Introduce Yourself

[djowel]

This is a nascent project, and the initial feasibility assessment will be gauged on engagement, so your involvement is crucial in these early stages. Join in and introduce yourself. Share your background, your interest in pitch-to-MIDI conversion, and whether you see yourself as a potential user of this product. This will help us gauge community interest and understand how many people might benefit from this project.

I encourage you to stay informed by clicking the Watch button to receive notifications:

[olavkvern]

Hi Joel,
As you might expect, I'm very interested. If this introduction is for others--because I think you already know what I do--I use CycFi pickups in my guitar robot:

https://www.guitarworld.com/news/guitar-string-picking-robot

I'm not so interested in pitch-to-MIDI conversion, but I use MIDI for a lot of other things--mostly audio effect control. I've never found a MIDI converter with onset detection good enough to handle input from the robot. Onset detection is critical, because my main interest is in applying new envelopes in filters and VCAs. That said, if I had decent pitch detection, I'd probably find a use for that, too. I've tried dozens of approaches, some with reasonably good pitch conversion--but poor onset detection makes them non-starters.

I would definitely be a user.

With my setup--I use an Expert Sleepers ES-9 to get 16 audio inputs at 65-sample latency--it might be possible to test your work without the need for hardware (i.e., via a plug-in).

Thanks,

Olav Martin Kvern

[ArielAram]

I find this project amazing and important for many reasons.

1. Because simple, clean guitar to midi converters are becoming scarce: after the GI-10 and GI-20, Roland stopped making pure guitar to midi converters, possibly because they were not profitable enough. (The GR55 is a "bulky" converter with synth). This project will allow such a simpler device again.

2. All EEPROMs (Flash ROM), and EPROMs chips have a finite data retention time. Typically 10-15 years and after that they just start to forget their data. A device using that technology for firmware storage will just stop working when it is old enough even if all other circuits are still good. Many guitar to midi devices are becoming old ( GI-10 is from 1995, GI-20 is from 2002, etc)

3. Since most people play electric guitar, midi converters that work with nylon guitars are even more rare. The most recent guitar to midi (Fishman) is for electrical guitars, not nylon strings.

4. This project can help acquire know-how on how to add midi capabilities to other types of guitars: The Roland standard 13-pin connected fits a 6 string guitar, which means that 7,8,10 strings guitars cannot currently use midi (unless you resorts to expensive workarounds).

5. Nowadays devices have shorter and shorter life spans: if you bought an Axon AX-50 guitar to midi converter chances are it will not run in windows 11. Software and drivers become obsolete faster and faster. By having an open source project you can always update to the latest version of an OS, or hardware.

6. Maybe this project will inspire other people to try to build midi converters for violins, arabic lutes, 12 strings guitars, greek bouzoukis, harps. Things that are not economically profitable (and therefore not viable for companies) can see the light of day thanks to open source initiatives.

7. Hardware is becoming cheaper and smaller. They key investment is in the software. The Sonuus G2M ( Guitar to midi), for example, is a little device, very practical, although monophonic. But because of being comercial, people won't be able to test algorithms for polyphonic convertion. This project can enable people to have open source designs of such devices.

I am willing to support this project as much as I can. I haven't programmed in 20 years, but maybe I can test and participate i some discussions.

[ArielAram]

For those that have a midi enabled guitar, using a Roland 13-pin DIN conector, you can use this "break-out"/"fan-out" boxes to get the audio signals:

https://www.synquanon.com/13-pin-output

or

https://www.separate-strings.co.uk/breakout%20cables%20and%20more.html

[pknoot]

Just a small note: Roland 13-pin guitars are not MIDI enabled......it so happens SOME of the Roland/BOSS GK synths can indeed output MIDI notes, but the guitar itself does not. Therefore if you connect your 13-pin guitar to the SynQuaNon 13-pin Input & Output modules, you do get the audio out, but MIDI is not involved. See this informative post about the GK MIDI misconceptions:

https://www.vguitarforums.com/smf/index.php?topic=14088.0

[ArielAram]

I did not write that the guitar alone outputs midi protocol. I wrote that "you can use this "break-out"/"fan-out" boxes to get the audio signals", not midi.

