2024 Chapter 11
Get Forked!
How do you start this story?
If it was a movie, I know how it would go. It would involve terribly evil corporate interests working to screw the lone innovator who came up with the idea, with either a cynical ending showing you can’t fight the big money, or an unlikely justice-will-prevail ending involving an internal coup and/or external grandstanding.
But real life isn’t a movie. And I’m not Abed from Community, able to only tell tales in modern entertainment metaphor. So let’s tell this story, and see how it goes.
Act 1: B.F. (Before Forkbeard)
Open on an idyllic scene: Schiit Audio, comfortable in What They Do and Where They Are. There are pressures, of course, for us to Do The Current Thing, which is to make a music streamer with a music player app.
Because that’s what all the cool kids are doing. Everyone in high-end has gone bonkers for music streamers, and everyone has to have one. And if you have a streamer, you have to have an app to run it.
Simple. Easy. Sell a streamer to all your customers, make a lot of money!
Yeah. Except for that app thing.
Because having an app means one of two things:
- You’ve ceded control of a big part of your company to whoever is doing the app. Sure, there are a ton of companies who will tell you how easy it is, how they can keep it up for you, yadda yadda, easy peasy…and then you find out it’s gonna cost a ton of money and they can’t do exactly what you want because of their framework, their standards, their other customers…
- You’ve done it yourself and become a software company by default. Now you have to update the app, chase bugs, fix it when the OSes change, manage different versions across a ton of different devices, hire people, have staff specifically for the app.
Which is why we said, “Yeah nah, that sucks, we ain’t doing it. We’re gonna stick with our antique tech, including old-skool IR remote controls, because there ain’t no software and it always works. Take your apps and shovel ‘em.”
At the same time, this idyllic era wasn’t perfect.
And our reliance on old tech wasn’t 100% wonderful and trouble-free. As in, we developed a ton of products with remote controls.
Different remotes.
And we ended up where you always end up in remote-land, which is in the
Land of Too Many ****ing Remotes.
Yeah. I know you’re there too. You have a random batch of stupid-ass remotes sitting on the ottoman in front of the TV.
You know what I’m talking about.
Remotes
suck.
And we weren’t immune. As in, some customers resorted to 3D printing remote racks with the names of products on them. Others painstakingly set up universal remotes (which also suck) with our remote codes in order to help stem the proliferation. And a few went so far as to have custom programming done to integrate their products into a modern multi-product, multi-manufacturer system, only to find out their spouse, partner, or progeny wouldn’t use the over-complicated mess at all, preferring to listen to a HomePod or a soundbar.
To summarize:
all remotes, all types, all approaches, all suck.
Full stop.
Aside: for the bros saying, “Universal remotes are the way,” no sorry they suck massive ass mainly because they need tons of setup no normal human wants to do, and as soon as you hand them to someone, they don’t know how to use them, and, oh yeah, they are fairly expensive and still not bidirectional and still, yeah, sucky. To be a bit more kind, they may work for you, you obsessive and intelligent person, but they are not a panacea.
So it’s in this world we live. By rejecting the Current Wisdom of Needing a Streaming App, we were sane. But by ignoring the problems with remotes and product proliferation, we were whistling past the graveyard.
Until one day…
Act 2: The Demo
Schiit Audio, like other companies with personality, opinions, and unique products, has fans. Sometimes the fans contact us. And most of the time that’s fine. Signing a few products, etc is no big deal.
But then there was Martin.
Martin was a super fan. He bought a lot of products. And he wanted to come by and talk about an idea.
Every once in a while, I can do this. In Martin’s case, I did. He came, we talked for a bit about an interesting idea (which we did not do) and that was that. I mean, yeah, we talked about high end audio a bit, and Schiit’s challenges, and I might have mentioned the whole remote thing, but that was the end.
Or so I thought.
A bit later, Martin contacted us again, saying, “I have something to show you.”
Now, this is the time I usually say something like, “Wow, I’m really busy (true) and I don’t know when I might have time to meet (usually semi-true),” and most sane people would tell themselves,
yeah maybe this is the end of all that.
