

[Cheumterling]

Quick follow up on an email exchange Electricityforprogress (Samuel) and I (Romain) had. Could be some extra pieces of information about it. We basically exchanged emails between 19 May 2022 and the 28 November 2022. The text was only edited to make the hyperlink accessible (as you might understand, I am not a native english speaker).

Hi Samuel,

I'm a biologist and an amateur musician. I can code in R but unfortunately can't read C.

I got a Instruo Scion and I understood this eurorack module is running under the code you made for the MIDI Sprout.

Long story short, I would like to know what logic is behind the sampling of random values when the EC triggers the Sample and Hold circuit. Are the random values taken from a white noise or anything similar?
I couldn't access this information on the internet and the people from Instruo could not tell me. However, as a scientist, knowing what is happening when dealing with randomness seems to be fundamental to me.

Thanks you in advance for you time,

Hello Romain and thank you for your interest in Biodata Sonification.

Back when I published my Biodata Sonification (MIDI Sprout) design back in 2013, many companies jumped and produced clone devices of their own. When I outreached to Instruo to congratulate them on the Scion, they pushed me off quite a bit and refused to share any information regarding the code or the hardware. As the internet goes ... Insruro put out a video about the Scion and during that video a bunch of 'b-roll' shots showed a bit of the codebase and some of the circuit schematic.

While I can tell you that the Scion definitely is using a 555 timer and my basic 'change' algorithm, I sadly can't comment on any specific functions of their device.

It is amazing to me how eager companies are to clone designs and how closed they are to provide attribution or share once they have taken someone's work. Don't get me wrong, I published my code to be shared and used ... but I hoped that the companies involved would at least participate in sharing with the Biodata Community.

Instruo lists the Scion as a 'random voltage' generator, and I know that they perform a variety of scaling, timing, and other parameters in the device. It is a cool module, but I believe the Scion masks the variety and timing of the biodata signal in order to make things seem more 'musical'.

I can certainly tell you how my code and circuit works, and you can review the designs on my github (https://github.com/electricityforprogress) and info on my website electricityforprogress.com

Hi Sam,

Thanks for the quick feedback, much appreciated.

No hard feelings about Instruo. I was at the Superbooth last weekend and because they had a booth I tried to ask them directly.
The first attempt was sold by “Let me go to ask him” and “He doesn’t know”.
The second (and last) attempt was a face to face with Jason Lin but he could not answer (despite trying to sell me the module and how cool it was).
The only piece of information I got was :
“We don’t know how it is working, but it is working. That’s the important point. We added a bunch of cool function such as the quantisation, the harmonic mode and the pulse output . For the randomness we don’t know. We talked to the people at MIDI Sprout but they had no idea either, because they asked external engineers to develop the code”.

Well, it is not a verbatim but the main message is here. And for the sake of the deontology, I am not saying this to pin them down but to give you a feedback on how they are acting, sadly.
Although they are not the only one to have this predating strategy. I rode somewhere that this “open-source stealing” also pushed away Emilie from the modular scene, albeit the end of Mutable Instrument.

With the hypothesis that the people at Instruo did not change the ground of the code (i.e. the random sampling), may I ask you to explain how the generation of notes (or voltages) occurs ?

You may want to have a look at a topic we started on Modwiggler, to get an idea where I and others synthesists are in the understanding of the module (Page 2, viewtopic.php?t=188612&start=25).

Finally, if writting down the explanation would take too much of you time, I can ask a colleague from my lab that use C extensivelly. I just tought asking the person who developped the code was the best way to understand.

All the best,

Oh cool I hope Superbooth was fun!

There is a bit of explanation on my electricityforprogress.com website and I also have the old MIDI Sprout forum preserved which has many deep dives into questions, issues, and experiences.

In short, the device reads a constant stream of pulses from the Biodata Sensor (555 timer), each pulse has a different width (timing) which represents the electrical conductance across the two electrode probes. These readings happen thousands of times per second. The values are averaged, the standard deviation is taken, and a 'delta' between the min and max values is calculated. If the Delta is greater than the Threshold (assigned by the knob) multiplied by the Standard Deviation, a musical note is output. Here the 'time' when a note occurs aligns with a 'change' happening in the plant. The note and duration of the note are assigned through some math which allows the note to represent the electrical conductance 'level' and still fit into a standard musical keyboard. Since electrical conductivity has many levels, and there are only 88 notes on a piano, the notes will 'wrap around' as the electrical conductivity increases and decreases. Overall it is a super simple algorithm which shows some interesting data over long and short periods of time. Attached is an image depicting MIDI data from an Orchid over long periods of time from the study by a student.

Good morning Sam,

Superbooth was really nice indeed ! Although a bit disappointing due to some makers and their way of seeing this niche as a potential money-machine. But overall it was really inspiring !

So when you say: "The note and duration of the note are assigned through some math which allows the note to represent the electrical conductance 'level' and still fit into a standard musical keyboard. Since electrical conductivity has many levels, and there are only 88 notes on a piano, the notes will 'wrap around' as the electrical conductivity increases and decreases."