I meant that guitars with 13-pin Roland connector are "enabled to be connected to a guitar to midi converter".

In any case we are missing the main point of what I was trying to say:

You may be well acquainted with the company Synquanon: in their page you read "The 13-Pin Input Breakout module receives signals from a Roland GK-3 compatible guitar and converts them to modular levels using low-noise audio op-amps."

"Converts"...that is the key point of my message.

A standard RCA line IN signal expects a voltage can reach 2 volts peak-to-peak with levels referenced to −10 dBV (300 mV) at 10 kΩ. Likewise a standard line output usually present a source impedance of 100 to 600 ohms.

I am not 100% sure ( Joel or you might bring light in this matter), but I believe that the analog signal that you get in the 13-pin connector does not conform to the standard voltages and impedances of the standard Line IN/OUT, and for that reason there are "adaptors" (aka breakout boxes, breakout modules,etc).

You need -apparently- an ELECTRICAL CONVERTION between 2 signals that are analog but with different voltages and impedances.

In other words it might not work to connect directly (with a cable) e.g. pin 5 of the 13-pin connector to a line-in input of a ADC card.

The device Joel is presenting in this project neither has a 13-pin Roland DIN connector.

That was my point: if you have a guitar with a 13-pin DIN connector you will still need one more device ( a breakout box) to connect to the device Joel is presenting.

Greetings,
Ariel//

[djowel]

Some thoughts:

• There will be 8 preamplifiers with input trim controls that allow a wide range of inputs.
• The input impedance will be high, to accommodate piezo.
• Maybe balanced? IDK. Maybe not.
• It is possible to have simple RGB LEDs on each input that indicates: Red: clipping 2: Green: signal is above -10dB 3: Yellow: Signal is below -30dB.
• The envelope is transmitted (somehow) and can be monitored by an external device/app to manage gain staging.

[loicreboursiere]

I don't know if it fits into your thoughts @djowel but if passive hexaphonic pickups like the submarine pro or the more recent PolyGuitars could also be pluged in, that would be awesome.

[djowel]

I don't know if it fits into your thoughts @djowel but if passive hexaphonic pickups like the submarine pro or the more recent PolyGuitars could also be pluged in, that would be awesome.

Absolutely!

[hobgoblina]

def interested in this project both as a user and a developer. it's been a few years, but I do have some experience working with audio/midi, c++, juce, etc.

I've got an 8-string touch guitar with nu pickups that I'd like to use with this project. I'm working toward realtime dynamic retuning & timbre shifting for playing with microtonal scales & inharmonic timbres that can be modulated during playing. i've previously taken a spectral modeling synthesis approach, but I've come around to pursuing a pitch to MIDI approach as that opens up a lot more possibilities than the SMS additive synth + noise envelope approach.

would be happy to contribute to this in any way I'm able to!

[djowel]

Welcome, @hobgoblina

[natewaddoups]

Hi everyone. I've recently picked up my guitar for the first time in about 20 years, and found CycFi while investigating the current state of guitar-midi interfacing.

I had a GR-50 back in the day, and found it barely usable. As someone who writes software for a living and for fun, I really like the idea of being able to contribute to a project like this one.

[Erol-80Hz]

Try this, you'll be amazed at how much better it is than the old GR-50 https://www.jamorigin.com/beta/ If you can pickup up a used Roland GP-10 it makes a cheap 13 pin audio interface to PC. If you no longer have a hex output guitar, it works with regular 1/4 inch output guitars too, just not as well.

[edellaq]

Fishman tripleplay user here.
Trakcing is ok, but sinche I'd use it mainly for transcribing, he latency while playing is unbearable...
I'd love to see a working pitch-to-midi-TO-STRING(or channel) device !

[hobgoblina]

yeah, I also need for each input to be mapped to its own channel for my use case

[pknoot]

@hobgoblina Indeed, the FTP software can do that.

[natewaddoups]

Share your background, your interest in pitch-to-MIDI conversion, and whether you see yourself as a potential user of this product. This will help us gauge community interest and understand how many people might benefit from this project.