Aside: I don’t like doing this. I do like meeting people. I think about doing even more—creating a curriculum to teach analog design, for example—but the mechanics haven’t worked out, or I haven’t wanted them to work out enough, so I don’t do a lot of these things.
Aside to the aside: engineering is taught almost 100% wrong, at least in my opinion. It starts with math and theory and lards on super-complex concepts and useless stuff like “here’s a circuit with 17 resistors, calculate the voltage at X,” to the point that most sane people would run and hide from the crazy people. I’d start with five parts and build something that really worked, like a headphone amp. I’d also start with P = VI, because watching something go up in smoke is a powerful teaching tool. Sure you can simulate it with 1 ohm resistors, but build it once and you see why you use 1k instead. 5 parts and a FLIR will teach you more in an afternoon than many engineers learn in years.
Aside to the aside to the aside: doing things—making things happen—is the real arbiter. You can have a dozen great ideas, but are you really ready to commit to making it through hurdle after hurdle—financial, practical, technical, social—in order to make the idea viable? When you find an idea you’ll walk through the fire for, that’s the right one. I’ll talk about toasters and teaching, but we’ll see when and if I feel like really digging in on those. Hell, ask me about Froot Froot Froot sometimes. I get lots of ideas. I don’t do most of them.
But, in Martin’s case, I said “Yeah, let’s meet.” Mainly because he was a smart guy with interesting ideas. And probably partly just because my timing aligned and it made sense. Sometimes the universe throws you a bone.
And oh what a bone.
Because, in the space of 5 minutes, Martin
changed the course of Schiit.
Here’s how it went down:
Martin showed up at the Valencia lab with a PC board, a power supply, and an iPad. On the PC board, there were a number of modules I recognized as Bluetooth or maybe WiFi modules, the kind of thing we’d played with when trying to get Bluetooth audio to not sound like ass. These modules plugged into the base board with simple pin headers; the whole thing was powered by a wall-wart. Next to each module was a series of LEDs, which looked suspiciously familiar, like something we’d do.
Martin plugged it in, and the Bluetooth modules lit up. He then opened an app on the iPad, which displayed a stack of Schiit gear, done in line drawing form (you know, like our manuals.)
Now I knew why the LEDs on the PC board looked familiar: they mimicked the front panels of several of our devices: Yggdrasil, Freya+, Loki Max, Tyr, etc.
“Is this what I think it is?” I asked, thinking,
Okay, this is some kind of remote control app.
“Probably,” Martin said, in his clipped, formal way. “But let me show you how it works.”
And then he showed me one of the most amazing things I’d ever seen.
“This is an integrated Bluetooth remote system,” he said, or something like that. “But it’s simple. No pairing. It uses broadcast to know what products you have and what you don’t.”
Martin unplugged one of the Bluetooth modules. One of the components in the Schiit stack suddenly went gray and unavailable. He plugged it back in and it reappeared.
“It also has complete bidirectional communication,” he said, pushing the button on the iPad stack and watching while the LEDs on the PC board changed, then pressing the button on the PC board and showing that it changed the input on the iPad stack.
“Volume works the same,” he said, pushing and swiping the iPad control and showing the volume changing on the PC board, and changing the volume on the PC board and showing it reflecting on the iPad.
My brain exploded.
This was
dead simple, intuitive, multi device, bidirectional remote control…literally the best implementation I’d ever seen of it, too. It was far, far beyond what anyone else was even thinking about.
And it was EASY!
Aside: **** you Apple, your crap is hard to use these days. Fire the buttheads doing 6x levels of menus for crap nobody cares about, and ***, offer some modes on your devices that are not geared to teen dopamine freaks.
Aside to the aside: and, for the love of god, put the new iPad back in the squisher camera-first and flatten down that damn bump. Anyone using an iPad as a camera needs to be slapped.
No, I mean EASY. I mean, like the glory days of the iPhone. Like stuff we aspired to make back when we were doing apps. Like the way things SHOULD work, in a perfect world without committees of people with opinions from designer nephews.
