
Yaroslav Tretyakov, Mikhail Yakovlev and Yuriy Zhayvoronokwhich were former competitors in the Imagine Cup 2009 part of the team SoundStreamer s represented Russia in the Embedded Development finals. Their team’s project ExoStream, uses the Windows Audio Session API to allow users to stream multiple sessions of music simultaneously via Wi-Fi to multiple speakers in the home. Each speaker is connected to a “Music Box” which receives wireless audio streams therefore converting wired speakers into wireless speakers, thereby removing the need to unplug and replug audio cables when switching between computers. The Music Box opens up several new possibilities for speakers such as sharing a speaker between two computers at the same time. ExoStream also has a neat feature of allowing music to “follow” you by automatically adjusting the volume on every speaker installed in the house based on the laptop’s Bluetooth signal strength so that music only plays loudest at where you are.
The team programmed the advanced audio functionality into Windows Embedded CE 6.0 R2 running on low-cost hardware, and achieved top 20s in the worldwide finals. Yaroslav discusses with me about the project.
Welcome here. First, let us know a little bit about your background.
Hello!
In brief, we’re just students, united by common enthusiasm. We did quite a lot of programming prior to IC09, but never had any unique experience in it. I guess we joined the competition like most of the teams: of pure interest and desire to do something big. Not that we actually hoped for something, really.
But then we found one thing that made us different from other competitors: it was the idea. Once we thought of the “wireless speakers” thing, we knew that should be our entry to the competition. The idea was unique indeed to Imagine Cup, it was almost cheating: not helping the humanity by our project allowed us to design a device that was too “shiny” to be trashed in first rounds. So, I guess not solving any millennium development goals was the main feature of our project.
ExoStream was really “out of the eBox”. It was very useful (we are still using it with pleasure), had a great business plan attached, and, of course, a sexy presentation with a killing demo. We fitted perfectly in all of judging criteria, except one – the main one.
I’d say ExoStream was our ticket to IC09 finals, a project we were destined to lose exactly there.
The “following sound” feature in ExoStream is pretty interesting. Was that an accidental discovery or did your team have that idea all the way from the start?
You named it perfectly – it was 100% accidental, a revolutionary feature we accidently bumped into while designing the automatic discovery subsystem. I remember me and Mike discussing the problem of a laptop and eBox not seeing each other because of the low Wi-Fi signal, and then one of us saying “Hey, we can MEASURE the signal and do nice things using that information”. The rest was purely technical, including the migration from Wi-Fi to Bluetooth.
But surprisingly the FS feature turned out to be not as great as it was shown in our presentation. The main reason for that is the instability and great noises in Bluetooth signal. The music can go up and down and even disappear for a time, while you just stand still with your mobile device. So we eventually disabled FS in the last version of the product (the one we ship to customers).
It seems like you’ve done some usability research for this project. Tell us about it.
Indeed, it was all about usability. We had to satisfy the customer, i.e. the listener. The usability research was not too complex – we simply based on a lot of feedback, including our own. I, personally, listened to music via ExoStream about 5 hours a day.
It is my strong belief that the best way to make a usable product is to use it personally.
What was unique about developing software for embedded operating systems for you?
Apart from the fact that it was our debut with embedded systems, the main difficulty was efficiency (not usability, no!). Playing music means handling millions of packets per second. In the environment of low-capacity embedded system that was a tough problem. The first working version played music pretty much like a jammed audio tape. We had to rewrite it carefully, and almost completely.
How do you see ExoStream changing users’ music listening behavior?
The idea is very simple: we show that audio playback devices and audio rendering devices basically have nothing in common and should not be associated with each other. That gives a lot of freedom.
What gadgets do you like and why?
Those, that make help save time and make life easier.
Do you think there will be a day where wired speakers will become a thing of the past?
Yes, it is 100% real, especially regarding the growth of wireless interface chips production. All that is needed is a nice standard for media transfer. Hopefully, standards are inevitable in massively developing industry sectors.
What do you think is the next big thing in digital media?
Forecasts in progress are almost always wrong, but in my opinion the next steps of media evolution will be the ability to stream any content wirelessly between any kinds of devices AND on-the fly smart media conversion (removing words from a song, adding images to video stream, etc…)
What plans do you have for the future?
Mostly those are business plans. Even though the ExoStream units are expensive (about $200) many people still want to listen to music from the speakers while lying on the sofa. We managed to sell a number of units and currently work on developing the nice new business.
End of Interview
Thanks Yaroslav for your time!
You can learn more about team SoundStreamer s’ project at the Imagine Cup website.
Recent Comments