GitHub discussions are a good way to gauge interest from potential collaborators on this project... But, participating in this thread requires a GitHub account, so I'm not sure it's a good way to gauge interest from users. The percentage of guitarists who have MIDI gear and have GitHub accounts might be small.

[natewaddoups]

I lean toward different tools for different purposes...

This medium is a great fit for finding collaborators since most of the potential collaborators probably already have GitHub accounts.

https://docs.google.com/forms/ would be a good way to gauge interest in products, since people can respond easily, and you can use multiple-choice questions to gauge interest in different features or usage scenarios. If you make a form you can share the link in your Facebook group, in the GitHub readme files, Discord, Instagram (if you have a page) and so on.

Discord or Facebook seem okay for brainstorming/crowdsourcing ideas. I personally prefer forums like vBulletin or Discourse (https://discourse.org/about) but I have to admit that they've tapered off and most of the discussion has migrated to Facebook. And Facebook is nowhere near as good for technical discussions and/or nuanced topics, because their threading model sucks and their seach is horrible, but don't get me started... :)

[djowel]

Awesome! Google Forms seems to be the best option indeed!

As for development related discussions, I think I prefer to keep it here. This is an open-source hardware project and I think this is the best. Once a product is available to the general would-be user, I think a targeted forum on site will be the best option, similar to that of, say, the Teensy users group.

Tell me what you think.

[natewaddoups]

Yes, I agree that this is the best place for discussion with people that are interested in participating in development.

[Erol-80Hz]

You might be wise to start a section at V-guitar forums. They have been around for YEARS currently 42,557 members with some of the best minds in the business. Anyone into midi-guitar has probably stumbled across this and joined. I'd say the current user base there are the people you should target! https://www.vguitarforums.com/smf/index.php

[djowel]

You might be wise to start a section at V-guitar forums. They have been around for YEARS currently 42,557 members with some of the best minds in the business. Anyone into midi-guitar has probably stumbled across this and joined. I'd say the current user base there are the people you should target! https://www.vguitarforums.com/smf/index.php

My only apprehension posting there is that I am not sure what the policies are for businesses. I recall in the early days that I was reprimanded in one forum (was it the 7 strings forum, IDK), when I posted, telling me that I cannot post business related matters.

[pknoot]

Peter Knoot here - very interested in how this develops. I'm a FTP user from day one and own FTP wireless, FTP Connect, and FTP Express. I'd be happy to do side by side comparisons.

[djowel]

I'm also considering providing the raw CV signals, including pitch, onset, offset and signal envelopes for each channel. I'm sure that's something that will interest you.

[pknoot]

That would be very interesting to me as well, Joel! However, you already know I'm a fan of audio-controlled analog/digital synthesizers (direct synthesis). The problem of synchronizing external oscillators through pitch CVs with the guitar frequencies is not trivial and real-time direct synthesis avoids having external oscillators trying to track along. I suspect that would not be part of this effort, right?

[djowel]

@pknoot probably not, if there's no interest. But let's see how it goes. It's a low-hanging fruit, since I already have the data.

[Orn3tte]

Hey Joel,
Maxence, professional saxophone player from France here, we talked a bit on discord already.
Playing a monophonic instrument, I'm mostly interested about the rapidity and accuracy of the pitch to midi converter (and of course envelop follower too, to map this like a CC#2 Breath Control).
My goal is to develop a musical project where I semi-improvise saxophone choir.
I'm currently using the Empress Zoia algo and those coded by the community of the Axoloti Core (currently named Ksoloti) and they're are gaving me ok-ish results but there's a lots of room for improvement.

