Wouldn’t you want to know if your audience is actually interested – or just being courteous? Well, we can help you out! As part of the PrintoCent’s PRINSE’20 seminar (Oulu, 29.–30.1 2020), VTT demonstrated a way to measure and visualize audience applause, during a speech – in real time. Derived from actual hand movements rather than audio, the data does not only show if the audience is clapping, but also how intensely. This provides insight into whether the reaction is genuine enthusiasm, or merely a polite gesture, sometimes also referred to as a “golf clap”.
With over 60 industry speakers, 95 companies, with 280 participants from 24 countries and offering tens of interactive demos at a demo bazaar, the PRINSE’20 is a significant event organized around the technologies of printed intelligence on a global scale. As such, it is also an excellent field test for the potential of ideas and technologies in development. We claim that ‘Smart Applause’ was one of the most attractive demonstrations in the seminar. Conversations withparticipants during the event confirmed that the interest in the demo and sensor itself was high, as can be seen from the fact that out of 193 provided sensors, only 73 were returned – a clear indication of the interest in VTT’s capabilities in both sensor tech and data analytics.
Technical description of the demo
The core idea of the demonstration was to measure and visualize the audience applause during the speeches in real time. Applause data were gathered by FlexDot – sensors, designed by VTT. These were provided to the visitors at the time of registration, with a request to attach the sensor on their wrist. The sensor itself is only c. 2 millimetres thick with a surface area of about 10×25 millimetres and can be loaded to run onboard algorithms – for our demo case it was an algorithm to recognise hand clapping from the sensor’s accelerometer data. In addition, the sensor measured skin temperature and transferred the data from the sensor to the cloud for visualization purposes through a built-in Bluetooth radio interface. As an output, the Smart Applause Demo visualized the intensity of applause on a real-time graph displayed on the big screens in the seminar rooms.
As for the algorithm, we measure the accelerometer Root mean square (RMS) value from the sensor, as we find it to be the most descriptive quantity of hand clapping intensity that the accelerometer can provide. The RMS value of the instantaneous values in a certain time duration relates to the power of the wave. For the period of time considered, this quantity gives an indication of the time-averaged power in the signal due to purely oscillatory acceleration sources.
Field test experiences and brief data analysis
For us as researchers, the event was also a field test. We were especially interested in connectivity, power consumption, ease of operation, algorithm design and data visualization. The demo gave us a lot of feedback concerning all of the topics. For example, use of the Bluetooth radio interface is quite challenging at events like this, as the system was bombarded with almost a hundred BT messages every second from different devices in a seminar room (apparently visitors’ wearables, smart phones, laptops, etc.).
And we wouldn’t be data scientists unless we’d also gather some anonymous behavioural data – so we did! The biggest applause measured during the event topped 982RMS, whereas an average intensity was 409RMS. With regard to skin temperature, the data analysis reveals that the highest temperature was 36.5 degrees Celsius, whereas mean temperature was around 33 degrees during the seminar day: A relatively cool audience! Taken as a whole, we were able to gather total 215 periods of applause in both seminar rooms, which is significantly more than the total number of speeches (46) during the 29thof January. This means the audience not only gave several rounds of applause to every speaker, but many seminar visitors also tested the ‘Smart Applause Demo’ system during the breaks.
All in all we got highly valuable feedback for our future research and development, and want to warmly thank all seminarparticipants and especially all of those who participated in the Smart Applause Demo – a big hand for all of you!
Samuli Heinonen, on behalf of the demo team
+358 40 068 8914
+358 40 584 9754
Printocent | VTT, Kaitoväylä 1, 90571 Oulu, Finland