“Where did you get the board?” I asked, pointing at the PCB with the modules on it.
“Oh, the modules are a standard BLE part—”
“No, I mean the board. The one laid out like our stuff.”
“Oh, I did that.”
“So you’re a hardware guy too?”
“No, I do apps. But I learned some layout software and did a simple board…”
My mind imploded.
“Wait a sec,” I said. “You
learned KiCad and
laid out a board and got it and
stuffed it and then
did this app so you could just show it to us, on a lark?”
Martin smiled. “I knew you didn’t like apps or Bluetooth, so I thought better to show you.”
OMGWTFBBQ.
I mean, this is something like I would have done in the Theta days, if we’d had online layout software and easy PCB ordering and if I knew a thing or a thousand more about apps.
Aside: guys, this is how you do it. This is having an idea and WANTING it. This is the way. Unless you’re sitting on a trust fund or have photos of the governor with donkeys, this is the only way you get ahead. Take notes. And find an idea worthy of WANTING it.
“Crap,” I said.
Martin looked worried. “What’s wrong?”
“I need to get Bluetooth on our products, pronto. And you have to show Mike.”
Act 3: Reality Montage
This is the schiit that movies don’t show you. You know, movies go like this:
- Dood gets bit by radioactive spider
- Dood finds out he has Powers
- Dood who has never sewn a sock in his life suddenly creates a superhero suit that actually cost $3 million and took a team of 47 practical effects guys to create—yes seriously, my wife glued the eyes on the spider-suit, she was one of many many people starting with sculptors and continuing with 3D printing gurus and ending with custom fabrics and such
- Dood shreds multiple suits fighting baddies
- Baddies are conquered, all is well
So yeah, that #3 thing.
That’s where we’re at now. Martin brought in his radioactive spider, we’re bit, we realize we have something really cool, now we gotta make it work.
Both of these pronouncements—needing to get Bluetooth on our products, and showing Mike to get his buy-in—are not trivial. Not by any stretch of the imagination. Hell, I could write a whole chapter on the thought process behind the Forkbeard module (which started as the BlueSchiit, or BS module, but I will condense it down here for sanity’s sake.
Subchapter 1: getting BlueSchiited
Bluetooth is not trivial or easy, no matter what anyone wants you to think. Bluetooth low energy control may be simpler than the morass of Bluetooth audio codecs, but it’s not exactly simple, especially if it’s gonna reside inside a stout chassis of steel and aluminum.
I mean, yeah, we could redesign with a plastic window, but that’s not so great, especially if its on the back of the product, facing away from the controller. Ideally the Bluetooth antenna should stick out of the chassis, which leads to the obvious option of having a Bluetooth antenna connection, with an external mini-pecker that will snap off and generally be a pain in the ass. Ah, no thank you.
So, from the start, I figured this should be a module that plugs into the chassis. A plastic dongle that hangs out the back, so the Bluetooth chip antenna could be exposed in all of its glory.
This approach had a bunch of things going for it:
- The same module could be used for multiple products.
- Only one Bluetooth certification would be required, rather than one for each product (ack).
- Tinfoil hat conspiracy theorists could choose not to have the module at all, defeating our evil plan to record their listening habits on our insidious app.
(3)—just kidding, we can’t take data like this without your permission, and you’ll notice we’re not asking for it. But if you don’t like the idea of controlling your stuff with an app, you can continue using IR, just like the Good Old Days.
The problem was: what module? What connector?
There are a limited number of standardized physical interconnects—USB, Ethernet, HDMI, etc. All of them invite a customer to plug the wrong thing into that socket and have a bad experience.
So from the start, I also figured: use a custom physical connection. One that was keyed, so you couldn’t plug the wrong thing in. Which is how we got to the weird semi-rectangular hole you’ve probably been wondering about on Kraken and Saga 2. There’s no way you’ll plug a Forkbeard module in wrong.
