“Why can’t headphones be a standalone intelligent device?”
That’s the question that inspired Max Wu back in 2014 as he stared at his non-electrical, phone-dependent tethered headphones. The idea of a smart wearable music-listening device excited the Massachusetts Institute of Technology MBA candidate, who partnered with friend and experienced entrepreneur David Zhu to found the AI-focused audio technology company Inspero.
There was no reference product back then. iWatch, which would evolve the wristwatch into a voice-controlled intelligent device, was still on Apple’s drawing board. And so Inspero started from scratch, and released the world’s first intelligent standalone headphones Vinci in 2016.Vinci 1.0 is a fashionable headset equipped with dual screens. USA today said Vinci’s audio quality “never sounded better”, while Forbes lauded the headphones “magical” voice-driven interface.
A year later, Inspero is about to push the boundaries with its second generation headphones Vinci 2.0, which inherit their predecessor’s features but are much smaller in size and aimed at the fitness community. “Joggers have to tie a phone on the body parts while running outside, which is an annoying burden. This is the problem that Vinci 2.0 can solve,” says Wu.
Vinci 2.0 is a workout-oriented headphone set with a unique pentagon shape that drapes around the neck. Its kickstarter campaign raised US$200,000 over the first 24 hours, 10 times its goal.
This week, Synced sat down with Wu, who is now Inspero’s CTO, and tried out the Vinci headphones.
Vinci headphones include built-in 3G cellular and WiFi connectivity, meaning all of Vinci’s features are available to users no matter whether their phones are nearby. While Vinci 1.0’s dual screens were used for network setup, Vinci 2.0 sets up via a smartphone app.
Inspired by Wu’s previous work as an Intel hardware engineer, Inspero engineers supercharged Vinci with Mediatek quad-core Cortex A7 processors (the same processor found in smartphones and tablets) with 1GB of RAM. The hardware enables features such as voice control, phone calls, calendar alerts, health and workout statistics, and third-party services like Spotify and Amazon Alexa.
Vinci headphones can run seven hours on a charge, which is reasonable for products running high computation tasks. Wu told Synced that the team made significant improvements on Vinci 2.0 to make it small without compromising the chips or battery performance.
Synced was impressed by Vinci’s incredible noise reduction — the smart noise cancelling mode that can silence different environments like home, office, train and airplane. In fact, Vinci’s noise cancelling works so well that Inspero researchers realized users could miss phone calls or messages. A bluetooth connection allows smartphones to relay call or message notifications to the wearer.
Inspero sees music recommendation as the most important Vinci feature. The headphones can recommend music based on the wearer’s heart rate, activity, and listening habits. Researchers also developed a conversational interface, enabling Vinci to respond to natural language requests like “I want some Justin Bieber”, or “Give me classical music”, or even “I’m not in a good mood.”
“Vinci will not help users order a pizza or book a flight to New York City, like Siri or Google assistant. Our focus is on a goal-oriented natural language processing system in the music domain,” says Wu.
Activated by voice control powered by speech recognition technology, users start conversations with the wake phrase “Hi Vinci.”
Fitness freaks are likely to use Vinci 2.0 in noisy environments like a gym or outdoors, and this can stymie voice recognition accuracy. To solve the problem, Inspero developed bone conduction microphones which rest against the user’s neck, receiving voice signals from bones instead of the mouth.
This technology optimizes separation of speech from ambient noise, improving speech recognition and voice services even in extremely noisy environments. Wu says Vinci is the first consumer electronic product equipped with such technology. However, only the Vinci 2.0 Supreme model, which costs US$160 more than the basic Vinci 2.0 Lite (US$89), has bone conduction microphones.
While voice control is efficient, it is not the most practical interface to use when, for example, requesting the next song. This is why Vinci headphones also accept physical commands. Vinci 1.0 users can swipe the right headphone screen to adjust volume or switch songs. On Vinci 2.0, an infrared gesture sensor on the right earbud can recognize hand swipes.
Building such a product wasn’t easy for a startup with just 40 people, and work remains to be done. For example, 3G cellular networks might not be up to speed in a non-WiFi environment, especially for cloud-based AI-driven tasks. Wu told Synced that Vinci will add 4G cellular as early as next year.
Vinci 2.0 has only three third-party services thus far (Spotify, Soundcloud, Amazon Alexa). Vinci plans to add additional services like China’s Xiami Music, which has millions of users. In addition, Google assistant will be added to Vinci 2.0.
Although Vinci 2.0 is significantly smaller and lighter than its predecessor, Wu wants to further reduce the size, and release different types of headphones. Maybe the Vinci 3.0 will arrive looking just like regular headphones?
Wu is convinced Vinci is the answer to the question he asked himself back in 2014, and will open the era of intelligent standalone headphones. “The smartphone is not designed for listening to music, but headphones are. We have smart voice-controlled devices like Echo and iWatch that are hailed by hundreds of millions of users. So why not Vinci?”
Note: Synced tech analyst Alex Chen participated in the interview with Max Wu, and contributed to the above content.
Journalist: Tony Peng | Editor: Michael Sarazen