Embedded - 471: Bicycle Built For Two

Episode Date: February 22, 2024

Where electronics meets music, there is a board called Daisy. Created by ElectroSmith, Andrew Ikenberry, the goal of the board is to teach computers to sing. Andrew joined us to talk about music, audio processing, instruments, product design, and electronic manufacturing.  See the Electrosmith website, specifically the Daisy Seed. The electro-smith github repository is extensive (with many Daisy Examples). Also see their YouTube channel. Electrosmith is offering 5% off until mid-March for folks with the coupon code mentioned in the show. We mentioned a number of synths but the CHOMPI is particularly nifty. Daisy Bell - Wikipedia (and where you might have heard that before (and if that doesn’t give “teach computers to sing” a creepy vibe, I don’t know what will)). Transcript Nordic Semiconductor empowers wireless innovation, by providing hardware, software, tools and services that allow developers to create the IoT products of tomorrow. Learn more about Nordic Semiconductor at nordicsemi.com, check out the DevAcademy at academy.nordicsemi.com and interact with the Nordic Devzone community at devzone.nordicsemi.com.

Welcome to Embedded. I am Alicia White alongside Christopher White. Our guest this week is Andrew Eikenberry. And let's talk about teaching computers to sing. Hi, Andrew. Thanks for joining us. Thanks for having me. Pleasure to be here. Could you tell us about yourself as if we met at, oh, I don't know, NAM booths? NAM. I'm very familiar with that. Definitely. My name is Andrew Eikenberry, and I'm an electronic

musical instrument designer. And I've also co-founded a couple of different music tech companies. I started my career when I was a senior at the Berklee College of Music in Boston. And while I was there, I co-founded a company called Qubit, which makes modular synthesizers. Over the years, Qubit's released somewhere over 100 products, and they've gone on to be used by a lot of incredible artists, sound designers, including people like Deadmau5, Martin Gore, Trent Reznor, and many, many others. And then a few years back, I kickstarted a new company called Electrosmith, which makes the Daisy platform, which is what brings me here today. All right. And we are going to ask a lot about the Daisy platform and some about some of the others. Well, I want to ask about dead mouse.

It's not mouse rat. Mouse rat? No, that's something different. Nevermind. Let's go on to the next one. Go ask about dead rat. I am sorry. Dead rat? No, it's fine. I was just going to ask how he drinks through the helmet, but that's fine. You know, I'm not privy to that information. Yeah, that's fine. I was just going to ask how he drinks through the helmet, but that's fine. You know, I'm not privy to that information. Yeah, that's fine.

Okay, before I can screw up that band name again, let's go with lightning round, where we ask you short questions and we want short answers. And if we're behaving ourselves, we won't ask how and why and are you sure and what about. Are you ready? Ready. Let's do it. What's your favorite instrument? Modular synthesizer. That's kind of a gift. That's too easy. Come on.

Favorite input to a synthesizer like keys, knobs, button sensors? It's going to have to be cutoff frequency. And I know that's a little vanilla, but got to do it. Cutoff frequency to a filter. Best booth you visited at NAMM besides your own or best stage thing you saw there, have seen there? Great question. Let me see. How could you choose just one? Right.

Exactly. Exactly. That's tough. But one does stand out. Novation had this booth one year where the whole floor, it looked like one of their launch pads. And all these different grids would light up and you're supposed to dance on it. And it was just an absolute blast. So I spent way too much time hanging out at their booth just dancing on this launch pad. Actually, could you describe what NAMM is?

Because we've used it a couple times, but I'm not sure our audience knows. Yeah, so NAMM is essentially definitely the country, if not the world's largest music trade show. And it specializes on the musical instrument and pro audio space. So you're going to see companies like,

well, you used to see companies like Fender, Gibson, Roland, Korg, etc. But it's obviously had a hard time since the pandemic. So your mileage may vary as far as which companies you see these years. But it is coming back. I was just there a couple weeks ago. And when you go there, it's not just the industry

manufacturers and stuff. A lot of famous musicians are just wandering around there. So it's kind of a weird experience. It is very interesting. Yeah. So, you know, ostensibly, it's an industry-only trade show. So you have to get a ticket. You can't purchase a ticket.

You have to get it through, you know, in theory, you're getting it through your job. Your work is going to get tickets. And then you're going to go negotiate. You're going to figure out what you want to order for the next year, see what upcoming products these brands have. But what ends up happening is the real business takes place on Thursday and Friday. And then Saturday and Sunday, everybody within driving distance gets a ticket from their buddy who has a guitar store. And then they drive down and walk around for the day and check out all the

cool stuff. Okay, back to lightning round. Sorry about that. That's not very short. Best Southern California beach? Ooh, I'm going to have to go with the one I live nearest to, which is San Comani. So let's go with T Street. Fish tacos, burritos, or tamales? Fish tacos. Musician, engineer, or CEO?

That is a tough one. You know, I'm going to go with my gut, musician. Complete one, engineer, or CEO? That is a tough one. You know, I'm going to go with my gut. Musician. Complete one project or start a dozen? Always one. Always one project. Favorite fictional robot? Bender from Futurama.

And do you have a tip everyone should know? Back up your files. Oh, a tip born of frustration and experience. It's like the tip where you shouldn't catch a falling soldering iron. Back up your files in two places. One of them not in your house. One is none, right? One is none.

