I've done some of this using the daisy seed. For time based effects like reverb, the memory/hardware constraints can be spicy. Definitely maxed out the seed hardware before achieving the (very long) level of reverb I wanted.
The hardware descriptions here seem on the light side. I'd want to be confident that it can handle intense time based effects.
It's promising that they seem to allow arbitrary write to the device, and only charge for tokens for the people that require the prompt playground.
Looking forward to see where this goes.
As an aside: building an ear-pleasing FDN reverb on an obscure-ish board with intense hardware optimization needs has been one of my favorite barometers for the abilities of new LLM models.
Haha, that's pretty clever. They get to sell $299 pedals, $20 plates, and upmarked "tokens" for their playground. Great example of selling shovels in a gold rush.
Well, the raw price of parts to build a simple delay circuit might be ~$3, but the hardware to build it into an actual pedal is another ~$10--nobody sells one for less then ~$20.
The real question is: do they have a real-time DSP implementing the AI FX? If not, it's worthless: if you want lag, just use any of the ten million VSTs with your computer; if it is real time, then it's basically a natural language interface on top of an FPGA/DSP -- in which case it could be useful, if it's got reliable presets and pro-grade durability.
I still think most people would take an IR-2, though
I put my laptop in my pedalboard, but after i stomped on it a couple times it stopped working.
I jest, but there is a place in the world for pedals with a small amount of latency. As i pointed out in another comment, a codable dsp pedal isn’t a new idea https://news.ycombinator.com/item?id=46727231
A cortex M7 developer board kit sells for about $40 on aliexpress. Throw in a few switches and a $10 case and you can have the same thing for ~$60 while doing almost no work yourself.
I don't know if they did anything fancy like increase the ram or storage or build a custom IC, so YMMV, of course.
My guitar teacher has a Line 6 HX Stomp multieffects pedal. In addition to programming effects patches use Line 6’s HX Edit desktop application, he also uses ChatGPT to generate patch files (they’re just JSON) by describing the effect or referencing a specific artist or song by name.
As I understand it, the difference between that and the pedal above, is that “patches” on the line6 are describing a chain of pre-existing effects. So like, phaser->delay->reverb. The polyend pedal you’re actually able to write custom dsp - so you’re able to build new units to chain together.
I do want something I can just buy and use out of the box. The hothouse seems to come close, but it'll add up to $180 once you include assembly and the daisy seed.
Hopefully if the concept catches on more there will be more options for hackable pedals on the market.
The polyend pedal being discussed is $299, plus whatever tokens you end up using. So $180 is a steal relative to that.
Also, those were just the three first ones from memory + a quick google, there are probably other options - but i doubt you’ll find anything under $150. Pedals are not cheap, except when mass produced - and a programmable dsp pedal is by definition niche.
Another option, for people willing to hack a bit, is that the FV1 dsp chip is used in a ton of pedals, and is pretty hackable if you wanted to load your own code: you could find a cheap pedal that uses that chip and hack it.
Usually, I'm not a big fan of Polyend products. They look cool, but they lack depth. Not the best tracker, not the best beatbox, etc. And also, I'm totally into Puredata and devices like Organelle, but the learning curve is steep. I get the idea of a vibe sound modeler. Not my alley, but that's interesting for a niche, I'd say.
They should have used a screen for the front plate and have the IA generate a visual. But then they'd pass the $20 per plate.
I don't like these kind of products, what I get in breadth I lose in depth, it's like having a enormous Steam library but only play the first half an hour of each game because I have limited time to invest and too many things. I'm already overwhelmed with my Katana 100.
I will say though, for a beginner, the breadth-first approach isn't bad--they don't know a good phaser from a bad phaser, but they know a phaser from a not-phaser; let them learn what a phaser is from a cheap DSP pedal, and if they like the sound, they can buy a real one.
This is more or less how i learned guitar effects, using cheap digital multi units from ~2005-2010; adding a natural language interface to that doesn't have to be bad, though I'd obviously prefer it explain what it's doing and not just presenting an un-investigable final output. Regardless, there is and always will be a market for beginner guitarists, and at the right price point, i could see this being good for them.
I agree a good multi-effect is useful to learn what the different effects actually do, but there are good entry-level multi-effect pedals that are cheaper than that. And this pedal can only have one effect at a time.
Also, it seems there's no preview in their AI playground, so you have to burn tokens and upload the effect to test it, and it may take lots of iterations to get what you want.
So I think this could mainly interest developers who are able to use it as a platform to develop their own effects without going through the AI thing, and beginners who want to be able to use different community effects to test things.