Does it means that the generated notes -representing the EC- are fitted (i.e. quantized) to a scale? Which is the default mode of your code? If so, would it be possible to output the raw values ? (after being transformed to fit a 0:+10V or -5:+5V range).

Additionally, thanks a lot for your input on the forum!
Would you agree that I copy/paste the 555 timer explanation on the topic, to better contribute to the understanding of the code?
Eventually after your answer to the above question. Although I would not publish our exchange on Instruo and the open source situation. I consider this as personal. Anyway, if you agree to it, I will send you the text for approbation.

We could as well continue the discussion there.

Thanks again and have a great day !

You are very welcome to publish any part of our dialog, and I am happy to share these questions and answers!

The modular and synth market is very challenging as there are small makers/designers who work directly with communities, and then there are larger companies who send 'representatives' and 'sales team'. The dichotomy of purpose from these two groups can be unnerving. Trade shows are a great time to show and share, but it is also a great time to spy on other organizations. There is a material benefit for sending a 'sales rep' rather than an 'engineer' to a trade show ... make sales and don't disclose secrets

Scaling - The electrical conductivity measurements are processed and assigned to a Musical Note. Consider 12 notes in a chromatic scale, and a series of octaves. The device contains multiple scales which can be selected, either Chromatic, Minor, Major, etc. Here the 'chromatic' scale is the most raw representation of the data and Minor makes a pretty lullabye. Your question about voltages, I generate MIDI notes and those notes are mapped to voltages in 1v/oct.

I am very often asked about the 'raw' signal. The raw signal is made up of tens of thousands of pulses per second each with microsecond widths. These 'real time values' are quite hard to work with, and there is too much data (too many data points happening too fast) to stream it to a serial monitor and 'look at' the raw data. This is why I take multiple readings and attempt to identify 'changes' and send notes out based on the change.

Humans and plants operate at different metabolic timescales, and there are false expectations for a mammal to be able to perceive extremely fast, extremely, slow, and extremely small changes. This is why we need a machine/algorithm to help show us activity in the plant.

It is 'possible' to take the 'raw conductivity' value and output a CV voltage BUT the true range of all electrical conductivity is quite vast while the voltage range of CV is limited, the data isn't very interesting unless changes are HUGE and occur over a long duration ... a change in conductivity for 33 milliseconds wouldn't really be knowable in a raw signal translation to CV, but with my algorithm a note can be pinged out to highlight the microfluctuation.

Hi Sam,
As a follow-up, would you agree if I get our email exchange published on the forum. I would ideally copy/paste the verbatim our all our email (with simply an edit on the hyperlink to make them accessible in the post).
Beside the data sonification material, the motivation to publish the verbatim is that I think stealing open source work is not acceptable and the modular community would benefit from pointing it out.
You can find the template that would be directly copy-pasted in the Modwiggler editor as attached file.

Hello Romain! Yes you are welcome to post our discussion and I'm happy to be quoted. One thing I would suggest is to modify the last sentence in your post, it is really important to keep a clean image especially when discussing something critical. Rather than referring to the Scion as random bullshit, I would recommend pushing the comment to be more along the lines of "It would be wonderful if Insruo can share their insights and research from designing and building the module so that the community can better understand how their product works and how it differs from other Biodata devices."

In the world of biodata and sharing the biggest risk is the criticism that a process or product is 'fake'. I find the only thing which I can do against these criticisms is accurately represent my intentions, explain the circuitry as I understand it, and describe the mechanics of the codebase. People desperately want a connection outside of themselves, in our world of constant difference and isolation, and while Plants and Biodata can help fill some of that void, there are many trolls whose only purpose is to invalidate and deride.

much love to you! Thanks for your interest, engagement, and outreach!

Last but not least, and considering that the original code has been altered, I would be very happy if Instruo could jump in to finally explain how the sampling process of "random" values works. Until then it's just an expensive module that spits random bullshit to me.
It would be wonderful if Instruo can share their insights and research from designing and building the module so that the community can better understand how their product works and how it differs from other Biodata devices.

[SpectralStimuli]

Thank you for sharing!

[lauprellim]

The gel-sensors used by the Scion are here to measure Electro-Conductivity (EC). Therefore water conductivity of organic subjects. To add up, EC is strongly influenced by cuticule thickness and gaz exchanges via the stomatal interface, which is measured using electricity through water conductance of the subject.
Not to cite all cell content (sodium, calcium...) that are constantly fluctuating within and among cells of a plant and therefore, influencing its EC between the electrodes.
But I'm not a plant biologist so please correct me if I'm wrong.

This thread has all of a sudden become much, much more interesting.

There are definitely many factors to bioelectric signals other than respiration, but it seems reasonable (from what I know) that respiration is one of many phenomena influencing conductivity across a leaf.

Thank you for your comment Cheumterling!

[Cheumterling]

To which I would add that, if one lands in front of a paywall when reaching for scientific/ peer-reviewed publications, please head to https://sci-hub.se.
You can then copy-paste the URL of the paper, or its DOI, and then download the article as a pdf (if it was uploaded on their server).
Its basically scientific piracy, illegal-ish, but fighting against money making and for free access to knowledge. A lot on the topic can be found on the internet