When I returned to Prague from my California trip, I had more motivation than ever to continue with the project, encouraging feedback from YCombinator, and fresh ideas to try. The aim was to create a final product that was small, ergonomic, and lightweight.
At this point, I was working on three facets of the project:
- Developing the hardware
- Improving the ergonomics
- Developing the software
The true innovation and novelty would be what we did with the microphone, the ergonomics, and the software. The electronics of the headset wouldn’t be something unconventional or original; that wasn’t the goal...at least, that wasn’t my plan at that moment.
I started with first principles, as usual. There were three essential elements of the headset to establish: where to put the transducers to get the best sound quality, how to create enough pressure on the transducers to transfer sound, and how to position the microphones.
The transducers had to rest just in front of the ear on the temporal bone. The microphones had to be as close to the mouth as possible, and there had to be enough pressure maintained on the transducers for the bone conduction to work properly. The best way to maintain pressure would be with some kind of hairband, or back loop.
I decided to test an assumption. We 3D scanned my head in various extreme positions to figure out the best shape of the headset to not interfere with movement, and to be so comfortable you could forget you’re wearing it.
Early in 2018, a mutual friend put me in touch with Peter, a BCI enthusiast (and software engineer). Peter had heard from that friend about some script kiddie that was trying to make an all-day interface out of a bone conduction headset, so he gave me a call.
He’d used bone conduction before, and even made his own offline AI assistants. Brain-computer interfaces was the topic of his masters thesis.
We met up for a beer, and everything clicked into place. We talked for hours. We ended up deciding to rent an office together. Peter continued working at his other job, but we wanted a place to meet to have random hackathons and share ideas. You might be familiar with the result of one of our hackathons: Sentien Launcher.
When we attended a BCI meetup in Prague, we explored the potential of BCIs. We moved on from that line of inquiry because creating a brain-computer interface was not at all feasible, for reasons we’ll get into in another post.
Today, Peter is both co-founder and Head of Software—I am glad he gave me the call, we met up and decided to work together.
Putting it together?
We’re up to Spring, 2018, and I was up to my eyeballs in obstacles and unexpected challenges. The entire process was taking a long time. I couldn’t get hold of a chip manufacturer to get Bluetooth chips.
As a stepping stone to prototype the fastest and ship our software, I had the idea that we could make kits and sell those to people who already owned bone conduction headphones.
They could make it fit better, and use the software to change the headphones’ functions. They’d achieve my goal of a seamless interface 100% controlled by the user.
Well, it didn’t take much testing before we realized the kits were actually a hassle to wear. We couldn’t make the hardware in the kit small enough to be comfortable. Once again, the only way to build exactly what we wanted was to make our own hardware from the ground up.
“People who are really serious about software should make their own hardware.” -Alan Kay
The most important objective was to get the ergonomics right. Darek (who was still in university while we worked) and I agreed to work together long-term. We continued making prototypes and I started researching health and ergonomics, as well as manufacturing processes.
By this stage I had been wearing a myriad of different headphones, both bone conduction and others, nearly every day all day. I wanted to learn about any possible health effects and sketched out all potential paths of effect.
From electromagnetic waves, to heat or vibrations. I was not satisfied by seeing that the headsets adhere to international guidelines. I was interested in how these guidelines were developed, what was their history, and whether they were relevant for this new age.
That branch of my research could easily morph into an entire white paper. The most comprehensive summary is a 250 page document done by the European Union on the effects of electromagnetic fields (EMF) on us. I found no negative effects on health from wearing a wireless headset.
I researched and read hundreds of pages on the negative effects of electromagnetic radiation, heat (such as the temperatures produced by the original Google Glass), metal sensitivity, and the possible effects of wearing a Bluetooth headset.
We studied material science to figure out what our headset should be made of. Could we avoid injection molding? I even sent an email to the University of Chemistry and Technology in Prague, asking if I could fund research into UV degradation of 3D printed plastics. They thought I was joking.
We started making decisions about the overall design as well.
Finally, we chose USB-C charging as the most compatible option.
We couldn’t make any of these decisions without also checking the viability of manufacturing based on our conclusions. It’s an iterative process, a sort of chicken-or-egg problem between design and manufacturing. Start with design and get feedback of the manufacturer.
In November, 2018, I attended the Electronica trade fair in Munich. I had a list of components and electronics that we needed to finally build this thing, and I met with anyone and everyone who would talk with me at the trade fair.
We wanted to make sure the microphone quality, sound quality, and reliability of the electronic components would be perfect, so we opted for more expensive components from well-known manufacturers like Qualcomm and Knowles.
Go big or go home.
Winter 2018-2019. Everything was steadily coming to a head. We had planned as far as possible. We had a list of components. We had a list of suppliers. Depending on our next steps, we could either be standing at a dead end or at the edge of a precipice.
There were two options. Should the three of us play it safe and leave it all as a research project and wait for someone else to make the audio interface of our dreams? Or should we go big with our own hardware and software product?
You can guess what we ultimately chose. To be honest, it was not even a question for me. We knew that current solutions were not solving the problems, and we knew we had a new take on the problem and believed in our solution. Soon enough, we had investors.
June 3, 2019. Sentien was officially established.
And in January, we were invited to speak on stage at CES 2020 with Voice Live from CES. CES 2020 was a great opportunity for us, not just to talk about Sentien Audio, but also to meet up with other startup founders and catch up on the latest innovations in the space.
Here we are, in 2020. By now, two more people have joined the team and we work with many contractors worldwide. We’ve caught up to present-day Sentien. I hope you enjoyed reading our story, or better yet, learned something new right along with us.
You’ve gotten an up-close look behind the curtain. You saw the entire process that turned a research project into a company launching its first product this year, including the major design decisions that we made to build the best possible experience.
A lot of thought, work, late nights, and sweat went into Sentien Audio. We hope you’re as excited about it as we are!
Part V? We’ll see you again in Part V when we announce that Sentien Audio has officially shipped. Until then, thanks for reading.
Imrich, and the Sentien team.