Never forget that. I'd like to thank our sponsor this week. Nordic Semiconductor specializes in ultra low power wireless communication. They have a wide technology portfolio, including Bluetooth low energy, low energy audio, Bluetooth mesh, thread, matter, cellular IoT, and Wi-Fi. They have thousands of customers worldwide with 40% market share in Bluetooth Low Energy. They have nearly 2 million system-on-a-chips produced every day. Sounds like a system that you can't go wrong with. So please check out nordicsemi.com. And if you would like to win one of the Power Profiling Kit 2s that our sponsor is so generously handing out, please send us an email that is show at embedded.fm

and tell us what your favorite PPK2 feature or spec is. Again, thank you to Nordic for sponsoring this week's show. Okay, so we asked you to talk about Daisy after one of my students used it in a project for my class and they were building, well, they were building a guitar pedal, but there were a couple other students who referenced it. So I get curious. And then I saw your tagline for a company is, let's teach computers how to sing. And tell me all about it. Yeah. So what ElectraSmith is all about is merging the worlds of technology and art. And I think sometimes in the tech space or definitely like the maker world,

we get a little too fixated on the process or the specific technology that we're using and sometimes kind of lose sight of the end result. And for us, you know, for us at Electrosmith, personally, we're really not just programming computers. We're doing what we think is, you know, one of the most awesome human behaviors known to man, which is making music, you know, or singing, right, metaphorically speaking. So it kind of just sums up really our mission statement, which is, obviously, let's teach computers how to sing. I almost said that, but it's really just merging these disparate worlds of technology and art. One of the interesting things about the Arduino platform was that it made technology accessible to folks who weren't computer scientists, who weren't engineers. Do you have the same philosophy or do you expect some level of technological understanding

from your audience? No, I would say we have a very similar philosophy. And you know, that stems a lot from my own personal experience, which is I grew up as a musician. I did not do particularly well in math or science classes. And you know, it's not because I wasn't studious. And it wasn't because I wasn't used to dedicating large amounts of time to learning something, it's not because I wasn't studious and it wasn't because I wasn't used to dedicating large amounts of time to learning something. It's just that I really didn't get excited about it. And I didn't see an application for it that made sense for me, you know. But meanwhile, I'm sitting in my guitar playing or sitting in my room playing guitar for, you know, eight hours a day and things like that, which I think a lot of musicians are used to.

So, you know, I think that there's, unfortunately, this disconnect between musicians and certain subjects like math and engineering, because they don't see an application. And one of our big, big, you know, driving forces is let's fix that. Let's show these musicians why it's important to learn these things. And not only that, let's enable them to create new instruments themselves. Because that's where when you really look at a lot of the innovation that's happened, it's always people who think outside the box. Leo Fender wasn't necessarily just thinking about a cool product. He was passionate about music. And that led him to build and create and design what he did. You know, and how do you how do you align? You need to understand signal processing, which is for you and how things work in modulation with not having a science background. Right. And, you know, that is a great question.

And largely what we like to say is that you don't need any to get started. You don't need to know any signal processing to get started. We've taken care of the nitty gritty, so to speak. And a big part of that is we have an open source DSP library, which is extremely accessible and anybody can use it. That way, they don't have to necessarily code up their own oscillator. They can just call our function, which does really all the heavy lifting for them. Now, on the flip side, I get this question a lot, which is, well, I'm a really experienced engineer, so am I going to be limited? And the answer is, of course not. I mean, the sky's the limit.

We're not going to hold you back. We just made it easier to get started. So the short answer to that question is you don't really need any knowledge of DSP to start. But as you learn and as you grow, it's going to come naturally. And then you don't have to worry about creating a baseline. You can just focus on the fun stuff, which is making sound and tweaking the algorithms. Do you need a warning on your product?

This may cause you to become interested in the Fourier space and mathematical concepts. We should. We should put something like that on there. As is our brand, we've gotten a bit ahead of ourselves. What? Yeah. So what is this product?

Oh, yeah. Yeah. There's this Daisy thing I heard is a board. Could you tell us about it, the specs and whatnot? Most definitely. Most definitely. So Daisy is an embedded platform for music or any audio device, really. It's not music specific. That just tends to be our bread and butter. And what the platform consists of is development boards, which are similar to a Teensy, Raspberry Pi, things like that. It has an ARM processor on it, an audio codec. That's one of the things that sets it apart, as well as a few other peripherals that make doing complex audio easy. One of those things being we have an external SD RAM chip on there, which gives you 64 megabytes of RAM.

So you can do loopers, you know, things with large buffers are inherently easy on the Daisy platform. And then the second part to what the platform is, is our software ecosystem. And what we like to think of ourselves as doing is just making it easy so that you don't really have to spend a lot of time designing hardware. You also don't have to spend a lot of time installing tools or creating a development environment. You can get up and running very quickly using our off-the-shelf solutions. And I saw that it can be programmed with the Arduino interface. And there was some sort of web thing that looks like it used an embed, like just drop a file in a folder sort of program on board? Yeah, exactly. So we support a wide variety of programming environments, all the way from max MSP or pure data on one side down to like low level C++ on the other. And of course, somewhere in the middle, there's the Arduino IDE, which we have full support for. And then the webpage you mentioned, the DAISY Programmer, is what we call it.

You can drop in a binary file and then flash any DAISY hardware. And now where this becomes really, really powerful is that people can share firmware really easily. And this isn't just users necessarily. A lot of companies take advantage of this because they can just put daisies in their products and then all of their users have an easy and effective way to flash firmware that doesn't involve getting a custom programmer or any weird cables or anything like that. Yeah, you just plug in USB on your computer, it shows up as a drive. You plop in your file and it programs itself. Exactly, yeah.

