A device comprising: an input for receiving a signal from a rolling-shutter camera which captures frames by exposing a lines of each frame in sequence, the signal comprising a coded light signal; and a signal processing module arranged to obtain a respective sample of the coded light signal from each of a number of said lines, being a substantially smaller number of the lines than exposed by the camera in each frame. The coded light signal comprises a message having a duration longer than this number of lines, and the message is repeated multiple times with a timing such that a different part of the message is seen by the camera in each of a plurality of different ones of said frames. The signal processing module is configured to time align the different parts of the message from the plurality of different frames, and reassemble the message from the time-aligned parts.