In car noise environments, speech quality is severely degraded and speech recognition systems fail to produce high recognition rates. Highpass filtering and Kalman filtering with a whitening process are applied to car noise corrupted speech signals and improvement both in speech quality and in speech recognition is achieved.