Cue-integration and context effects in speech: Evidence against speaking-rate normalization