End-to-end latency - the time a computer system needs from an input event until output is displayed - directly influences task difficulty and user experience.
It is therefore an important topic in HCI research.
Different human-computer interfaces require different ways to measure latency as it is influenced by all involved hardware and software components.
However, many approaches to measuring latency rely on professional lab equipment and are therefore hard to replicate.
We propose a method for accurately measuring the end-to-end latency of traditional setups with a button-equipped input device and a display.
To this end, a microcontroller closes the electrical contact of a mouse button to trigger an input event, and captures the screen response via a photo-resistor.
Our approach combines parts of different existing methods to measure latency and only relies on cheap and off-the-shelf components to allow for easy replication.
The latency values measured by our device are very close to those measured with a high-speed smartphone camera (240 Hz). The maximum error is below 2.64 ms - lower than the camera's temporal resolution and the screen refresh periods of high-fps computer displays.
Therefore, our approach allows for repeatedly and reliably measuring end-to-end-latency.