As an embedded developer, supporting multiple ways of programming something sounds like a nightmare. It's horrible. It is absolutely horrible. I'll tell you, and this goes back to one of your lightning round questions, which is really good, is we've learned a lot of lessons about supporting too many things. And that's just the honest truth. Sometimes if you support too many things, you decrease the user experience across the board. So it is something we're working on moving forward is kind of narrowing down specifically what most users are using and focusing more development time on that specific language rather than trying to get even coverage

across the board. That makes a lot of sense because there's this layer of impenetrableness when you have too many options. It's like, just tell me what to do and I'll do it and then I can make my choices. Right. Exactly. And that's one of those things where it's a better product if we tell you this is where you're going to have the best experience and that's what we give you. Not pick your poison. Pick one of these six different languages and your mileage may vary as far as how well it works out for you. So yeah, we're definitely having these conversations daily as we move forward with the support for the platform. If I want to program it with a JTAG style programmer, SWD probably, can I do that or do I need to go through the other way? You definitely can. And it's something we highly

recommend. You know, we get a lot of first time programmers coming on board using it and they're, you know, using printf to debug, quote unquote, you know. And so we always like to tell them, hey, there's a better way to do this. You just have to buy this little JTAG programmer debugger hardware, which we get directly from ST. And it's called the ST-Link Mini V3 something or other. It's on our website. And that's what we recommend using. And then we have kind of built-in support to the VS Code IDE

for getting it running when you're writing in debugging code. Okay. What's in the name Daisy? I thought you would have called it Moog Lite, except that was probably trademarked or something. That is a great question. And I'm surprised when people ask me this, people that know so much about the history of electronic music and this and that, and they're

still kind of like, why Daisy? It's kind of weird. So Daisy comes from this song called Daisy Bell, and in parentheses, bicycle built for two. It's this old song from the late 1800s, I believe. And it was actually the first song that was ever played by a computer. It was programmed at Bell Labs by Max Matthews, sometimes in the 50s, I believe. And it's this cute little song, Daisy Bell. You can look it up online. But we really wanted to kind of highlight the heritage of what we're trying to do here, of making computers sing and where it all started and hopefully take it to a whole new world.

Because obviously, Max Matthews was using a computer that was the size of, I don't know, five houses or something like that to do this. And now we can do it on this tiny little stick of gum-sized dev board. And a lot of people are probably familiar with this from the Space Odyssey 2001 film, where Hal is, you know, they're about to shut him down, and he starts to, he asks if he could sing a song. And then he actually sings the Daisy Bell song. It's this really touching, you know, beautiful moment in the film when the computer is singing the song and then slowly kind of dying as he's singing it.

Yes, that's beautiful. No, I mean, yes, and I know the sound. I wondered why I knew I could hear the song in my head as you mentioned it. And it's because of that scene, not because I actually knew anything about the song. Right, right. Okay, so one of the things I noticed with projects I've seen that use the DAISY is that there are a lot of wires. And that seems to be... A lot of people build things on breadboards.

They build things on breadboards, but then you usually have a moving component because you have sensors in multiple moving components. Potentiometers or keyboards. Or buttons or keys or sensors of some sort. And then they go to the daisy and then audio comes out somewhere. Where does the audio come out? Let's go with that first. Okay. So the audio is coming out of the onboard audio codec. And this particular part, the one we're currently using, can go up to 192 kilohertz on the input, 96k on the output, that's the sample rate,

and up to 24-bit resolution. And that particular chip that I see is taking the digital ones and zeros and converting that to analog audio. And that's what you're going to actually listen to, is you're going to connect the voltage that's coming out of the dev board, coming from the codec, to some sort of connector, which will take it elsewhere to your system. Like a powered speaker or headphones?

Exactly, exactly. And, you know, I actually kind of have a funny story about the audio codec on the Daisy. I like to call this the codec saga. And I tell everyone who will listen, some people get bored faster than others, but I think your audience will hopefully enjoy it. So flashback a few years to 2020,

and this is when we launched the Daisy on Kickstarter. And it was right before what we now know of as the chip shortage so of course we didn't know that when we're launching the daisy or we probably would have waited i don't know done something maybe something different um but the original audio codec that we were using is the ak4556 which is made by akm they're a very high fidelity japanese audio codec company they specialize in audio chips and we were very happy with it. Everybody was happy with it. We designed it on the board.

We got our first shipment of, I don't know, 5,000 or so. And, you know, we kind of had scheduled POs over like two years because we started to get really scared with chip shortage and everything. Now, closer to the end of the year, I happened to see some headline online. It was probably like some clickbait thing, like Japanese semiconductor factory burns to the ground. And I'm like, huh? And I click on it, of course. And it says AKM. And now I'm like, oh, no, this is not good. You know, AKM factory burns to the ground, blah, blah, blah. And I'm like, okay, that's terrible. But what are the chances that it's the exact same factory our chip gets made in? Slim to none,

of course, right? So next week, I get my sales rep emails me, hey, factory burned down. So it's going to be six months before you get your next purchase order. Now, this is like a disaster for us. We're really young. We're just getting the platform going. We don't want to go out of stock. So we locked up all the open market inventory we could, cleared out the catalog suppliers. If you couldn't get the AK4526, it's probably my fault. I'm sorry. We needed them. We did what we could, waited the six months or so that they said it was going to

take to get the factory back online. Then we get another email and they said, AKM actually looked at it and they decided to make this chip obsolete. You're not going to get any of your purchase orders. Sorry. They recommend using this other thing, which was not pin-to-pin compatible.

It was some QFN package, and ours was some sort of TESOP package, right? So this was like a huge nightmare for us. We were freaking out. We were never going to get any more. And at that point, the open market stock had skyrocketed like 10 times the price. So it was just this crazy situation. So what do we have to do? We got to design it out, of course, right?

So at that point, we designed in a Wolfson codec. I think Sirius now, they bought Wolfson, but it's the WM8731, which if you've ever made or looked at a Eurorack module, it's kind of this ubiquitous codec in the Eurorack space. But it was also used on one of the early iPods. So it's kind of a fun fact about that codec. We designed that one in.

Specs were a little bit worse, but mostly comparable. And so it fit most applications. We designed that one in. Everybody's happy. We schedule out our POs. And we're sitting fine for a few months. And the next thing we know, Wolfson emails us or Sirius emails us, hey, this is going obsolete.

