Department of Speech-Language Pathology, University of Toronto, Toronto, Canada.
Department of Rehabilitation and Special Education, University of Cologne, Cologne, Germany.
Clin Linguist Phon. 2022 Mar 4;36(2-3):292-300. doi: 10.1080/02699206.2021.1974094. Epub 2021 Sep 23.
Nasometry is used to assess nasality in speech but it is unclear whether overly frequent recalibration of the instrument adds measurement errors. The goal of the present research study was to describe the effect of the Nasometer 6450 calibration on the nasalance scores of repeated recordings. In a first experiment, the Nasometer calibration values stored in the computer's registry were manipulated to investigate the impact on nasalance scores. In the second experiment, a set of pre-recorded speech samples was re-recorded 40 times with a Nasometer 6450 in 4 different calibration regimens: Short-term repeated recordings without (R1) and with recalibration (R2C), and long-term repeated recordings over 10 days without (R3) and with recalibration (R4C). The first experiment showed that, compared to a calibration value of 1.0, a value of 0.9 resulted in nasalance scores that were on average 3 points lower while a calibration value of 1.1 resulted in scores that were 0.5 points higher. The results of the second experiment showed test-retest differences of less than 2 nasalance points for 91% of the data for a non-nasal stimulus. For a nasal stimulus, 91% of data were within 5 points for R3 and R4C. The results suggested that frequent recalibration of the Nasometer may slightly increase test-retest differences of nasalance scores. An alternative procedure for verifying microphone balance without recalibration is suggested.
鼻音计用于评估语音中的鼻音，但目前尚不清楚仪器是否过于频繁地重新校准会增加测量误差。本研究的目的是描述 Nasometer 6450 校准对重复记录的鼻声分析值的影响。在第一个实验中，对存储在计算机注册表中的 Nasometer 校准值进行了操作，以研究其对鼻声分析值的影响。在第二个实验中，使用 Nasometer 6450 以 4 种不同的校准方案（R1 无和 R2C 有、R3 无和 R4C 有）对一组预录制的语音样本进行了 40 次重复录制。第一个实验表明，与 1.0 的校准值相比，0.9 的校准值导致平均鼻声分析值低 3 个点，而 1.1 的校准值导致分数高 0.5 个点。第二个实验的结果表明，对于非鼻音刺激，91%的数据测试-重测差异小于 2 个鼻音点。对于鼻音刺激，91%的数据在 R3 和 R4C 时在 5 点以内。结果表明，频繁地重新校准 Nasometer 可能会略微增加鼻声分析值的测试-重测差异。建议采用一种无需重新校准即可验证麦克风平衡的替代方法。
