In speech synthesis, the filtering effect of the vocal tract on the source signal can be imitated. A number of filters, of which frequency and bandwidth can be manipulated, are connected in series, such that the source signal first passes through the formant filter with the lowest frequency. This results in the first formant F1. Then the sound, attenuated by the first filter, passes through the second filter, resulting in F2, and so on. In this way, the spectrum shows peaks at the frequencies of the formant filters, but also shows that the higher in the spectrum, the more the signal is attenuated.