Time for your lifetime buy. Right? So we're flabbergasted. I mean, I just can't believe the luck at this point. And so, of course, yeah your lifetime buy. Right? So we're flabbergasted. I mean, I just can't believe the luck at this point. And so of course, yeah, lifetime buy, we lock up, I don't know, 20,000 or so chips, right? Lifetime buy, we're going to be good. Around the same time, the Daisy sales volume took off. So our lifetime stock went, you know, wind up being like six months stock. And all of a sudden, we're redesigning another audio codec onto our board. And that's where we actually wound up looking at this TI codec, which is the one that's currently installed, which is the PCM3060.

And it actually has better specs than both the AKM and the Wolfson codec, which is great. But we had to do another redesign. This just happened, this must have been, I don't know, six or seven months ago. We finally got that into production. And that is the current codec that's on the board. And for everybody who's had to make minor changes to their code, I'm sorry. This is the reason.

Hopefully this will clear it up a little bit better than our forum posts did. How different are they between, aren't they just, it's naive, aren't they just nice I2S on one end and... Voltage on the other? Sort of. I'm sure there's config registers and stuff. There's a lot more stuff to it, yeah.

For that reason, we've tried to simplify the hardware connections and not take advantage of the extra things we can control digitally and just make it simple. Backwards compatibility is really important for us, of course. But each codec is just a little different specs-wise, which could also affect your noise that might pop up on different layouts. And then the output impedance is kind of an issue. So those are things we've learned. different specs-wise, which could also affect your noise that might pop up on different layouts. And then the output impedance is kind of an issue.

So those are things we've learned to deal with because of having to switch so many times. But that's it. That's the codec saga for you. And the chip shortage. And the chip shortage. But you managed to get... You managed to get a processor. Yeah. Yes, yes.

And that was, you know... At the time that the chip shortage really came into effect and we saw our lead times pushing out, there was a lot of companies depending on the daisy to ship their own products. And so we kind of looked at things and just bit the bullet and locked up tons and tons of stock, you know, with the risk of being, hey, maybe we won't make it through all of this, of course. But there were so many companies that it wasn't just us depending on it. It was a lot of other companies that were depending upon that revenue. And so for me, in order to build trust in the platform, I thought it was really necessary to just kind of, you know, double down and lock up the stock. I mean, it's an ARM Cortex M7? Correct. Yeah, it's the H750 series. Is it? Yeah, just so you know how bad it was.

So these were on allocation. So they would quote you like something around 52 plus weeks lead time. But that didn't mean you were going to get them. Because at the end of that 52 weeks, if you didn't understand how the whole semiconductor supply chain works, your sales rep is going to say, hey, like, yeah, you weren't in the bucket because you weren't on the allocation. And I'm sorry. That's just the way the cookie crumbles. Because what happens, of course, is they have to take care of their best customers, the semiconductor companies.

Right. And that's what that's why chips go on allocation. And that's kind of the way it ends up happening is the lead times go out the window. They don't mean anything. They mean that's a minimum. And if you weren't in line by now, you can't have it until then. And being in line is not obvious. Exactly. Exactly. What most people tell me is, look, if they're quoting you over a year, it probably means they don't really know. It probably means never.

Yeah, over a year is go away. It doesn't matter. Stop asking. Who knows? But that was an ST part, if I look at your schematics. And yeah, an STM32 part, those were chip shortage catnip. Just very hard to find. 100%. I can say we definitely developed a much closer relationship with STMicro around that time.

Not coincidental. But I can honestly say we didn't go line down once. So we were very connected and taken care of with STMicro. I can't say enough good things about them throughout that process. Good for you. So this started as a Kickstarter program, and then you fulfilled your orders, and then it continued on as a small business

and then it kind of took off. Is that the right trajectory? What can you tell me about how the business is going? Yeah, so I still am currently running Qubit, the modular synth company. It's a big part of what we do. And the first customer of Electrosmiths to design the DAISY in was, of course, Qubit. So every Qubit product that we've released for the last

three years, three or four years, has used the DAISY at its core. And then a lot of other companies in similar audio spaces, be it Eurorack Modular, FX Pedals, Desktop Synths, they also started designing it into their products. So the Kickstarter was very successful. It was above and beyond what we expected. But the next kind of huge bump of revenue actually came from companies more so than the hobbyists, at least in the post-Kickstarter time period, one, one and a half years. And then once people started seeing it in all these products that they were using, that's when we saw a really big uptick for the first time post-Kickstarter of the hobbyist and education space. That's funny.

Kickstarter worked then. It did. It definitely did. And I started to ask you about the wires that I often see with the hobbyist use of the daisy. Do you see that a lot? Do you have ideas for how to fix that? Is it just part of creating something? I think part of it is the nature of audio somewhat.

I mean, there's, you know, it just tends to require lots and lots of cables. And anybody who plays guitar or especially uses synthesizers can attest to that. Not that that's not just an engineering problem. At large, it is, of cables. And anybody who plays guitar or especially uses synthesizers can attest to that. Not that that's not just an engineering problem at large, it is, of course. But the audio space, there's so many different protocols and different types of connectors and jacks and sensors that everybody wants to use that it's just a little bit the nature of developing with this type of platform. Now, with that being said, a big thing that we've tried to do to mitigate that has been making these breakout boards or example hardware designs, which will

eliminate the need or at least minimize the need for doing a lot of breadboarding. And we'll take common applications, say, if you look at like the Daisy pod, this is just a line level breakout board for the Daisy. And it just so it has a headphone output, it's got a couple pots, a couple switches, a couple jacks, and it's line level audio, and then you just power it from your computer. So that's probably like a daisy 101 project, you're going to just break out the breadboard, wire up all these things. But why do that if you can just buy the pod and kind of eliminate all the wires that you're that you're talking about. So that's actually an an extremely popular product because many people are saying, you know what, screw the breadboard, let's just start with that and I won't have to use all these wires. That makes sense.