Aside: although it wasn’t called “Forkbeard” until later. BlueSchiit was the codename for a while, until I decided that was too cheeky, so Forkbeard, the son of Bluetooth, is what we went with.
Aside to the aside: for the history buffs, yeah I know, Forkbeard is not a nice guy. Neither was Bluetooth. Neither were many, many historical figures. Life is not ideal. Sorry to burst your bubble.
And, very early on, we started playing with the physical BlueSchiit/Forkbeard connection. Heck, I’m surprised nobody noticed it on the board shots for Kara. The catch was: those were early, experimental implementations. It didn’t work 100% on Kara, or on early Vidars. But those led to the Kara F and Vidar 2F that we have now. Bottom line, if there’s a Forkbeard port on the back, the product is Forkbeardable. (Argh, is that a word? Probably not.)
Beyond that, the module itself presented its own headaches, including documentation ranging into the hundreds of pages, incomprehensible instructions only resolved by experimentation, and several revs to get it to work with the multiple microprocessors we use in our products.
Aside: yeah, you have fun ensuring a module will work on 5V or 3.3V, PIC24 or PIC32 or DSP or (redacted), at data rates beyond your usual I2C for future stuff, it’s not all fun and games. Shoutout to Stephan and Evan, who worked through some of the earliest firmware issues, and built the first run of modules by hand, respectively.
But the module—making it small, making it universal, making it work with most everything we make that has a microprocessor, revising every product so it can use the module, yeah, that’s child’s play when it came to the next challenge.
Subchapter 2: getting Mike on board.
“I don’t have time for this!” Mike said.
Which is a frequent refrain. Mike wants to spend time with his family. He wants to enjoy a pastoral life on his acres in California (no, seriously, he’s in the boonies down a dirt road on a well—neener neener, we’re going to be down a paved road on a rainfall capture/solar house in Texas soon, Mike is 100% right to GTFO of the cities).
And Mike, like me, hates Bluetooth. And apps. And all that implies.
Aside: because apps mean we are a software company.
Aside to the aside: With Forkbeard, that’s OK.
But I knew I had to get Martin in front of Mike—Mike had to see what BlueSchiit could do. Because Mike’s team does what Mike wants, and if Mike said that BlueSchiit was bunk, BlueSchiit was dead in the water.
So Mike blew off one meeting, and didn’t make another.
I started to sweat.
Finally, I figured out a cool-but-shady deal: have Mike come by Valencia for a meeting, but don’t mention Martin’s demo.
That worked. Mike came by. Martin was there. Mike swore a bit about Bluetooth and how we’d never be a software company and how apps suck and all this internet of things crap is crap and more along those lines…
…and five minutes into the demo, he looked at Martin and asked, “Can we work on consignment?”
Yeah.
Walls down.
Paradigm shift.
We were gonna be a software company.
Now, the story wouldn’t be complete without discussing this consignment thing, because some of you are crossing your arms and huffing about horrible capitalists and such.
Here’s the thing: Dave still chooses to work as a consultant with us, with a vig on products he designs. That’s consignment. Also, I chose to work with Mike at Theta on consignment, which earned me enough money to start my own company. Don’t like consignment? Don’t do it, or change the rules.
Now with Martin, no sorry you get the anti-Hollywood ending. No consignment in sight. We hired him. Because that’s what he preferred.
Aside: have a great idea? Show us. You may be surprised.
Act 4: Oh Gawd What Have We Done?
Where movies don’t go is into the details. You know how it goes: hero decides to do something, assembles a team, team does some horrible fake soldering on ancient PC motherboards, and next thing you know you have a total breakthrough, a world-saving device, a shiny object not producible on the planet at all (seriously, who made the helmets in 3 Body Problem?), they use it, there’s some nailbiting moments, but bing bang, everything’s fine.
Except it never works that way.
I mean, I knew what I’d signed up for, at least intellectually. It started with:
Revise all of your products.
But it took a bit to sink in. By “revise,” I mean,
revise.
As in:
- New PCB layout
- New BOM
- New chassis
- New firmware
- New full testing and qualification
…for
everything.
