When we speak about a frequency band, we traditionally describe this in terms of a center frequency and a bandwidth. Bandwidth can be expressed in Hertz, in octaves or in Q. For a filter we usually define bandwidth as the frequency range where the filter has damped no more than 3 dB of the iput signal. Bandwidth expressed in frequency is technically the simplest. For example if we have a band centered at 1000 Hz with a bandwidth of 400Hz, the band covers the area from 800 Hz to 1200 Hz. However, this may not be the most intuitive scale to use, since the perceived width of a frequency band will be related to the center frequency. Expressing bandwidth in octaves means that our previous example has a width of approximately 0.6 octaves (0.585 to be more exact), similarly the same bandwidth expressed as Q equals approximately 2.4. For more info about bandwidth calculations, see http://www.rane.com/note170.html
Frequency range identification
Here we play some sound examples of bandpass filtered white noise.
A bandpass filter will allow a specific frequency band to pass. The sound will have most energy at the center frequency of the filter, gently tapering off upwards and downwards in the spectrum. The bandwidth of the filter is expressed in octaves here, meaning roughly that the bandwidth in frequency is equal to center_frequency*octave (actually it is a little bit less, as it is correctly defined as the lower frequency bound times the octave).
Listen to the sound examples and determine center frequency and bandwidth: