In a series of previous experiments, it has been shown that when required to produce the back rounded vowel /u/ with a lip-tube perturbation that prevents lip rounding, speakers reached the auditory goal associated with /u/ by altering their tongue position. In the present study, the importance of sensory feedback was further investigated by combining lip-tube and auditory feedback perturbations. Fifteen adult francophone speakers produced 5 blocks of ten /u/ tokens under various conditions (following and preceding a baseline condition). First, tokens were produced when the speakers had a lip-tube in place (predicted to increase F1 and F2) as well as a real-time auditory feedback perturbation (designed to cancel the acoustic effects of the lip-tube, that is, to decrease F2 and F1). Next, only the lip-tube perturbation was applied, with and without white noise. Formant values for each condition were extracted at the vowel midpoint. Although most speakers produced a larger compensatory response with the lip-tube alone condition than with the combined lip-tube and auditory perturbation condition, all speakers produced altered formants in the latter condition. This suggests that speakers try to reduce auditory and somatosensory feedback errors during speech production, with a speaker-specific weighting of each sensory modality.