EVERY. THING.
And that’s while we we also needed to do:
- An app, at least for one platform
- Hopefully apps for more platforms
- A Bluetooth module, including an injection-molded chassis for it
- Signup and qualification with the Bluetooth Authorities
- Solidification of features, including security and some advanced things we thought about (more on this later)
- Documentation for possible provisional patent application (because of the advanced things)
- Testing of multiple products in multiple scenarios
Oh and by the way, this was also gonna hit in the year that:
- We were developing key new products that would be affected by Forkbeard
- We were also buying a building in San Antonio for expansion
- We were also moving California production to Texas
- Rina and I were selling the Cali house
- And building a house in Texas
So how do you do this?
Well, if you’re a normal company, you don’t. It’s just…
too much.
Too crazy.
Too much chance for something to go very, very wrong.
But we’re not a normal company. And, to date, the massive fallout from all of this insanity has been…
I’ve been a bit scarce around here.
Here’s the reality:
People can accomplish the seemingly impossible…as long as a certain set of rules are followed.
“What rules?” someone pops up to ask. “Rules like working 16 hours a day 7 days a week, subject to the tantrums of a crazy dictatorial boss, yeah no thanks to that.”
No.
Rules like:
- Everyone is personally involved. As in, there are no pencil-pushers with Gantt charts sitting back and sucking their teeth whenever something goes off track. Everyone has to play. In our case, that meant me, Mike, the digital team, Alex, Tyler, the tech staff, production, everyone. But it started with me.
- The impossible task has a limited duration. As in, “it’s gonna be hell for 8-12 months, but after that it’ll go back to normal. Ish.” You can’t push forever. That’s a lesson a lot of tech companies (and ad agencies) learned the hard way.
- It’s acknowledged that this is crazy. As in, “nobody would do this, we’re trying for something that no sane person would do.” As in, “we know we have a chance of screwing this up, but we’re going to try anyway.”
- There are reasonable options in case of a screwup. As in, if the entire operation fails, the whole company doesn’t implode. It’s OK to fail. It’s OK to scale back. But that’s not what you’re shooting for. In fact, you’re shooting for more.
- The Impossible Operation has a big payoff. For the company, personally, money-wise, whatever…but it’s not just “for the hell of it.” There’s a reason—a real reason—for the big crazy.
For me, this meant going back into “startup mode.”
Startup mode means, no sitting on ass, no watching TV, no doomscrolling, no sitting back…literally get up, do some PCB layout, go to work, do stuff needing to get done (new runs, etc), get home, fire up the screens again, and start working on the hundreds-of-items-long to-do list. Free time on the weekend? You can do documentation. A bit of quiet? Finish up more boards. Waiting for a meeting? Build that prototype. Yes it’s gonna be interrupted by the new building and the house and current stuff being stupid and a hundred other things, but you keep pushing and eventually you get there.
And that immersion matters. Without that immersion, we wouldn’t have had some of the truly crazy, truly breakthrough stuff that came about due to Forkbeard.
Sounds insane?
It is.
But it ends.
Aside: and if you think anything about this is nuts, go back to Martin’s demo story.
Back to “what we got ourselves into.”
From the start we ensured #4 was never a problem. As in, if Forkbeard totally faceplants, we’re fine.
How?
Oh this is fun.
- We developed a single Forkbeard Bluetooth communication module for all products. That’s the red widget sticking out of the weird hole on Forkbeard stuff. One module means one Bluetooth qualification, one development for all products, NOT development and qualification for EACH product (shudder), one common codebase, one common interface. This dramatically reduces the work…and the chance for problems.
- We made it modular. Don’t want Forkbeard? Don’t get it. Go ahead and use the IR remotes you have. Go ahead and 3D print racks to hold them with. Go ahead and train a universal IR remote. No problem. It even saves you some money.
- We made the physical connection standard—and foolproof. Forkbeard uses a custom keyed hole in the chassis, so it can only be inserted one way. It also can be hot-plugged, tolerates products with 3.3V and 5V processors, and has tons of protection from ESD and such built in. Oh, and by having the module hang WAY out of the chassis, we get great Bluetooth range—no worries about something being blocked by all the metal around it.
