All studies which analysed voice quality after T1 and T2 glottic cancer employed sustained vowels for evaluation. Sustained vowels can often be produced more clearly than expected in a severe degree of hoarseness in running speech. In this investigation voice quality was analysed by measuring the signal-to-noise-ratio (SNRc) in running speech. For evaluation these data were correlated to the scores from an auditive evaluation and from a clinical examination of the larynx considering functional parameters. 39 patients were examined after different therapies (primary radiation, chordectomy, frontolateral partial resection). The grading of the three different methods for characterisation of the voice quality correlated well. Therefore the SNRc is presumed to be a valuable additional method to the psychoacoustic voice evaluation. Comparison of voice quality after different therapies provided no significant difference between radiotherapy and chordectomy. The functional results after frontolateral partial resection were significantly less satisfying as expected.