My coding knowledge is really limited but I'd be really happy to help in anyway I'm able to

[djowel]

Welcome here, @Orn3tte

[Susorl]

Hi Joel,
I'm Suso Romarís,, processiing electric guitar signal with computer from quite a few years (mainly with MAX/MSP and Pure Data, but every DAW counts on this task...).
This project is fantastic and it will help the world of electric guitar in general. I will try to help in some way.
This sentence that you said: " I've been ambivalent about audio-to-MIDI conversion, opting instead for direct synthesis without the intervening conversion to MIDI", contains an important aspect that needs to be maintained, because real time MIDI conversion is a form of direct synthesis also, the particular class that uses the Pitch Note and Velocity like basic parameters (and the term Velocity reflects the connection with the velocity of the Hammer on the Piano). Still, MIDI2, that is an extension, or update, remains like a subclass of the broader concept of Direct Synthesis (that you mention..)
What I see here in this project is very promising, but in my opinion we need to differentiate in the 2 aspects of the device (like allways) : Hardware and Software. Both things are important, but this has a strong influence on the efforts of the global project.
That's my first message on the forum, and let's see the evolution. Thanks a lot, Joel, for the efforts and the commitment on this.

[djowel]

Thanks, @Susorl

Indeed, this is the hardware aspect. The software aspect is already under the Q DSP umbrella.

[Susorl]

Ok. The Hardware aspect contains some blocks that can introduce some differences on the global project. Like you said to Peter: "...providing the raw CV signals, including pitch, onset, offset and signal envelopes for each channel. " That sounds very promising and it contains like a more advanced concept that MIDI itself (but MIDI is like a part of that...). From the Pickup, hardware for signal conditioning and Analog-to-Digital is the first harware block here. You have done the Infinty hardware, that contains an example for this, but adc could be different depending on the target.

Because we are focused on the concept of multichannel, or monopickup per string, I think that is good idea to conceptualize first the problem here and define the model. I recommend to simplify the concept in the sense that we can use 1 String only, and create the building functional blocks around the simple case of 1String, and then the adaptation to a particular hardware solution with multiple Strings (from 1 to 8 for example...) is a matter of particular electronic definition. (You created a in Infinity a group of 3 Strings....). The important is to define a model, a protocol that includes and contains the purpose and the goals.

Like someone said: ".. Every Solution has a Problem.", and that reflects the feedback process of designing real things.

Thanks.

[Susorl]

And the same applies to the Software, when it's defined (and refined) the algorithms for 1 String, it could be used in different setups (number of strings for example), but also in different kind of Hardware also. It's good to maintain an open door to the concept of building bloks, that you already apply in Nu design...

[djowel]

And the same applies to the Software, when it's defined (and refined) the algorithms for 1 String, it could be used in different setups (number of strings for example), but also in different kind of Hardware also. It's good to maintain an open door to the concept of building bloks, that you already apply in Nu design...

This is always how I work. take a look at Q DSP, if you haven't yet: https://cycfi.github.io/q/q/v1.5-dev/index.html

[Susorl]

Yes. I know. In fact you have done over the years the complete development of what is needed here. And it's prepared and ready to be implemented. I can contribute with some conceptual ideas that I'm working on... Thanks. Let's start.

[loicreboursiere]

Hello Joel,

Loïc Reboursière here, thanks for the oppportunity. I've been reading your articles, wow that looks awesome, can't wait to test it.

For the others here, I've done some research on augmented guitar and guitar playing techniques detection 10 years ago (wow that seems like ages ago...) and finalized my Phd on hexaphonic guitar 3 years ago.
During the Phd I gave guitarists an hexaphonic multieffect and one also used the playing techniques detection tool which he used to trigger soundfiles and apply effects to them. Even if only a few playing techniques were working it was already incredible the new ways of playing it can contribute too (so is to say about hexaphonic multieffect, but I guess @pknoot would have a lot to tell on that regard).

I'm interested into jumping in the development of Hz especially to see if playing techniques detection could be integrated into that framework as well as how they could be integrated in the new MIDI standard. That would be a great opportunity to take those algorithms out. Getting to understand some of the C++ concepts of the Q library would also be of my concern :).
I know C++, JUCE, python and embbeded audio througth Teensy and Daisy and a bit with STM32.

[loicreboursiere]

I've just read back your comment UP avoce @djowel and what I'm sharing is more on the Q side I guess... Do you see any relevant point I mentioned that could be of some use for this hardware project ?

[djowel]

I've just read back your comment UP avoce @djowel and what I'm sharing is more on the Q side I guess... Do you see any relevant point I mentioned that could be of some use for this hardware project ?