- We absolutely minimized the cost of implementation on every product. As in, most products have 1-3 parts added…a $0.08 connector, and two $0.01 resistors. That’s 10 cents to enable Forkbeard from a hardware perspective. The cost of adding the hole is nil, since all our products are tooled. Of course, all products need different firmware to communicate with the Forkbeard module, but they already have the microprocessors and sensors in place. Forkbeard just makes them visible…and interactive.
So, let’s say Forkbeard is a total fail. That leaves us with a bunch of products that we spent 10 cents more on, and have a weird hole in the back (covered with a plug). And that’s it!
Aside: but I don’t think it’ll go that way.
But even with all of this modularization, even with this very safe and sane approach, we had to pull back the scope at bit at first.
Why? Because it was too big.
Tell an engineer “hey just revise everything to a massive change that is gonna be a moving target for a year or so,” is recipe for complete brain lockup. And that’s pretty much what we got. Dave stared blankly at me when I announced it and spluttered about what and how and when and schedules. And even I balked when I realized that I had well over a dozen products that could be “Forkbeardified.”
Aside: the name “Forkbeard,” comes from Svein Forkbeard, Harald Bluetooth’s son. You can thank Rina for it, as well as the Schiit name. I came up with a horrible logo that used the runes for Svein Forkbeard, then simplified it significantly to create the current diamond.
So that’s why we very quickly scaled the project back…to focus on what we called “the GigaStack.”
The GigaStack was 3 products:
Kara F was just a Kara, proven for Forkbeard. So it was pretty much a gimme.
Gungnir 2 was the single digital product that Dave could concentrate on.
Wotan was the single analog product that I could concentrate on.
Now suddenly things felt a lot better. Three products wasn’t too insane. Three products would let us show how Forkbeard worked across multiple devices. It also provided a truly insane, high-value stack of DAC, preamp, and BIG power amp…for a price less than “entry level” integrateds by some brands.
Aside: no seriously, let’s talk about a True Multibit differential DAC with our all new Multiform Topology, a fully discrete relay ladder attenuated preamp, and a 200WPC Nexus stereo power amp with our first-ever switchable feedback system, with Forkbeard control…for $4400? Ah come on, why do you buy anything else?
But, even pared back, that still meant a ton of things to do.
One of the first things on the list was wireframes for the app. I didn’t know Martin super-well at that time, and I’d done a lot of wireframes for apps in my day, and I wanted this one to be right. I’d seen a lot of simple interfaces turned to garbage by too much input, good or bad—
Aside: hey Apple, reminder you still are, ah, well, gettin' really complimicated.
—so I went ahead and did a bunch of provisional wireframes and sent them to Martin.
Which again, oversimplifies things.
What really happened was:
- I decided to do some wireframes
- I started putting them together and quickly went down a rabbit hole of “oh gawd should it be more like the product or more like a modern interface”
- And at the same time had a couple of crazy ideas that could make Forkbeard sooooooo much better
- So I did all of it and added a bunch of notes and sent that mess to Martin
Who got back to me quickly, with a lecture about size of elements in iOS, and consistency, and stuff like that…but he generally agreed with where I was going, and liked the simple (less product-centric) approach.
And then we discussed the crazy ideas.
Because, while I was doing the wireframes, I quickly realized we could use the current sensor in the amps as a instant power output meter—kinda like a VU meter.
Aside: ORT, here you go. Finally. Because they are useful to know you’re your amp is clipping. Unfortunately it’s on your phone. Stick the phone to the amp. Done.
Aside to the aside: and anyone complaining about getting their phone, cost of phones, etc, please bite me and look at the cost of refurbished iDevices, or go through your drawer and find an old one to use. Done and done. Or wait for Android and Windows. Get a Fire Pad for $59, and laugh at the $220 universal remotes out there.