And then there's a submodule and an init. What do these do? Yeah, so the patch submodule is taking that same concept just a little bit further because the patch sub-module is basically an entire Eurorack module. It includes power conditioning, CV input output circuits, audio scaling circuits. Anything that you would have to do to make a Eurorack module goes away, and it's all just right there on the board. So you could actually, if you wanted to, you could just connect all of your jacks and your pots to your front panel and then just hard solder wires directly to the patch submodule board. And it's a functional Eureka module.

You don't need any parts. You don't have to do any hardware design or any circuit design whatsoever. So it kind of just takes that concept and takes it to the next level. Oh, I see. The patch submodule doesn't need a daisy plugged into it. It is a daisy with different IO configurations and more stuff. Exactly, exactly. It's just the daisy seed schematic plus all of the circuits you're going to wind up doing anyways. Okay. So you have two additional things that I want to ask you about because they go to my theory that ESD doesn't exist.

I don't know about that. You have the field and the patch, which are these really cool big enclosed... With actual knobs and buttons. Yeah, you turn them into a little desktop synthesizer module or a Eurorack module

with a front panel and jacks and knobs and things. And they're not cheap. And the daisy just plugs into the front of them in a socket. So I worry about this daisy in a bar on the floor. Next to Christopher's drums as the sticks are flying. Or on the rug where he plays. Exactly.

Yeah, that's a valid concern. That is a valid concern. And, you know, our design goal for this was really to show what's possible with the Daisy and make it easy to develop. We weren't necessarily trying to compete with Strymon's Reefers, if that makes sense. Now, they'll probably fare better than you would expect, maybe, in that sort of environment with all the lights on them and everything. They'll probably sound better than you might guess. But at the end of the day, you know, it was more of a development platform than an actual product as such. Yeah. They're very cool, though.

There's something cool about just having the electronics totally visible and outside. Well, the field has a case and the daisies on the outside. But other than that, it's enclosed. Everything else is relatively open and free to hang you know, manufacturing enclosures is kind of a headache and a nightmare. So, you know, if you don't have to, why bother? And, you know, we're happy to help our customers design their own enclosures, which is more often than not what we're doing. And we do include, you know, a lot of mechanical files to kind of ease that process for our customers, you know, like 3D models of the DayZ and recommended cutouts for various like the screen, for example, things like that. But yeah, I'm not a huge fan of enclosure manufacturing. That's all right. We'll just include enough files they can 3D print it themselves. Exactly. Is that what most people end up doing? It depends. And, you know, it really depends on the audience because, you know, we sell to a lot of companies. Now, they're not going to be 3D printed enclosures.

But then we sell to a lot of hobbyists. Even for the pod, people are making these really cool 3D printed enclosures. We see everything. We see all different types. I have a couple of listener questions I want to get through. First, Bailey asks if polyphony is supported. It is. It is definitely supported.

And you'd probably be amazed at how many voices of polyphony you could get on the Daisy. It's a really common application. Anyone familiar with the Chompy sampler that just got released? Yeah, so that has a Daisy inside. Oh, okay. I don't know the exact number of voices of Polyphony, but it has a lot, and it's doing a lot of effects

and a lot of other stuff, and it runs just fine. It's kind of amazing. Audio is not a difficult task for modern microcontrollers. I have some sense, like this one here, the Blofeld, I don't think that's got anything very exciting in it,

but it's a sampler with, it has tons of voices and yeah and i think daisy probably has way more processing power than but we saw the wiggler which was a daisy that was a choice i mean that was somebody's monophonic yeah okay yeah that was their choice but that's been a big trend too you mentioned you know micro modern micro controllers are so good at audio and they are and what we've really seen happen in the last five to 10 years is we don't have to use purpose-built DSP chips anymore. It used to be you'd get your Shark DSP on there, and then you'd still have some other microcontroller handling all of your controls and your ADCs and all that stuff. But the cool thing about the Daisy,

and one of the reasons we chose the particular STMicro processor we did, is that modern general purpose microcontrollers are so fast. You can just use those, and they come with everything else you need anyways. Our chip actually has 16-bit ADCs on board, which is extremely unusual. I mean, even general-purpose microcontrollers don't always have that. Well, and it doesn't have any DSP blocks, but it's got a lot of DSP-specialized code. Right, right. And that's under the code that you provide. Oh, like arms.

CM-Sys. Neon stuff or whatever. Yeah. Another question from Bartholomew. Are gold-plated contacts worth it? This should have been a lightning round question. No, but oxygen- free cables definitely are yeah i don't i don't know how far down this road i want to go because it's kind of a polarizing

dangerous topic so to speak um so let me split the middle okay um we only use gold we only use enig boards so that's a gold-plated surface finish on our PCBs. They are substantially superior and better in every single way than all the other options, Hassle, what have you. Now, when it comes to hi-fi audio setups, I think there have been double-blind studies where 99% of people can't hear the difference. Take that for what it's worth. Between a garden sprinkler controller cable and a $10,000 gold cable. Exactly. Yeah, exactly. So that's my two cents on it.

So sometimes it matters, but most of the time, probably not. Well, you're already doing things at 96K, and if anybody can hear above 40 kilohertz, just let us know. There's reasons to do computations at that speed, but not anyway. I had a thought. What was it? Oh, well. Back to Bartholomew. What are your recommendations for strain relief for corded things? Gosh, I don't know that I'm the person to ask that question to. Okay. Yeah, I would say if you're performing on stage, definitely pull your cable through your strap and then plug it in.

But I think that's about the most expertise I can pull in that area. You mentioned playing guitar in high school, and then you went to music college, but you didn't major in guitar. Yeah, so it's kind of complicated because at the time I went to Berkeley, you had to have what was called a principal instrument. And this could not be a computer. They've since changed that, which I think is incredible. But at the time I went there, you had a major, but that was still separate from your principal instrument. So I was both a guitar principal, as they say, and then I was also majoring in electronic production and design. So I kind of did both.

