In this paper, we show that the performance of voice activity detection algorithms (VAD) can be highlydependent on the type of background noise and we introduce a new VAD algorithm that is based onrelative energy measurements in different frequency bands. The obtained experimental results arecompared to the results obtained with two other spectrum-based VADs and it is concluded that a VAD,configured to use around 3 frequency bands can cope best with a large variety of background sounds.