Aside to the aside to the aside: and now we have the best touchscreen on any audio product…because it’s NOT ON THE PRODUCT. And you can upgrade it whenever you feel like it. Neener. Neener.
I think in my original mockup, I had the instant power output meters as a subscription-based extra cost item, but that’s a joke. We’d never do that. That’s truly evil.
And I quickly realized that we could go beyond power meters. We could do system health—temperature and current reporting. We could do tube integrity. We could let you know if you’re running in Class A or not, on products with high bias.
Aaaaaaaannnnd, I realized, with an entire system, you have a whole lot of data and control and you can do a whole lot more. And that’s why some of the most mindblowing stuff is still coming. (And no, you won’t pay for it.)
Oh yeah and did I mention that Forkbeard is inherently multi-product AND multi-stack? It’s fine if you have a living room system, a bedroom system, a listening room system, and a garage system. Organize all of them and control the one you’re listening too.
Aaaaaaaannnnnnnddddd…
…some of you have noticed that weird “ADVANCED AND EXTENDED” leg of the Forkbeard overview diagram.
Yeaaaaaaahhhhh…
I actually took all of the following out of a draft of this chapter only a couple of days ago. Because I didn’t know if we’d have anything ADVANCED AND EXTENDED to show at, like, the show. Not a big deal, because Forkbeard was already a big deal, and we would just add it later (and for free, more on that in a bit).
But we do have the first of those ADVANCED AND EXTENDED features to show.
What is it?
Something we call Visual Volume™.
Visual volume looks at the total gain, output, and capability of your exact Forkbeard-enabled system, and
tacks yellow and red ranges on to the volume knob.
As in,
you always know how much headroom you have.
For your exact system.
Yes, even if you change gain. Even if you change your stack to a different amp. Want to know if Aegir is enough for you? Well here you go. Want to know how deep into party mode you can get Wotan? Not a problem.
Oh and by the way, that’s just the start of the inter-product features.
Starting to get WHAT A HUGE FOOKING DEAL THIS IS?
That’s when I said,
crap, this may be patentable.
And, after a few convos with friends deep deep in the tech world, I decided: yeah we need to go for this. Because Forkbeard is unique, and what we’re showing now is just the tip of the iceberg.
But I am getting ahead of myself. This was (and is) still a big, crazy, all-consuming project, that led down so many rabbit holes that it’s amazing we pulled it off…and ended up with more Forkbeard products at launch—many more—than just the “giga-stack.”
I mean, we literally just achieved the first communication between the Forkbeard module and the internal microprocessor late in 2023 (thanks, Stephan), and the first working demo of a single button-press on an iPhone affecting a Forkbeardified Schiit product in January 2024 (thanks, Martin). Along the way, we had to overcome a mountain of gotchas, including:
- Manufacturing of the Forkbeard modules. These use a standard Bluetooth module with some glue logic, etc, but we needed to produce a hundred or so of them for testing. This led to us getting a reflow oven, and also dramatically enhanced the production of prototypes for dense boards like Gunnr. Thanks Evan!
- Dealing with different microprocessors. Some of our products use 5V microprocessors, some use 3.3V microprocessors, some have internal nonvolatile memory, some don’t. Changing everything to a standard microprocessor would be a nightmare, both for testing and firmware development reasons and for stocking reasons—we have significant stock of both parts. So Forkbeard had to accommodate all variations. After some hardware and firmware changes, it did.
- The reality of security. Although we really really loved the initial concept of Forkbeard—turn it on, it sees your products, and you start using them—that doesn’t take into account situations like neighbors in the apartment sharing a wall with you having Forkbeard as well. So we added a PIN to the startup. But that also meant…
- …managing the need for change. What if someone else buys your gear? Now we need to release the PIN. Sounds easy? Yeah not so bad, except some of our newest products run the 3.3V all the time in sleep mode, so you can’t count on that reset. After a while, we settled on having you unplug the product and then turn it on with a button press to accomplish this.
