Since the introduction of total laryngectomy for laryngeal cancer in 1873 it has become a popular treatment for carcinoma of the larynx. But all patients who have received total laryngectomy have significant disability in their psychological and social lives. To solve these problems, various procedures for voice rehabilitation after total laryngectomy have been developed. The methods for voice restoration is conventional esophageal speech and the tracheoesophageal shunt speech are general. These two methods vibrate the same site(pharyngoesophageal segment), but the air sources are quite different. This difference leads to various different acoustic features between the esophageal and the tracheoesophageal voice. To evaluate which method is more accurate and efficient for communication, we set up 3 groups. Group I consisted of Amatsu tracheoesophageal shunt speakers who use tracheoesophageal shunt voice. Group II is composed of esophageal speakers who use their esophageal voice for primary communication. The authors have analyzed the acoustic parameters group I, II speakers using the Leipzig's voice rehabilitation grading system. The authors also compared the acoustic parameters of normal laryngeal speakers for control. The authors used the data of 16 normal laryngeal speakers as a control. The results obtained were as follows : 1) The fundamental frequency was significantly higher in the group I than in the group II. 2) The intensity had a higher value in group I, followed by group II. 3) The maximum phonation time of group I was longer than that of group II. 4) In the number of counting/breath, group I had more than group II. 5) In the number of syllables/breath, group I had more than group II. 6) In the number of word/breath, group I had more than group II. 7) In the number of word/minute, group I had more than group II. 8) The intelligibility of group I was higher than that of group II. 9) The parameters of the normal control group was higher than the other groups, but fundamental frequency and intensity were lower than in group I. In conclusion, the tracheoesophageal voice, compared with esophageal voice, has some inconvenience in that the patients must occlude tracheostoma with their finger during phonation. But it can produce nearly normal phasing without interruption and a louder voice. Thus the Amatsu tracheoesophageal shunt voice is more familiar and effective in communication to that of esophageal voice, and its also has additional merits in its high success rate, quickes learning possibilitis, and rapid voice restoration. For these reasons, the Amatsu tracheoesophageal shunt is the recommandable method for voice rehabilitation after total laryngectomy.