Streamer properties such as their velocity, diameter, intensity and density, can be obtained by analysis of temporal and spatial resolved ICCD imaging. In this paper, experimental results on streamer generation and propagation as a function of several high-voltage pulse and reactor parameters are described. Experiments were performed on a large scale wire–plate reactor in ambient air. The set-up allows for independent variation of the parameters over wide ranges. The minimum gate time of the ICCD camera is 5 ns, allowing for a high temporal resolution. The camera can be triggered with a precision of 1 ns. Both negative and positive polarity pulses are investigated. The most important conclusions are as follows. (1) The streamer velocity ((0.5–2.5) × 106 m s-1) increases if the applied electric field and/or the voltage rise rate is increased. (2) The same is true regarding the velocity ((0.2–1.2) × 105 m s-1) with which the streamer diameter (0.7–3.0 mm) increases during propagation. (3) Typical properties (velocity, diameter, etc) of negative and positive polarity streamers vary less than 25%, especially when the applied electric field is high. (4) As long as the dc bias voltage is below the dc corona onset value it does not have a separate effect on the visual streamer properties. Only the total voltage (peak voltage + dc) is of importance. (5) A simple model was used to determine the electric field in the secondary streamer channel. It was found that in the light emitting part of the secondary streamer the electric field is approximately 21.5 kV cm-1. In the remainder (dark part) of the channel the electric field is around 6.5 kV cm-1. This paper shows mainly experimental findings. Not all observed relations and phenomena could be explained. This is partly caused by the fact that current theoretical and numerical models are not yet able to describe the experimental situation as used during this study.