They have a public github repo with examples of code, that’d be a good place to start - but you could also check out the electrosmith daisy seed, a little audio dev board. Someone has made a pedal enclosure for it so you can diy something VERY similar to what this is (minus the vibe coding tools).
Split the audio signal in two, offset the phase of one copy by a varying amount centered around 90 degrees, and mix the two signals back together. Let me control the speed that the phase varies, and the depth of the phase change.
Something like that?
Now, knowing that is something that would sounds cool - that’s the hard part :)
From a cursory glance it appears to be a physical guitar pedal that lets you program virtual effects. The "vibe coding" aspect is likely a system directive + effects library SDK docs fed into an LLM along with the user prompt that generates the appropriate C++ which is then compiled into an effect and run on the pedal.
Note: Which is still very cool. The previous programmable guitar pedals that I've seen were all pretty low-level.
Ive been interested in doing this with a raspberry pi. Ive plugged my guitar to my pc and used FL Studio, a daw, and can add effects to it live and was curious if someone would code a os (i guess) that only ran VST (the filters) and had a screen and knobs to control things. I know its very possible, I just didnt have the time to learn how to do it.
Lol, you read my mind. I’ve been wanting a generic-looking, wood-grained “tablet display” covered with a dozen PHYSICAL faders, sliders, and knobs that you can leave permanently hooked up to a DAW that interfaces with virtual synths for over a decade now!
When you switch to a different VST, the hardware’s display would dynamically update all the text around each dial and button to match the corresponding virtual control.
Slightly related, there was a programmable guitar pedal based on the Pi Zero called the Pedal-Pi a little while back that might interest you:
> I’ve been wanting a generic-looking, wood-grained “tablet display” covered with a dozen PHYSICAL faders, sliders, and knobs that you can leave permanently hooked up to a DAW that interfaces with virtual synths for over a decade now!
Just one of several. These have existed for at least two decades, save for "dynamically update all the text around each dial", which has a variety of complications that I won't go into here.
Yeah I should have clarified - I have plenty of generic MIDI controllers. The special sauce is reflecting the "VST" rendering/presentation of its own sliders/dials onto physical ones.
This means not having to look up and down constantly between your computer monitor and the physical hardware since the knobs/dials each have small screens/displays are 1:1 matches (so Frequency Range, Sub Audio, Clamping Point, Oscillator Frequency, etc).
VSTs are rather inscrutable and I think it would be difficult to design in an agnostic way that played nicely out-of-the box with the majority of them. Doesn't stop me from lusting over the possibility though.
I am a little confused. I think you the mean reverse of the usual mapping: (a) from the surface to the plugin GUI (b) the plugin GUI is drawn to look like the surface. Right?
Interesting idea, but creates a bit of a coding conflict: the plugin developer writes the plugin GUI (typically feeling they've lavished a lot of love on it); they're not in control of the layout of a control surface (and indeed, may have no way to know what it is). So a job that would really be the job of the control surface manufacturer can't be done because that's the domain of the plugin developer.
It's fairly easy to imagine a single control surface offering this for a tiny subset of all possible plugins, but getting beyond that seems pretty much impossible to me. There was a protocol that Digidesign/AVID bought back in the mid-oughts which did maybe 60-70% of this, in the sense that it provided negotiation between the plugin and the host/surface. Problem was, it was so complex that almost no 3rd party plugin developer or control surface developer was willing to get involved.
Yeah I'm not explaining myself very well. I don't have a lot of knowledge around the inner workings of how the GUI aspect is specified on a VST but it seems to be incredibly diverse which is why while I'd love to see something like this - I just don't think it's really feasible.
It's all for the love of physical dials - that tactile ability to play with a synth is such an underrated thing.
I've got tons of VST recreations of older synths like the Minimoog Model D, Prophet 5, etc. but it's just not the same fiddling with controls using a mouse...
Hmm. I don't hate the idea by any stretch, but it low-key feels like an IR-2 gets you almost all the way there with a lot less money and effort. Not to mention i trust Boss to survive decades of touring.
Like i said i don't hate the idea. I think it's just a difficult market for this kind of idea.
I play guitar, I've built many pedals, i worked in music retail for the better part of a decade.
I think "one thing does all" pedals are hugely attractive to beginners, with good reason, and I even used to recommend such things to beginners specifically. When i was young, a digital "i can try on every effect ever made!" was an amazing value proposition--assuming the product was cheap enough for my parents to buy it for me. That was usually a Zoom 606 or DigiTech GP50 -- not anything anyone would want to gig with, but an amazing birthday present for a junior guitarist.
