This paper presents a copy synthesis method devoted to the control of the Klatt synthesizer. Our method allows speech stimuli to be constructed very easily from natural sounds. We chose the parallel branch of the Klatt synthesizer, which allows us to control formant amplitudes. The analysis strategy of a speech signal is straightforward: the fundamental frequency is calculated so that voiced regions are known and the frication energy is set to the value of the spectral energy above 4000 Hz. After formants have been tracked, the amplitudes of the resonators are measured on a spectrum obtained by an algorithm derived from cepstral smoothing called "true envelope". This algorithm has the advantage of approximating harmonics very accurately. The values of the most important acoustic parameters of the Klatt synthesizer are provided by these measurements. Stimuli which have been created by means of this method have a timbre close to that of natural speech. This copy synthesis method is incorporated in our software for speech research called "Snorri". Therefore, the user has at his disposal a versatile tool for creating stimuli in the context of the Klatt synthesizer.