@loicreboursiere thanks! I definitely am eager to work with you to bring your playing techniques detection tools to C++ and the Q DSP library.

[natewaddoups]

Random thought... I've been viewing this mostly as a way to get guitar-to-midi conversion, but if the hardware is packaged in eurorack modules, the modular synth community might enjoy it. Could be sold as single-channel pitch-to-midi modules, or stereo-pair modules.

There's already the GK-13-to-7x1/4 module, and one of those plus six single-channel Hz modules or three stereo-pair modules would constitute a guitar-to-midi system. Non-guitarists could use the pitch detection modules in whatever combination suits their needs.

[djowel]

The form factor is something worth discussing. Right off the bat, I was thinking more like a bare PCB board similar to Teensy. With this, anyone can make their own eurorack module.

[pknoot]

@natewaddoups I agree it would be one convenient form factor that serves a market larger than the guitar synth community! However we need to think it through: if this module outputs MIDI, then in most cases one would have to use MIDI-CV conversion in order to synchronize the typical Eurorack VCO modules. This is a frustrating exercise in tuning, particularly if you have 6-8 channels and temperature drift. I've tried and gave up. A much better solution is to have the module output direct synthesis waveforms and hard sync these to the VCOs. No tuning issues! The other advantage of direct synthesis is for analog subtractive single-string processing along with the guitar channels - very exciting prospect.

I see an opportunity for two types of hardware: one for generating MIDI notes into DAWs and conventional synths, the other for generating waveforms that play nice with Eurorack systems.

[natewaddoups]

I have to confess I have zero experience with modular synths, so I'll gladly defer to your expertise there. It sounds like you're saying that a eurorack module should have an analog output that sends a sine-wave at the detected frequency? Would it be useful to also have a DC output that represents volume / amplitude of the incoming signal?

The only thing that got me thinking about Eurorack is that I'm on the waiting list for one of CycFi's Thor guitars, which has pickups for each string.... So I'm going to want 6 channels of compression and 6 channels of overdrive/distortion, and Eurorack modules will take less space than traditional rack-mount gear.

[natewaddoups]

The form factor is something worth discussing. Right off the bat, I was thinking more like a bare PCB board similar to Teensy. With this, anyone can make their own eurorack module.

That sounds like the best way to start - minimum cost, maximum flexibility. There might be a larger market for complete products (with enclosure, power supply, standard I/O sockets, etc) but that would take a much bigger investment.

[djowel]

@natewaddoups I agree it would be one convenient form factor that serves a market larger than the guitar synth community! However we need to think it through: if this module outputs MIDI, then in most cases one would have to use MIDI-CV conversion in order to synchronize the typical Eurorack VCO modules. This is a frustrating exercise in tuning, particularly if you have 6-8 channels and temperature drift. I've tried and gave up. A much better solution is to have the module output direct synthesis waveforms and hard sync these to the VCOs. No tuning issues! The other advantage of direct synthesis is for analog subtractive single-string processing along with the guitar channels - very exciting prospect.

Yes, I can also synthesize raw waveforms (e.g. pulse, square, saw, sine).

[supersational]

This is such an important project, pitch detection on embedded devices/web would enable so many applications.

I'm a PhD student in music AI and would be keen to help. I'm using pitch detection for sensory augmentation, giving myself perfect pitch through haptic feedback. I also want to use this to help teach people to play guitar and other instruments by analysing their performance and linking it to sheet music. So compiling this to webassembly would be high on my priorities (easiest way give it to users is via a website).

I'm currently collecting a massive dataset for multi instrument transcription (to train e.g. MT3) which may be useful for this project? I imagine machine learning could fine tune the parameters for the detection of various instruments.

The articles you've written are really great btw.

[djowel]

Thanks, @supersational and welcome!

[supersational]

Nice, I was wondering whether for this project you're focusing first on the multi-channel guitar pickup application or on the more general case of pitch detection for other instrument types?

Also are there any examples you know of the Q DSP library running on embedded hardware like ESPs, Arduino or Teensy? it seems like a good fit for their resource constraints but I haven't seen anything myself (although it is hard to search for a library called "Q"). What hardware are you running it on usually?