Anyway, the reason i never pursued building guitar gear as a job, even though i built plenty of gear i used personally, was mostly because of what i mentioned:
1. Reliability above all else. Charging professional prices for guitar gear means that shit better survive being stomped on 30x/week and a couple a three beers being spilled on in its lifetime.
2. TPB, no question. Use high quality switches, switching caps as needed. Silent, un-loaded transition between on and off is huge -- though, if you're offering delay effects, tails might be desired, the user should have an override, which would require buffered bypass instead. Just be sure to communicate which you use so people know (publishing raw input/output impedance is good; designing to work with vintage impedance-sensitive fuzz, even better)
3. I think most high end guitar players are hyper-picky, and a "Jack of all trades" unit doesn't appeal--the market for that kind of device is decidedly mid-market, and must be priced appropriately to succeed.
It's cool, though. I remember using some GNU audio real-time FX app and plugging my guitar straight into my SoundBlaster and having a great time, despite the unusable lag it all induced.
I think you could capture serious players by offering presets, IR import/copying a given input sound, of course durability, and multi-functionality: is this can replace one effect at a time? Cool I guess, but i have a pedalboard. Can this replace ALL my pedals, including routing and stacking? Vastly more appealing. If it's all software, stacking effects is probably already in the codebas3, and routing would be trivial.
I've done some of this using the daisy seed. For time based effects like reverb, the memory/hardware constraints can be spicy. Definitely maxed out the seed hardware before achieving the (very long) level of reverb I wanted.
The hardware descriptions here seem on the light side. I'd want to be confident that it can handle intense time based effects.
It's promising that they seem to allow arbitrary write to the device, and only charge for tokens for the people that require the prompt playground.
Looking forward to see where this goes.
As an aside: building an ear-pleasing FDN reverb on an obscure-ish board with intense hardware optimization needs has been one of my favorite barometers for the abilities of new LLM models.
Haha, that's pretty clever. They get to sell $299 pedals, $20 plates, and upmarked "tokens" for their playground. Great example of selling shovels in a gold rush.
Yeah, I'd imagine you could point claude code at the github and it'll do just as good or better, for a much cheaper price.
> A simple delay could cost $1.00 or $2.00, whereas a complex granular looper might cost up to $5.00.
These prices don't seem reasonable unless there's some really special sauce in their ai.
Well, the raw price of parts to build a simple delay circuit might be ~$3, but the hardware to build it into an actual pedal is another ~$10--nobody sells one for less then ~$20.
The real question is: do they have a real-time DSP implementing the AI FX? If not, it's worthless: if you want lag, just use any of the ten million VSTs with your computer; if it is real time, then it's basically a natural language interface on top of an FPGA/DSP -- in which case it could be useful, if it's got reliable presets and pro-grade durability.
I still think most people would take an IR-2, though
I put my laptop in my pedalboard, but after i stomped on it a couple times it stopped working.
I jest, but there is a place in the world for pedals with a small amount of latency. As i pointed out in another comment, a codable dsp pedal isn’t a new idea https://news.ycombinator.com/item?id=46727231
A cortex M7 developer board kit sells for about $40 on aliexpress. Throw in a few switches and a $10 case and you can have the same thing for ~$60 while doing almost no work yourself.
I don't know if they did anything fancy like increase the ram or storage or build a custom IC, so YMMV, of course.
My guitar teacher has a Line 6 HX Stomp multieffects pedal. In addition to programming effects patches use Line 6’s HX Edit desktop application, he also uses ChatGPT to generate patch files (they’re just JSON) by describing the effect or referencing a specific artist or song by name.
As I understand it, the difference between that and the pedal above, is that “patches” on the line6 are describing a chain of pre-existing effects. So like, phaser->delay->reverb. The polyend pedal you’re actually able to write custom dsp - so you’re able to build new units to chain together.
But there are other pedals that do custom code, rather than just custom patching, see my other comment https://news.ycombinator.com/item?id=46727231
This is really cool. I have a Line 6 Helix, and have wanted to explore something like this. Do you know if there is anything online about doing this?
I don’t have a specific recommendation, but a quick search found threads on Reddit and the Line 6 forums about generating patches with ChatGPT.
I hate it.
A pedal you can define with code? Kinda cool, definitely already exists, but kinda cool.
A pedal where you buy tokens to feed the ai monster to generate code to customize your pedal? Ugh. I want off this ride.
Edit - other hackable pedals:
https://www.electrosmash.com/pedalshield
https://www.op-electronics.com/en/dsp-multieffect/696-diydsp...
https://clevelandmusicco.com/hothouse-diy-digital-signal-pro...
