In the field of speech adaptability to environmental conditions, the intelligibility of hyper auditory and visual speech versus normal speech was investigated. The « Lombard » reflex was used to obtain auditory sequences produced with vocal effort for a set of French /b, d, g, p, t, k/ plosives. Analysis of plosive identifications showed an improvement with the auditory hyper condition in the plosive place of articulation for dentals and velars. Instruction to produce hyper-visual speech provided hyper-articulated labial movements for a set of French /a, i, y/ vowels and /b, v, z, J, l, r/ consonants. Analysis of visual identification scores showed a gain for the hyper-visual condition on the /i/ versus /a/ contrast, but to the detriment of identification of lip movement for consonants - results which were moreover confirmed by the lip shape analysis.