And I spent a lot of time playing guitar in college. Cool. That leads me to the question. You went to Berklee School of Music. You obviously did some computer stuff there for computer production, for music production and stuff. But how did you go from that to working on designing boards and firmware and working with, you know, microcontrollers and stuff? Signal processing. Yeah, you know, so at the time when I was at Berklee, I knew I wanted to do electronic music

and I knew I loved synthesizers. I didn't have any yet because I was at Berkeley, I knew I wanted to do electronic music and I knew I loved synthesizers. I didn't have any yet because I was still playing guitar. But for me, the thing about guitar, it kind of morphed into this like pedal experience. And it wasn't really about playing at a certain point. It was really about just tweaking sound and creating new timbres with the pedals and not necessarily utilizing the guitar primarily. So my head was kind of already in this space. And then when I got to Berkeley, I took a class on C sound, which is a programming language specifically for music. And it really got me thinking that the ultimate way to make electronic music is to actually make your own instrument and then make the music with the instrument. In hindsight, this is like kind of crazy. And like, I don't really recommend this to a lot of people because I think you can...

Only suckers buy their own instruments. You know what I mean? Yeah, exactly. That's just so wild. That is so inaccurate and not true. And I think it kind of shames like, oh, don't use presets. And that's just, come on. Apex Twin uses presets and he makes them sound amazing. Let's not get upset about presets. Life is too short not to use presets. Exactly. Exactly. But for me, I think this led to, even though my thinking was a little flawed, I think it led to a very interesting career because I got excited about learning how to write code and how to make electronic circuits specifically to make my own instruments,

which I could use for composition. And so it started with C sound, it went to Max MSP, and then I started getting turned on to the embedded stuff. I started learning, you know, of course, Arduino, Makey Makey, very entry point stuff. But then one day, I was in a class with Dr. Boulanger at Berkeley. Everybody's very familiar with his work. And he held up this, this like small rectangular board. And he said, this is the Raspberry Pi. And this must have been, gosh, 2011, 2012. It was basically right when it first came out, right? So nobody had heard about it. And he said, the reason that's important is because it runs Linux. And of course, on Linux, you can run C sound among a number of other open source music programming languages. So that really got the light bulb going off in my head, which was, okay, this is $35, and it can run C-Sound, and it's already embedded.

I can write C-Sound code, and I know these modular synthesizers cost like $300 a piece. There might be something there. And that's really what kind of kick-started my whole career was being crazy enough to think that, sure, I'll just figure that out and learn some engineering enough

to release a product. And I did, and people wanted it. And it was really a fortuitous time in the industry because it was really early on for Eurorack synthesizers. There was almost no digital designs at all. And so when we came out with this Braswell... Oh, really? Oh, yeah. No. I mean, back then, people were basically... Analog and modular were like synonyms. Yeah, yeah. It was the same thing for them. And all these people hated digital.

You know, honestly, they really didn't like it at all. So when we came out with our product, sure, there was some kickback. But most people were just craving these more advanced synthesis techniques that they'd never had access to in the hardware domain. Question from Tom Anderson. Comment first. I've had good results with the Daisy calling the API from C++. Do you have any advice for guitar builders who are looking at building it into their guitars?

The power draw is about one watt question mark. I don't know. What is the best way to get rid of the low level 11.5 kilohertz tone at the output? I feel like, Tom, maybe you need to go to the forum for that part. But let's go with the first part. Advice for guitar builders. Yeah, no, I'll hit both because these are both really great questions. So the first thing to note, effects pedals are extremely sensitive to noise.

It's crazy. And this is something we didn't, we had a little more experience on the synthesizer side of things. So when we started doing effects stuff, it was kind of like a wake up call. Like, whoa, this is a whole different ballgame because you're, you're, you're ample. First of all, I mean, first of all, let's just talk about, you're using these single coil pickups, which is a crazy archaic technology, which almost just inherently have hum. And then you're magnifying it massively

to get up to effects pedal level. And then you're sending it through all these things. And then you're running it through an amp, which is again, another crazy piece of technology. So it's just more sensitive in a nutshell. So you have to be extremely careful, specifically when it comes to doing a layout. Like PCB layout becomes imminently important

in a way that it isn't in a lot of other applications. So I know that's not great advice for people on the breadboard, but it's the truth at least. So let's start with that and then let's go to the second point. So how do I get rid of, he's saying 11 something kilohertz tone. The more common tone that people are going to be talking about is a one kilohertz tone, the dreaded 1k tone. That's a drone. That is not good. So the thing about the 1k tone is it's being caused by your audio callback. So every time your audio callback functions, it's going to pull extra hard on your ground. You know, there's going to be this big current spike and it's going to create a transient on your power supply. And this is going to work its way through to your output circuits

and you're going to get what more commonly is a 1 kHz tone if you're using a standard block size, right? And capacitors. Right, right, exactly. Well, the technically correct answer is, okay, well, let's look at your ground scheme and let's get all these caps and let's do some pie filters and let's do all this.

Okay, yes, that is true. But the quickest and some would say maybe dirty way to fix this is if you're not maxing out the daisies capability, which most people are not, take your block size. Now, this is the number of samples in your callback, but it's just in libdaisy if you're using our code. Just change the block size. Bring it down to a really, really small number. Set it to two. Now, what that's going to do is it's actually going to take that tone and it's not going to make it disappear, but it's going to make it so high frequency that humans can't hear it. And so for like, you know, 90% of your applications, if you really need those cycles, we can have that conversation. It's more

involved. But if you're really not pushing every last cycle on the daisy, just change your block size and it's going to go away. Because you're pushing it above human hearing. Right, because we can only hear 20 hertz to 20 kilohertz. And let's be real, most of us can only hear up to like 15 kilohertz. At the best. Yeah, yeah, yeah. I'm a rock fan. It's depressing going to those hearing tests.