I do want something I can just buy and use out of the box. The hothouse seems to come close, but it'll add up to $180 once you include assembly and the daisy seed.
Hopefully if the concept catches on more there will be more options for hackable pedals on the market.
The polyend pedal being discussed is $299, plus whatever tokens you end up using. So $180 is a steal relative to that.
Also, those were just the three first ones from memory + a quick google, there are probably other options - but i doubt you’ll find anything under $150. Pedals are not cheap, except when mass produced - and a programmable dsp pedal is by definition niche.
Another option, for people willing to hack a bit, is that the FV1 dsp chip is used in a ton of pedals, and is pretty hackable if you wanted to load your own code: you could find a cheap pedal that uses that chip and hack it.
I think I'm with you on this.
Usually, I'm not a big fan of Polyend products. They look cool, but they lack depth. Not the best tracker, not the best beatbox, etc. And also, I'm totally into Puredata and devices like Organelle, but the learning curve is steep. I get the idea of a vibe sound modeler. Not my alley, but that's interesting for a niche, I'd say.
They should have used a screen for the front plate and have the IA generate a visual. But then they'd pass the $20 per plate.
I don't like these kind of products, what I get in breadth I lose in depth, it's like having a enormous Steam library but only play the first half an hour of each game because I have limited time to invest and too many things. I'm already overwhelmed with my Katana 100.
I will say though, for a beginner, the breadth-first approach isn't bad--they don't know a good phaser from a bad phaser, but they know a phaser from a not-phaser; let them learn what a phaser is from a cheap DSP pedal, and if they like the sound, they can buy a real one.
This is more or less how i learned guitar effects, using cheap digital multi units from ~2005-2010; adding a natural language interface to that doesn't have to be bad, though I'd obviously prefer it explain what it's doing and not just presenting an un-investigable final output. Regardless, there is and always will be a market for beginner guitarists, and at the right price point, i could see this being good for them.
I agree a good multi-effect is useful to learn what the different effects actually do, but there are good entry-level multi-effect pedals that are cheaper than that. And this pedal can only have one effect at a time.
Also, it seems there's no preview in their AI playground, so you have to burn tokens and upload the effect to test it, and it may take lots of iterations to get what you want.
So I think this could mainly interest developers who are able to use it as a platform to develop their own effects without going through the AI thing, and beginners who want to be able to use different community effects to test things.
Very cool! Would love to know more about the audio processing backend that drives this type of thing
They have a public github repo with examples of code, that’d be a good place to start - but you could also check out the electrosmith daisy seed, a little audio dev board. Someone has made a pedal enclosure for it so you can diy something VERY similar to what this is (minus the vibe coding tools).
https://clevelandmusicco.com/hothouse-diy-digital-signal-pro...
Could you even describe a typical guitar effect if it’s not something that already has a name? How would you describe a phaser?
Split the audio signal in two, offset the phase of one copy by a varying amount centered around 90 degrees, and mix the two signals back together. Let me control the speed that the phase varies, and the depth of the phase change.
Something like that?
Now, knowing that is something that would sounds cool - that’s the hard part :)
This looks interesting. Would love to see if there are examples of pedals already vibe-coded.
I don't think that's what this is.
From a cursory glance it appears to be a physical guitar pedal that lets you program virtual effects. The "vibe coding" aspect is likely a system directive + effects library SDK docs fed into an LLM along with the user prompt that generates the appropriate C++ which is then compiled into an effect and run on the pedal.
Note: Which is still very cool. The previous programmable guitar pedals that I've seen were all pretty low-level.
Ah I see.
It is still cool though.
I'd LOVE to try it out, or see a demo in person.
My wishlist includes asking the platform to "generate a pedal that sounds like the lead guitar in Comfortably Numb" and it generates that.
I'd pay good $$$ for it.
(It might still not make a Gilmour out of me though :-( )
Ive been interested in doing this with a raspberry pi. Ive plugged my guitar to my pc and used FL Studio, a daw, and can add effects to it live and was curious if someone would code a os (i guess) that only ran VST (the filters) and had a screen and knobs to control things. I know its very possible, I just didnt have the time to learn how to do it.
Lol, you read my mind. I’ve been wanting a generic-looking, wood-grained “tablet display” covered with a dozen PHYSICAL faders, sliders, and knobs that you can leave permanently hooked up to a DAW that interfaces with virtual synths for over a decade now!
When you switch to a different VST, the hardware’s display would dynamically update all the text around each dial and button to match the corresponding virtual control.