- The race to Kraken, Saga 2, and anything else that might release before Forkbeard. Releasing any product that had an open Forkbeard slot meant that the firmware standardizing communication between the products’ internal microprocessor and the Forkbeard module had to be tested, broken, fixed, and 100% reliable before we shipped a single product. Which is why that part of the Forkbeard code was done in May…long before we even had a cosmetic Forkbeard app.
- The long and short, or ugly and simple, debate. The earliest version of the Forkbeard app displayed everything…literally every product we were working on that might end up with Forkbeard. It also showed, in gory detail, the exact temperatures, operating currents, and any other fault states, right down to several decimal places. This is a HORRIBLE customer experience, and it’s HORRIBLE for us, as well. Constant questions about “well I hear so-and-so is running at 43 degrees C, but mine is 49, and I see one channel is running 110mA, but the other is 105mA, and nothing changes even when playing music, etc…NO NO NO NO. Hence the simple icons that let you know when there is a problem, and power meters tuned to the reality of what the amp is doing.
- Multiple. Mindblowing. Moments. Like when I jokingly said, “It’s too bad we can’t update Gungnir 2 via Forkbeard instead of the SDcard in Autonomy, and knowing instantly his blank, vaguely surprised look meant that WE COULD. Again, to be clear: you update Forkbeard products via the app. No sending them in, no SDcard, nada. The app lets you know if there’s an update. And then all the other crazy ideas, stuff you’ll hear more about next year.
But this is still underselling it. It was a big crazy project, but in about a year, we had products talking via Forkbeard (that first working product was the subject of a cryptic post I made on Head-fi), and then when we had that first product running and reporting current in real time…that’s arguably a bigger deal than anything we’ve ever done since the start of the company, and then when we all started using it, all the time, on nearly everything…once you start using it, you don’t stop.
“Wait a sec,” you say. “On ‘everything?’ I thought this was a 3-product deal.”
Yeah, and a funny thing happened on the way to the impossibility. As in, whenever I had a free moment, I converted a product to Forkbeard. So we have a lot more than I expected at launch. Here’s the complete list:
- Aegir 2F
- Vidar 2F
- Wotan
- Kraken
- Saga 2
- Kara F
- Freya+ F
- Gungnir 2
“But what about Tyr?” someone asks. “What about the rest of the DACs?”
Sigh. We get it. We’re working on it. What I can say is that Stjarna will have Forkbeard from the start…so start thinking about adjusting your gain and loading from a screen using actual human-readable values, rather than squinting at tiny print and bright LEDs.
And more. Soon.
Act 5: The Era of Forkbeard
When we were getting down to the wire, I wondered:
How the heck do we introduce something like this? Multiple new products, multiple old products relaunched, the whole Forkbeard idea, the modularity, the app…
It was too much.
So I decided:
pull back.
The single thing to focus on is this: we’re now in a new era, the era of Forkbeard.
This is literally the biggest change in our company since we started. We’re now a software company—a fully committed, internal-resources-invested, dead-serious-about-this software company.
And everything changes now. If you go to our website, you'll notice a few new things...as in, all of the Forkbearded products are live and available to order or pre-order. The Forkbeard modules themselves are waiting on the official Apple annointment of the app, which we (not me) expect around the end of the month. And yes, we are seriously selling a Gigastack. Yes, you can even add cables to it.
Forkbeard changes everything. This one unique multi-product, modular approach to control, measurement, visualization, and inter-product coordination now gives you unified control—and a unified overview—of all of your systems…and unique inter-product features like Visual Volume. And soon, it will give you even more, as we add functionality, and products, to the system.
Oh yeah, and one last thing: “as we add functionality” is sometimes greeted by crossed arms and skepticism, at least in part because there are butthead companies that think it’s a great idea to charge you monthly for seat heaters.
So here are two more promises:
- Every function added to Forkbeard is free. No initial charge. No monthly charges. Period. For anything we come up with. Forever.
- We don’t take any personal data, we don’t track usage, we don’t sell your data. Period. Again, because we aren’t assholes.
Stunned? I know I am. I didn’t think we’d ever get here. But we are.
All because one fan walked in with a great idea.
