At Thalmic Labs, we’re transforming the way people interact with computers in an entirely new way: by reading the electrical activity in their muscles. To do this, we use EMG (electromyographic) sensors that we’ve developed; similar to the kind used in hospitals for reading your electrocardiogram (ECG – heart check up).
Before the Myo armband, capturing the type of EMG data we use would have required a $10,000 desktop machine that uses special electrodes and conductive gel, plus some seriously big amplifiers. We’ve managed to put that technology inside a wireless armband that costs around $200, and doesn’t require shaving your arm or applying conductive gel.
As a developer making applications for the Myo armband, you may have noticed that our SDK does not provide a view of the “raw” EMG data. We offer “classifier outputs”: data that’s already been processed by algorithms inside the Myo armband. Rather than exposing the signal directly in your muscles, the classifier output shows you what that signal means: a fist, a wave motion, spread fingers, and so on. Similar to how a mouse outputs the relative x and y position of the cursor, not the raw values of the optical sensor inside. The Myo armband provides the output command.
As of December we’re going to make this new data stream available to developers in experimental form free of charge: they’ll have access to a stream of raw EMG data straight from the Myo armband.
There are a number of reasons we made classifier outputs available and raw EMG hasn’t been available:
- Battery life. Processing the raw data on-board using classifiers that are optimized for power efficiency results in significantly better battery life than streaming that amount of data through Bluetooth. Transmitting raw data efficiently over this signal has been a huge technical hurdle, but overcoming it means amazing new uses for the Myo armband. Prosthetics, fatigue and hydration monitoring, sleep state monitoring, identity authentication based on unique muscle signals… We’ve heard compelling arguments from developers for why it is worth the trade-off to have access to this data.
- User Experience. Working with EMG signals is hard. Building an application that works for you and your friends is pretty straightforward, but building something that will work for everyone is another question entirely. The signals that are produced when different people make similar gestures can be wildly different. Different amounts of hair, fatty tissue, and sweat can impact what the signals look like, and this is compounded by the fact that the Myo armband can be worn on either arm and in any orientation. It’s incredibly difficult to transform a person’s hand into a controller, and we need to make sure that the applications in the Myo Market yield great experiences for users.
That being said, we’ve been listening to your feedback. Many of you in the developer community feel strongly that working with raw EMG data straight from the device is an important part of unlocking its potential, so in December we will add experimental functionality to make raw EMG data available to developers.
How you use the raw EMG data is completely up to you. Developers have shared their plans for all sorts of EMG-based applications in fitness, health, and music creation to name a few. But opening up raw EMG data to our developers doesn’t necessarily mean you’ll see it in the Myo Market any time soon.
This is a screenshot of two raw EMG signals from two different people. They are making exactly the same gesture. This complexity is compounded when you consider that the signals made for your ‘fist’ may be nearly identical to the signals produced from someone else tapping their index finger to their thumb.
As our developer community continues to grow, we’ll be able to work through solving these challenges and adding more and more functionality (and gestures) with your support. Data is king, and if we get more of it, we’ll be better able to map your muscle activations to useful outputs to control your technology.
In the meantime, we probably won’t be allowing applications in the Myo Market that employ custom gestures. While they may work on the individual developing the application, it’s unlikely they will work across all of our users, and that’s not the user experience we want. We’re making this data available to encourage people to play with it, and see what they can come up with. We’re hoping to make exceptions for high-quality, compelling uses of the raw EMG data, but for the most part this feature is experimental for now.
Right now, we’re all about the experiment. Everything we’re doing at Thalmic Labs we’re doing for the first time. You early developers have been a part of this journey with us since day one, and now we’re giving you a new tool to play with. What comes from it remains to be seen, but you’ve never failed to surprise and amaze us. We don’t think this will be any exception.
Thanks for supporting us so far and for the great feedback. We’re listening, and working tirelessly to build the best experience for you. We’re a team of hackers/developers ourselves, and love nothing more than creating new technologies and exploring the limits of human-computer-interaction.
— Steve, Matt and Aaron