Browsing by Author "Kousa, Markus"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Design, implementation and evaluation of a low-cost, high accuracy feedback latency measurement system
    (2017-12-11) Kousa, Markus; Kämäräinen, Teemu; Perustieteiden korkeakoulu; Ylä-Jääski, Antti
    A touchscreen is a commonly used medium for the interaction between a user and a device. Response to user's action is often indicated visually on the screen after a certain delay. This interface latency is inherent in any computer system. Studies indicate that the latency has a major contribution on how users perceive the interaction with the device. While modern commercial touchscreen devices manifest latencies ranging between 50 ms and 200 ms, research indicates that the user performance for tapping tasks deteriorates at considerably lower levels and users are able to discern the latency as low as 3 ms. In this Thesis we present a novel solution for Android operated mobile devices to expose factors behind the feedback latency of a tap event. We start by reviewing the main components of the Android operating system. Next we describe the internal system elements which partake in the interaction between the user's touch input event and its corresponding visual presentation on the screen of the device. Propelled by the obtained information, we implement an affordable, fully automated system that is capable of collecting both temporal and environmental data. The constructed measurement system provided revealing results. We discovered that most of the feedback latency on a mobile device is accumulated by the internal components which are involved in presenting the visual feedback to the user. We also identified two main user action patterns which impose a huge effect upon system's responsiveness. Firstly, the location of touch is reflected in the amount of feedback latency. Secondly, the interval between two consecutive touch events might cause even unexpected results. Our study demonstrated that the latency can vary a lot between different devices by ranging from no effect on one device to a five-fold difference on another device. The study concludes that, despite the feedback latency is affected by multiple factors, the latency can be measured very precisely with the system that can be built even by an average Joe.