[loicreboursiere]

There is one on the Daisy seed
https://github.com/recursinging/daisy_q.

Just a template project using Q pitch détection.

[djowel]

Nice, I was wondering whether for this project you're focusing first on the multi-channel guitar pickup application or on the more general case of pitch detection for other instrument types?

Also are there any examples you know of the Q DSP library running on embedded hardware like ESPs, Arduino or Teensy? it seems like a good fit for their resource constraints but I haven't seen anything myself (although it is hard to search for a library called "Q"). What hardware are you running it on usually?

To nae a few:

https://github.com/gcardi/PitchDetector
https://www.instructables.com/Making-a-Guitar-Tuner-and-Audio-Spectrum-Analyser-/
https://www.github.com/profdc9/GuitarPico

My (hardware DSP) consulting work also exclusively uses the Q DSP library.

[djowel]

What hardware are you running it on usually?

STM32 series is my preferred MCU. I design my own hardware.

[brotherdust]

This is a timely project and I'm so excited this even exists! There's not a lot of choices for MIDI guitar, and, with the exception of perhaps MIDIGuitar v3, not a lot of innovation.

I hope to contribute somehow.

My background is in network engineering, specializing in service provider and microwave communications. Corollary to this, I have also done technical documentation on a professional level for over ten years.

I also have experience in:

• Parametric CAD, such as Fusion, and Solidworks
• Programming languages such as Python, C++

Obviously, I'm not an electronics engineer, but I do like to tinker endlessly and am keenly interested in guitar MIDI. I'm wrapping up a project to modify and permanently install a Fishman Triple Play Connect to my Strandberg Boden. This involved extensive modifications to the hex pickup as well as to the bridge. During this project, I found myself just wishing the Nu pickups had a native way to do MIDI conversion without a GK unit; and that might just happen now!

I've seen in this thread that there is also interest in non-MIDI synthesis. I share this interest, even trying (failing) to learn DSP concepts and implementing in my favorite DAW, Bitwig.

So, long story short, count me in! I can help with docs, testing, design ideas, etc.

[Susorl]

For me MIDI is a particular case, the most simple one, of the Audio to Synthesis broad scenario. It's the reduction of the musical Action to only 2 Numbers ( Note Number and Velocity), in a given moment. And it's incredible to see how almost all music is based on that, it's becaus eit's the Score. I want to see this project like a the path to obtain valuable Metadata from the audio, and use MIDI 2.0 (why not?) like a possible protocol. But, the original audio, and the post-processing, is good to maintain. The information that is there is clearly more than 2 Numbers. Interestingly, most of the music of today is done with these 2 numbers, also with a few chords...!!!

[djowel]

Welcome aboard, @brotherdust

[dingodoppelt]

Interest +1
Hi, I'm Nils, professional saxophone player from Germany. I've used pitch tracking on the saxophone and other instruments before and have become more interested in the technical details over the years.
I'm not a programmer and a less than mediocre guitar player but still I came across your work twice while searching for MIDI guitar things and DSP library things. And since you asked for interest: Here it is!
cheers, nils

[djowel]

A warm welcome to you, @dingodoppelt!

[rfxDarth]

Hey Joel!

Feels like I'd be an oddball in here in that I'm neither professional player, nor researcher, nor tinkerer, nor even just good at playing guitar. I just love the instrument and all the possibilities one could do with it. Got fascinated one day with that one Burr Johnson's demo of Graphtech/AXON system, and it got me hooked into this whole midi and hex pickup rabbit hole, which subsequently led me to Cycfi blog, which I came over and over to as it's one of the few places really pushing the boundaries out there. Keep rocking on!

The main thing that particularly stands out to me is how open you guys are about your stuff, as opposed to e.g. Fishman, Roland, or Jam Origin, etc. Posting research progress, ideas, sources and schematics, it's something that piques my interest and brings me over and over again every time I'm even thinking of hunting down any gear from the aforementioned names. And I want to support that.