I've done it. It's really depressing. And if you add a lot of filtering to get rid of a tone in your signal chain you're also affecting your signal and 1kHz is kind of an important frequency range that's an important frequency range but above about 15 you're pretty good smashing those exactly

we've talked to a lot of engineer engineers that don't have experience with audio and they're like just get a brick wall filter right there at 1kHz and we're like yeah, I don't think you understand that would sound interesting it's kind of important and they're like, well, yeah, just get a brick wall filter right there at 1K, and we're like, yeah, I don't think you understand. That would sound interesting. It's kind of important. Maybe Tom's problem is he set the block size to 4.

Yeah, so start with block size, and then move on to layout and things like that. I would mention here, too, and I know a lot of people on the forum and Discord already know about this, so it's only sort of secret. We have a new form factor of the daisy coming already know about this, so it's only sort of secret. We have a new form factor of the DAISY coming out. It's called the Seed2DFM. DFM stands for Design for Manufacturer. This particular form factor, it's already been installed in a bunch of products for like two years,

so it's very field tested. We're just wrapping up the documentation, but the reason I bring it up now is because it has differential outputs instead of single ended outputs. And in the effects pedal environment, this pretty much just eliminates all of your issues. And so we've kind of got to the point where if you're running into the tone and your block size is, you can't really adjust your block size and you don't really want to mess with the layout too much, just use this other form factor because it's going to basically solve the problem for you. When I asked you in lightning round about musician, engineer, or CEO, you took a pause.

When you go in on Monday, what are you most excited about working on? Is it the hardware? Is it the structure? Is it planning, software, music, signal processing tools? What is it that you like to do with your company and with Daisy? You know, I would answer this question differently depending on the month you asked me, I think. And part of that is just that, is just that the needs of the businesses changes over time. And depending upon which business I'm focusing on, it changes too. Because of course, when I'm wearing my Qubit hat, we're usually just designing instruments. And so I tend to get really excited about the DSP and the musical applications and what the users are gonna be able to do.

Then when I come over to Daisy, it has to shift a little bit. So let me answer it with my kind of current obsession. And this might sound kind of lame, but my real obsession and what gets me the most excited at work these days is the actual manufacturing process of the Daisy boards. Not a lot of people know this, but we actually do the assembly ourselves. We have an SMT line in shop, in-house, and we do everything, everything except fab the boards or make the chips. So this has been something that I've just been kind of endlessly tweaking. But lately, what happened was our daisy volume got so high that we actually had to install a new SMT line just to handle daisy boards.

And so it gave me a chance to kind of from the ground up structure an entire manufacturing paradigm that was really tailored and tooled up specifically for only two or three boards, which makes it extremely powerful. I don't know if you, you know, if you have any manufacturing experience, the fewer SKUs you're dealing with, the easier it is to achieve great results. So currently, you know, this week, I mean, that's what I've been obsessed with, you know, getting the reflow oven profile just right for the daisy or, you know, evaluating all the different solder paste options and things like that. Okay. Manufacturing engineering wasn't on the list. Why did you bring assembly in-house?

And are you doing this in San Clemente in Southern California? That's right. Yeah, we're doing it here. Wow. You know, there's a lot to say. Let me give you the short answer. I think when it comes to manufacturing my own products, I'm a little bit too much of a control freak, maybe. But it really is about the fact that, to me,

Daisy isn't just a schematic. And I get a lot of companies coming to me and they want to license the schematic because they just want to take advantage of the circuit design and they want to take advantage of the code base. Right. If we're going to license the schematic, you know, because they just want to take advantage of the circuit design. They want to take advantage of the code base. Right. If we're going to build it in, we'll just build it in, not by your board and then have to deal with connectors. Yeah. Okay. But to me, like, DAISY is so much more than that. And knowing how it's built and being able to say, I personally stood there and tweaked that

reflow profile. I know what your field defects are going to be like. They're going to be very, very low, if not non-existent, because I tried so hard to improve it. And not only that I tried so hard originally, it's that we take every daisy defect we find, we work at improving it, and you can't really improve it if you don't have control over the manufacturing stage. And so it creates this kind of closed loop system where all the defects get fed to the engineers, which then gets fed to the manufacturing team, which then improves the defects, but then so on and so forth. You know what I mean? So it's really this process of constant improvement, which enables the platform to

get better every single year. And if we outsource manufacturing to some nameless overseas company, we're not going to get that. I know we're not going to get that. And it might not even be, it might even be consistent from production run to production run. And that's important. Of course, 100%. Especially with, you know, with all the companies that are relying on us to put the daisy inside their boards, we really don't have any margin for these things failing. And the reputation of the platform depends on how well it was manufactured. So at the end of the day, do we really want to give that out, trust that with somebody else?

Probably not, you know. So we've been talking about the hardware itself, but the other big component to the platform is the software. How much of that do you do versus how much do you have engineers who do it versus how much is community managed? I was very lucky early on to have engineers come on board that are way better than I could ever hope to be. So I do very little to no programming these days. I have fantastic engineers that do a great job on that. So as far as the code base overall, I would say a good 90% of it

is still internal firmware that we're writing, software tools and whatnot. But we do have, if you check out our GitHub repos, we do have a lot of community engagement and a lot of external contributors, which are really helping take the projects to the next level

and keep us on task with that as well. What kind of projects do you see DAISY being used in other than Qubit and the things you work on yourself? So, you know, of course, there's the modular sense, that's the bread and butter, the effects pedals, the desktop sense. There's been a few really interesting areas that have gotten me really excited lately. One of them is in the educational space. We've seen it being used in courses for all sorts of things, whether it's creative coding, intro to programming, intro to hardware design.

