A speech signal pitch analyser based on the wavelet transform is proposed. The pitch period estimation is realized both with a time and a frequency domain method. We describe part of the analyser using the accurate resolution of the wavelet transform in the low frequencies. The use of progressive wavelets allows to extract from the complex wavelet coefficients information given with a phase and a modulus, like Fourier coefficients in the short-time fourier transform. The instantaneous frequencies calculated from the low frequency channels of the wavelet transform ( 50 to 500 Hz ), give an estimation of the fundamental frequency of the speech signal or one of its harmonics during the voiced segments. The paper closes with some results wich confirm the validity of the method.