Unfortunately, throughout my life all this cool stuff was mostly out of reach for myself, financially speaking... But, it is now! So I'll definitely be closely monitoring this project for any updates.

[djowel]

Welcome here, @rfxDarth. And thanks for the kind words!

[Erol-80Hz]

Hi Joel and everyone,
Erol Sommer here. I've been into hex audio processing since the first Roland V-Bass came out. I've installed 13 pin systems on about 14 basses, used Roland, Ghost, RMC, CyCfi, and Ubertar pickups. I started doing some guitar about 5 years ago and now have about 5 guitars with some kind of 13 pin output. I have owned/used all the following guitar/bass synth units: Roland V-Bass, Roland VB-99, Roland GR-50, Roland GR-55, BOSS GP-10, AXON 100mk2, Antares ATG-1, Yamaha G-50, Fishman FTP, MIDI GUITAR 3 HEX. So far, the best tracking is MIDI Guitar 3 HEX, followed closely by FTP and AXON. AXON is much more self contained than FTP and easier to gig with, but GR-55 is really set up the best for live performance. That said, I want the best tracking! So currently I'm gigging on guitar with the MIDI GUITAR 3 HEX ( a bit cumbersome for live). I wish I could have a GR-55 unit that tracked as well as MIDI GUITAR 3 HEX. Ideally, what I want is a better tracking GR-55 type floor unit with 13 pin analog input, sampled sounds loadable via USB or SD card, setup with 2 expression pedals, 6 patch change button, Bank up/down, one CTRL button. Below is a pic of my current gig rig Antares ATG-1 into GP-10 for USB out to computer with MG3HEX (mini pc sits underneath the ATG-1 with mini display resting on the TC Harmonizer) Be nice to get this in one unit!

[djowel]

Welcome here, Paul. Alas, the requirement you are asking (a better tracking GR-55 type floor unit) is beyond the scope of this project. The target of this project is a bare-bones PCB board with standard headers and MIDI/USB out. It will not have any PCM sounds. It will not even have an enclosure, nor power supply. Other developers, or possibly Cycfi in the future, can work on such things, with the Audio-to-MIDI board as its heart.

[Erol-80Hz]

I understand, but good luck on the project!

[spDuchamp]

Howdy! My name is Stefan Powell and I'm a musician and experimental instrument maker, as well as a community collaborator at the Responsive Ecologies Lab (RE/Lab) at Toronto Metropolitan University.

I'm currently working on two guitar related projects in the lab, both are assistive guitars for disabled musicians.

One musician I work with currently uses a custom guitar based in the GK hex pickup that feeds the signals into a Bela board with a multichannel interface. The Bela is running Puredata code that does threshold detection and envelope tracking and sends MIDI. Chord changes are handled on the computer using a custom Max4Live patch. I'm building a clone of this guitar for him that will use Nu pickups and move Bela components that currently sit in an external blackbox, to inside of the modified guitar.

The other guitar project is a more open and customizable version, somewhat based on the first project. This one is for another musician who has the dexterity in one hand to play but doesn't have the strength in the other hand to form or hold chords. The condition they have is degenerative, so we are looking to design an assistive guitar they can learn and master, but that we can adapt and update over time to accommodate changing conditions. For this project, we are also looking to use the Nu pickups, but are aiming at making a custom controller that outputs MIDI2.0. We're also looking to design a chord changing controller that attaches to the guitar, interfaces with this new microcontroller pickup that holds the chord tables, and apply the fret transpositions (and alter envelop duration and brightness for eg bar chords) based on the selected chords.

For both of these projects we need to do threshold detection for Note On/Off, anbd envelope tracking for Aftertouch, but we don't need any kind of pitch tracking. Each string sends its defined MIDI value, with the envelope of the string vibration sent as aftertouch control. The musicians I'm working with do not have the strength or dexterity in their fingers to pitch bend, so that is handled using MIDI control in their software.

So, yeah, this project is very much of interest, and I'm happy to contribute wherever I can.

[djowel]

Welcome aboard!

By "threshold detection", I suppose you mean onset detection?

