An RTLA-typed articulatory synthesizer is constructed to evaluate the performance of the inverse solution of speech production based on perturbation theory. Vocal-tract area function is derived for a given formant trajectory target by applying the inverse solution. It is then used to control the RTLA synthesizer. Formant mimic synthesis and formant copy synthesis are implemented to validate the effectiveness of the method for both artificially specified formant trajectory targets and estimated formant traces of vowel-to-vowel transitions. The output quality of the synthetic sounds are found to be very good.