本研究測試六十個聽力正常而且聽覺敏感度較好的受試者透過外語用聽輔儀聽取過濾的語言，並且找出在何種特定頻率帶過濾的語言聽起來與正常語言相近，從研究結果可看出每個語音都集中一個或二個最適當頻率帶，國音中鼻音子音比相似英語子音低，其他子音則無明顯的特性。國音中子音和母音在180-4000Hz的頻率範圍即可做最佳的語音接收和辨識。母音中單韻符頻率範圍分布無較明顯的集中趨勢，複韻符的最適當頻率在中心頻率710, 1000, 1400, 2000Hz便能清楚的接收，結合韻符則有更明顯的趨勢，集中於中心頻率1400Hz，也就是1000-2000Hz的頻率範圍。研究所得的結果，將國音依最適當頻率帶由高而低可把母音和子音分成低、中低、中、中高、高五組組成國音的音調模式。 子音 /ㄐ/、/ㄑ/、/ㄓ/、/ㄔ/ 和Peterson的研究相一致，送氣子音比不送氣子音的音調高，其他子音則無此現象，子音中有鼻音比捲舌音低的特性。子音也有擦音 /ㄒ/、/ㄕ/、/?/ 音調排序最高，鼻音 /ㄋ/、/ㄌ/ 音調排列最低的特性。 國音中母音所得最適當頻率帶的結果與英語相近之母音互相比較結果顯示單韻母、複韻母並沒有比英文高或低的特性。比較國音中五個主要母音的音調排列順序結果， /ㄨ/、/ㄚ/、/一/ 可明顯排出由低而高之音調順序。 研究結果顯示國音中基本頻率並不是判斷音調的重要因素，與Harbold（1958）和Guberina所提的觀點一致。國音中 /一/、/ㄨ/、/ㄩ/、/ㄚ/、/ㄛ/、/ㄜ/、/ㄝ/、/ㄦ/ 八個母音的接收在第一、第二、第三共振頻率即可涵蓋接收母音的足夠訊息，此結果也與Miner和Danhauer（1976）所得結果相一致。

　　　　　The purpose of the study was to investigate the optimal octaves for Mandarin phonemes. Sixty college students were randomly selected from those who passed on hearing sensitivity. The subjects listened to filtered phonemes while they adjusted center frequencies of the filtered bands on the SUVAG LINGUA (System Universal Verbo-Tonal Audition). The optimal octave was defined as the band which produced the most similar sound like the unfiltered phoneme. The main findings were summarized as follows: The optimal octave for each phoneme was centered on one or two frequencies. The optimal octave for nasals was lower than the similar one in English. The speech perception and identification for phonemes were focused on the optimal octaves through 180 to 4000 Hz. The optimal octaves for vowels were distributed at a wide range. Diphthongs can be perceived clearly in the ceter frequencies of 710, 1000, 1400, and 2000 Hz. The compound vowels tended to centralize in the center frequency of 1400 Hz. The tonality for voiced phonemes /t/、/t'/、/ts'/ were higher than their voiceless counterparts, but other consonants did not show the same characteristics. Affricates and nasals had the highest and the lowest tonality, respectively. Foundamental frequency was not the crucial factor for tonality judgement. The optimal octaves for vowels could be perceived through the frequency ranges of F1, F2 and F3.