Instrumental measurement of voice is often viewed as providing objective measurements of voice parameters that can be repeated in an orderly manner. However, there are many factors that influence whether or not, for example, one can objectively compare one set of results obtained on one occasion with those obtained on another occasion. There may have been faults in the equipment. Perhaps a recording microphone was not positioned at exactly the same distance from the speaker’s mouth on each occasion. Perhaps different computers were used, each with a different sound card which may have slightly different response times for sound processing, and so on. Perhaps the client wore different clothes on each occasion – perhaps one set was softer and quieter, the other a stiff cotton that rustled as the client moved involuntarily, thereby adding interfering background noise.
As with any assessment, there is likely to be what is referred to as a reactive effect. This is sometimes known as the observer’s paradox. In essence, the presence of an observer/assessor will influence a person’s performance: he or she may feel tense; they may feel that the assessor is expecting him or her to speak in a certain manner/style, and so on. As a result, the person being assessed may consciously or unconsciously adopt an unusual way of speaking/using their voice. Consequently, the data obtained may not be representative of the client’s typical vocal behavior as the client subjectively reacts to the assessment procedure.
Moreover, as with any instrumental measure, for the data to be useful, it must be interpreted by a human. Thus, the experience, training and skill of the therapist in interpreting the figures introduce an arguably unavoidable subjective element.
The key to increasing the objectivity of instrumental measurements is, of course, to hold as many variables constant as possible, e.g. use a head-mounted microphone to keep the distance from the mouth to microphone as constant as possible; use the same computer/recording device on each occasion (including the connecting cables, microphone, amplifiers, sound cards, and so on); make all recordings in the same room (at the same desk, in the same position, and so on).
If one is using instrumental measurements for research purposes then it is essential to keep all variables constant. Often a sound-proof recording booth is used to create a sterile acoustic environment that minimizes environmental noise. For most everyday purposes, however, this rigid control of variables is not possible neither is it necessarily essential. This is because instrumental measurement of voice is simply one of several tools that assist in the diagnosis of voice disorders. It is not the only tool, and the assessment of voice requires the consideration of a variety of different data types (drawn from laryngoscopic examination, perceptual assessment, case history and instrumental measures) by a suitably trained, experienced and skilled speech-language pathologist/speech therapist (voice therapist).
Also, an important use of instrumental measurement of voice is monitoring a client’s progress over time. Even if it is not possible to compare my measurements against measurements made in another clinic elsewhere in the world (because we are each attempting to control variables differently), I can use the measures to compare each of my client’s results one with another. This allows me to document outcomes and progress over time.
Wherever possible I can compare my findings with so-called normative data. Normative data represent the normal or average scores obtained by people who are alike (e.g. comparing men with men, women with women). Unfortunately, normative data is not always available for each measurable voice parameter. However, Table 1 sets out some norms for adults when assessed using the Praat software.
|jitter:||<= 1.040 %|
|shimmer:||<= 3.810 %|
|adult males||adult females|
|mean pitch:||128 Hz||225 Hz|
|minimum pitch:||85 Hz||155 Hz|
|maximum pitch:||196 Hz||334 Hz|
Table 1. Suggested Normative Data for Praat Measurements
As an example, Figure 1 shows the pitch contour of a recording of me saying, “My name is Graham Williamson and I live in Billingham.”
The length of this recording was 3.646 seconds. You can see from the left-hand axis that my pitch never rises above 200 Hz and only just dips below 100 Hz, towards the end of the utterance after about 3.0 seconds (actually when I start to say, “Billingham”). The Praat data for the solo connected speech measures for this utterance are given in Table 2.
|Mean pitch:||125 Hz|
|Standard deviation:||19 Hz|
|Minimum pitch:||89 Hz|
|Maximum pitch:||188 Hz|
|Pitch range:||99 Hz|
Table 2. Praat Data for Solo Connected Speech Task
By comparing these data with the normative data for Praat measurements, you can see that (on this occasion) my pitch, as assessed on this functional task, was within normal limits. This suggest that, using this measure, I do not present with any laryngeal pathology that is affecting my pitch.
For most everyday purposes voice therapists will seek to control as many variables as possible when conducting instrumental evaluations, cautiously compare their findings with available normative data, and use the results to establish a baseline measurement against which client progress can be monitored over time.