Slightly related, there was a programmable guitar pedal based on the Pi Zero called the Pedal-Pi a little while back that might interest you:
https://www.electrosmash.com/pedal-pi
> I’ve been wanting a generic-looking, wood-grained “tablet display” covered with a dozen PHYSICAL faders, sliders, and knobs that you can leave permanently hooked up to a DAW that interfaces with virtual synths for over a decade now!
https://faderfox.de/
Just one of several. These have existed for at least two decades, save for "dynamically update all the text around each dial", which has a variety of complications that I won't go into here.
Yeah I should have clarified - I have plenty of generic MIDI controllers. The special sauce is reflecting the "VST" rendering/presentation of its own sliders/dials onto physical ones.
This means not having to look up and down constantly between your computer monitor and the physical hardware since the knobs/dials each have small screens/displays are 1:1 matches (so Frequency Range, Sub Audio, Clamping Point, Oscillator Frequency, etc).
VSTs are rather inscrutable and I think it would be difficult to design in an agnostic way that played nicely out-of-the box with the majority of them. Doesn't stop me from lusting over the possibility though.
I am a little confused. I think you the mean reverse of the usual mapping: (a) from the surface to the plugin GUI (b) the plugin GUI is drawn to look like the surface. Right?
Interesting idea, but creates a bit of a coding conflict: the plugin developer writes the plugin GUI (typically feeling they've lavished a lot of love on it); they're not in control of the layout of a control surface (and indeed, may have no way to know what it is). So a job that would really be the job of the control surface manufacturer can't be done because that's the domain of the plugin developer.
It's fairly easy to imagine a single control surface offering this for a tiny subset of all possible plugins, but getting beyond that seems pretty much impossible to me. There was a protocol that Digidesign/AVID bought back in the mid-oughts which did maybe 60-70% of this, in the sense that it provided negotiation between the plugin and the host/surface. Problem was, it was so complex that almost no 3rd party plugin developer or control surface developer was willing to get involved.
Yeah I'm not explaining myself very well. I don't have a lot of knowledge around the inner workings of how the GUI aspect is specified on a VST but it seems to be incredibly diverse which is why while I'd love to see something like this - I just don't think it's really feasible.
It's all for the love of physical dials - that tactile ability to play with a synth is such an underrated thing.
I've got tons of VST recreations of older synths like the Minimoog Model D, Prophet 5, etc. but it's just not the same fiddling with controls using a mouse...
The MP Controller is about as close as you'll get unless you go for an AIO solution like an Astrolab
https://mpmidi.com/u-he-plugins-mp-controller
https://www.arturia.com/products/hardware-synths/astrolab/as...
Hmm. I don't hate the idea by any stretch, but it low-key feels like an IR-2 gets you almost all the way there with a lot less money and effort. Not to mention i trust Boss to survive decades of touring.
Like i said i don't hate the idea. I think it's just a difficult market for this kind of idea.
I play guitar, I've built many pedals, i worked in music retail for the better part of a decade.
I think "one thing does all" pedals are hugely attractive to beginners, with good reason, and I even used to recommend such things to beginners specifically. When i was young, a digital "i can try on every effect ever made!" was an amazing value proposition--assuming the product was cheap enough for my parents to buy it for me. That was usually a Zoom 606 or DigiTech GP50 -- not anything anyone would want to gig with, but an amazing birthday present for a junior guitarist.
Anyway, the reason i never pursued building guitar gear as a job, even though i built plenty of gear i used personally, was mostly because of what i mentioned:
1. Reliability above all else. Charging professional prices for guitar gear means that shit better survive being stomped on 30x/week and a couple a three beers being spilled on in its lifetime.
2. TPB, no question. Use high quality switches, switching caps as needed. Silent, un-loaded transition between on and off is huge -- though, if you're offering delay effects, tails might be desired, the user should have an override, which would require buffered bypass instead. Just be sure to communicate which you use so people know (publishing raw input/output impedance is good; designing to work with vintage impedance-sensitive fuzz, even better)
3. I think most high end guitar players are hyper-picky, and a "Jack of all trades" unit doesn't appeal--the market for that kind of device is decidedly mid-market, and must be priced appropriately to succeed.
It's cool, though. I remember using some GNU audio real-time FX app and plugging my guitar straight into my SoundBlaster and having a great time, despite the unusable lag it all induced.
I think you could capture serious players by offering presets, IR import/copying a given input sound, of course durability, and multi-functionality: is this can replace one effect at a time? Cool I guess, but i have a pedalboard. Can this replace ALL my pedals, including routing and stacking? Vastly more appealing. If it's all software, stacking effects is probably already in the codebas3, and routing would be trivial.