Cheers,
--Joel

[spDuchamp]

Yup. Exactly. Super low latency, separation of strings, and tracking of string vibration amplitude is exactly what we need for building assistive guitars. These are essentially expressive MIDI controllers with real strings. All the guitar sounds are synthesized. For the disabled musicians that I work with, they have to do everything through the computer, and the assistive guitar is an expressive input device for composition and performance.

Here's a look at the original version of the Kellycaster.
https://www.drakemusic.org/technology/instruments-projects/the-kellycaster/
We're working with John Kelly to develop improved versions of this assistive guitar that are robust for working disabled musicians, and that we can adapt to specific and ever changing needs. Then we aim to teach others how to build these.

We may have a lot of overlap in our projects, so we're happy to share and help out however we can.

[djowel]

Such a noble endeavor! I applaud your dedication and support for individuals with disabilities.

[Will0169]

Hi!
I'm Will, guitarist and Sunday DIYer.
I have quite a few guitars but I mostly play my DIY headless guitar (24" scale) with wireless FTP, all plugged into a Zynthian (rpi4 based).
I really like the idea of ​​the NU single string pickup, the great flexibility of the concept (I dream of a Nu Multi XR Spectra Combo 7 strings fanned for my future headless guitar)

I came across two "projects":

https://youtu.be/qh59mj0z5zc?si=2AzWNpM159WE_5oe
https://github.com/relic-se/pico-guitar-midi-tracker/tree/main

https://youtu.be/Fu0Qsz2h3HE?si=2a5iCn0vkVX-GNq4
https://github.com/ripxorip/pico_guitar_to_midi

And I thought: single string pickup to a single pico pi audio/midi to a usb hub = polyphonic audio/midi converter.

What do you think you?

I ordered enough to make 7 cheap single string pickups and 7 pico pi's to do some testing.

PS: I'm not bad at electronics but ignorant at coding.
PPS: Sorry for my broken English, I use Google Translate

[djowel]

Welcome here, @Will0169

I am not familiar with the MiGiC algorithm, but from the video, it seems to be based on waveform feature detection, which can be rather erratic, unfortunately.

And indeed, a quick search gives me this:

https://homerecording.com/bbs/threads/bituosity-migic-erratic-notes.386931/

[Will0169]

Thanks for the welcome!

Indeed, the MiGic algorithm seems a bit random, at least, in 2018 (date of the discussion in link), I read that there had been quite a few optimizations since then.

I have several small bits of information, in bulk, to put here, if it can help:

• https://github.com/jomarkow/Audio-to-MIDI open sources based on teensy
• https://github.com/geraldmwangi/GuitarMidi-LV2 a lv2 linux plugin
• https://github.com/mattmart3/pick-midi 9 years old but can be useful
• https://gitlab.com/anthonio/pythonaudioanalyserdemo no code but 2 tests on different techniques
• https://www.vguitarforums.com/smf/index.php?topic=13591.0 very interesting discussion (with videos) with a Jamorigin member, before he joined the team, on a project based on teensy 3.1

Apart from that, I dismantled my wireless FTP (2015 model), but I don't know if I can legally put the information found on this thread...

[Will0169]

I just saw some videos of the Boss Vg800/Gm800 with the new Serial GK interface, which promises imperceptible latency but also the possibility of "virtual" alternative tunings and the simulation of any type of pickup (Tone Block like software? Variax like?).
Pitch to midi + individual pitch mods....the holy grail!

[Susorl]

I really think that the concept of "virtual" alternative tunings, that is not new, but creates the transition to a different paradigm of digital processing that changes the approach of the future of Electric Guitar. Thanks for the previous links. Software and Hardware, from an open perspective, are already available for a transition from Analog to Digital. Let's see who is pushing on that direction. The Holy Grail could have different qualities, but it needs to accessible and implemented in a generic path .

[Will0169]

"but it needs to accessible and implemented in a generic path", Yes, and that is certainly the first condition.

[Susorl]

Agree. And we need to understand that MIDI (or MIDI 2.0) is necessarily a Digital technology, but the Digital paradigm transcends MIDI itself, it's more a "Processing" framework. That's why, for me, the Multichannel Digital Audio Tansmission is part of this discussion.