Video-encoders in telemedicine systems need to adjust their coding methods for operation on heterogeneous networks on which bandwidth fluctuates. We evaluated MPEG-4 compressed video-pictures of three trauma patients. We compared the original video-frames with compressed video-frames in terms of the peak signal-to-noise ratio (PSNR). In a qualitative evaluation, three emergency specialists scored the quality of the video-images blind, on a five-point scale (1= bad to 5 = excellent). The PSNR increased as the bit rate increased from 0.2 to 6 Mbit/s. When the bit rate was fixed, in other words at a given network bandwidth, a higher PSNR was obtained at the expense of spatial resolution and frame rate. The video quality was highly affected by the amount of camera shake. Emergency telemedicine systems require a high bit rate, high spatial resolution and a high frame rate to achieve optimum video quality. However, if the bandwidth is limited (i.e. the bit rate is fixed), temporal resolution becomes more important than spatial resolution.