An evaluation of tonal components in noise using a semantic differential approach yields several perceptual and connotative factors. This study investigates the effect of culture on these factors with the aid of equivalent listening tests carried out in Japan (n=20), France (n=23), and Germany (n=20). The data's equivalence level is determined by a bias analysis. This analysis gives insight in the cross-cultural validity of the scales used for sound character determination. Three factors were extracted by factor analysis in all cultural subsamples: pleasant, metallic, and power. By employing appropriate target rotations of the factor spaces, the rotated factors were compared and they yield high similarities between the different cultural subsamples. To check cross-cultural differences in means, an item bias analysis was conducted. The a priori assumption of unbiased scales is rejected; the differences obtained are partially linked to bias effects. Acoustical sound descriptors were additionally tested for the semantic dimensions. The high agreement in judgments between the different cultural subsamples contrast the moderate success of the signal parameters to describe the dimensions.