It's being used at RISD, which is actually just a design university. It's not an engineering college as such, right? They're using it to teach how to design modular synth modules, right? So it's really branched out to all these different areas. That's one area that's been particularly exciting for me just because of my passion for education and especially in the non-engineering space. And one other project that popped up on the forum lately, which I hope continues, and if anybody out there is an amateur radio operator, a ham radio nut,

please make things with the DAISY. It's such a great platform for that. I was about to ask you, have there been non-musical signal processing applications? Not a ton. But I did see a software-defined radio, an SDR project popped up last week, and I'm a ham. I'm a KN6

TAU, if you see me on the airwaves, that's me. So that gets me really excited, because it's not music-specific. Now it's just audio, you know, and it's so powerful, but I don't think it's gotten the notoriety in that space that it really deserves. There's a lot of things that need signal processing. Yeah, yeah. Well,

in audio range. Right. Yes. Which is still a lot, but... Yes. Yeah, yeah. Well, in audio range. Right. Yes. Which is still a lot, but... Yes. No, it's a cool platform. Sorry, I've gone off into design land myself, so that's not good radio for sure. So Christopher, drummer, he plays bass guitar, plays some guitar and he plays

piano and he does embedded software, um, but he hates it. So take that with a grain of salt. Um, and he's the cohost of this show. So, and he does all of the making a sound good part of it. What should I get him from your website? You should get him the field. Without a doubt. Despite his fear

of the exposed PCB, I think it would still suit his purposes best. It has a lot of buttons and he does really like buttons. Oh, and they light up. I like light up buttons. Yes.

Got light up buttons, got lots of those, got lots of LEDs. Ins and outs and cables. Ins and outs. And yeah. And that's the big thing about the field too

is it can process audio, but it can generate audio. So if you're both a piano player and a guitar player or a bass player, you can run your bass through it or you could connect your MIDI keyboard up to it. He has a few of those. Sorry. Stop looking at my desk. We're in the studio, which is surrounded by instruments and whatnot. Do you get to play music very often? I always hate this question because I think I disappoint people so much when I answer it,

but I'll be honest. I'll be honest with you. So I work a lot of hours. And when I get home after a 10, 12 hour day, messing with synthesizers is not always top of my list. So I don't really patch as much as I used to or nearly as much as I used to when I get home, but I have actually gotten into playing the ukulele. And it's been really interesting. It's been kind of this therapeutic way to reconnect with music in a way that doesn't use electrons. It's been this kind of, I guess, cathartic. It's like this real therapeutic thing I could do on the weekend. So I do play music. I'm making a lot less electronic music than I used to,

which I'm working on. But I have been playing a little ukulele. I'm a huge advocate of low-friction musical instruments that you can just pick up and that are fun to play, don't require a lot of technique necessarily. Sure, sure. So when you're tired, you can actually still do something. Well, and there's some push and pull.

If you spend your day, like I do some writing and I enjoy writing, but if during my job, I am doing a lot of writing, like I'm finishing a technical book now, so I'm doing no writing for fun. But if my job was all programming and hardware and I didn't have any writing outlet, I would probably write for fun. And I have in the past. So there's some, if your job is all about making instruments all the time, going home to think about instruments. I mean, you might still get to think about music. Thinking about instruments, probably that part of your brain may get tired. Some other part of your brain needs a turn. Exactly, exactly. And avoiding

burnout is so important. I've been really excited to see more talk about this since the pandemic, but it's extremely important. And I think when you make passion your career, it's even more dangerous for this very reason. Yes. Well, I usually say if you make your hobby your career, then what are you going to do in your free time? Exactly. Exactly. Or what are you going to do when you burn out and now you don't like your hobby anymore or your career?

You don't have either. Exactly. Do you have any advice for staying away from burnout or after you've gotten into it, getting out of it? That's a good question. You know, I would say one thing is it's best to think of it as a marathon, not a sprint. And for me, whenever I do find myself kind of in the throes of burnout, it's good to kind of reconnect with what got you excited about it in the first place. So for me, sometimes I'll try to find an instrument that I didn't design or I didn't have any connection to that is exciting and play with it.

So for me, I've mentioned this before a couple times, but Seat Lombard is this modular synth manufacturer who makes things that are so wild and different than what I normally make. There's been a few times when I get super burned out, I'll just plug one of those in and play with it. And it's just so different from what I'm used to that it reconnects me because I get excited about electronic sound, but I don't have to play with my own stuff. And that helps. It does help because if you're playing with your own stuff, you see bugs in it.

Oh, I see the bugs I battled all day at work and then I have to play with them at home. Or even if things are going well and there are no bugs, you still have to be thinking about how a customer might approach this. And if you are the customer, there's some critical part that can turn off. There's definitely a novelty factor to things. Especially with music, I find like

that's why gear acquisition syndrome exists, right? It's like I'm bored with all this stuff, but if I get something else then maybe I'll be inspired. And like you said, with the modular thing you were playing with. But if you're so close to something and you're building it, I mean, it would be hard to get any separation from the concepts even to, yeah, it would be easy to get lost in that, it seems like. Well, Andrew, it has been wonderful to talk to you. Do you have any thoughts you would like to leave us with?

Yeah, you know, I'd like to share one of my favorite quotes, which is from Arthur C. Clarke. It says, Any sufficiently advanced technology is indistinguishable from magic. Now for me, like being a lifelong musician, growing up, music was magic, pure magic. It's just endlessly thrilling to me. Now coming into technology a little bit later in life,

this combination kind of just blew my mind and continues to do so. And, you know, being able to come to work every day and combine the magic of music with the magic of technology is really just an absolute dream. Our guest has been Andrew Eikenberry, electronic instrument designer and music tech entrepreneur. Andrew is the founder of Electrosmith, Qubit, and 2HP. Thanks, Andrew. Thank you. Thank you to Christopher for producing and co-hosting. Thank you to our Patreon listener Slack group for their questions.

And of course, thank you for listening. You can always contact us at show at embedded.fm or hit the contact link on Embedded FM. And I can't really think of a quote to leave you with. So I guess I should probably tell you that ElectroSmith has offered us 5% off if you type in Embedded FM in their coupon code. And this is good for a month after the show goes up. So if you're hearing this before about mid-March and you want to check it out, 5% off Embedded FM.