Skip to Content

Department of Linguistics

SPEECH PHYSIOLOGY

Laryngography



Aim

To compare laryngographic data for different vocal qualities.

The laryngograph is a device used to provide information on vocal fold contact. The technique is non-invasive and easy to use. Two electrodes are placed on the throat positioned on each side of the thyroid cartilage. A weak but constant voltage is passed from one electrode to the other allowing the current to fluctuate depending on the contact variations between the vocal folds. When the glottis is closed there is maximal conductance. These variations can be recorded for analysis.

You should note that the terms Laryngograph (Lx) and Electroglottograph (EGG) are effectively equivalent terms (only a very slight semantic difference) and might be used interchangeably in these and related notes.

Data Collection

  1. During several workshops, digital copies have been made of the laryngograph traces from students producing a sustained /a/ vowel using the following voice qualities:
    • modal
    • breathy
    • creaky
  2. These Lx recordings have been processed outside of class times and results will be retained for each speaker for whom there is a satisfactory trace for each of the three collected voice qualities (modal, breathy and creaky).

Results

The raw data collected during the workshops was processed by software especially written for the purpose. This software separated the Lx signals from the accompanying sound data. Speakers were identified who had readable Lx traces for all three voice qualities.

a) Time Scale

The time scale varies for each graph as the Fundamental frequency (F0), and therefore the period, varies for each person PLUS voice quality combination.

b) Voice Quality Types

Modal quality is, for most of these speakers, fairly similar to the waveforms described in several of the references listed below. For some speakers the modal wave shape is similar to breathy quality. This might occur if this person's habitual quality is rather breathy.

Creaky quality Lx waves vary greatly.

Determining Times from the Lx Traces

For each of the Lx traces you will notice that there is a different time scale. Because they are of differing duration the time scale is different.

The time scale (in milliseconds) is displayed at the bottom of each graph.

To calculate time values:

Make measurements with a ruler to determine the numer of milliseconds per millimetre.
Measure the length of the X axis in millimetres, then calculate the number of milliseconds per millimetre by dividing the number of milliseconds by the number of millimetres.
For example, if the total duration of the trace is equivalent to 20 milliseconds and the length of the trace is 200mm, 1mm = 0.1msec.

Measure the distance between the relevant points (specified below) in millimetres to determine the time in msec.

To determine the fundamental frequency (F0) for a single glottal cycle we would use the formula:-

Note that because the glottal period is expressed in milliseconds (thousandths of a second) the formula is 1000 over the period.

Which Participants' Traces will I Analyse?

You should measure the Lx traces for the three voice qualities of the participant whose traces can be found here.

Analysis

For each of the measurements listed below, measure the values for the first three full glottal cycles displayed on each graph for the selected speaker. Your starting point should be where the first full cycle begins its closing movement (point a). Note that all three cycles have already been marked up for you.

a) Voice Qualities

Carry out the following procedure for the speaker and make comparisons between voice quality types. For three consecutive cycles of each utterance the following landmarks have been identified:

  1. Onset of closure
  2. Full closure
  3. Onset of opening
  4. Maximum opening (determined by the 3/7 rule)

Calculate and tabulate:

  • the glottal cycle duration (period) in milliseconds (ms) for each of the first three sequential glottal periods. From this, determine the average glottal cycle period (by dividing by 3).
  • the average fundamental frequency (F0) for these three cycles.

Calculate and tabulate the average percentage of a glottal cycle involved in:

  • closure - onset of closure (a) to full closure (b)
  • opening - onset of opening (c) to maximum opening (d)
  • open time - maximum opening (d) to onset of next closure (a)
  • closed time - full closure (b) to onset of opening (c)

Calculate the following ratios expressed as a percentage:

  • average Open Quotient OQ (ratio of open time [from (d) of one cylle to (a) of the next cycle] / Period)
  • average Speed/Skewness Quotient SQ (ratio of closing time [(a) to (b)] to opening time [(c) to (d)]).

Compare these values between the different voice qualities.

Report Presentation

The assignment should be submitted in the format of a brief scientific report. That is, following a brief literature survey you should very succinctly outline the aims and any hypotheses of this experiment (dot points recommended), then you should present your results concisely using tables and graphs followed by a brief discussion and concluding statement. You don't need to repeat the methodology in your assignment.

  • BACKGROUND

    (Survey of the literature): This section discusses any past work in the area and examines the theoretical background to the topic.

  • RATIONALE

    This section includes the aims and hypotheses based on your literature review.

  • METHOD

    It is not necessary for this report to describe the method.

  • RESULTS

    In this section you will present the results from your analyses and will use appropriate tables and diagrams to illustrate. It is absolutely essential that you provide tables of values upon which you base your analyses.
  • Discussion

    Relate your findings back to the theoretical issues discussed in the literature review.

    • Discuss patterns associated with different voice qualities as described in Laver (1980).
    • Explain how Lx reflects vocal fold contact patterns for the individual voice quality types.
    • Discuss whether the results of your voice traces agree with those of other researchers (e.g. see readings below). It is not a requirement for you to examine all of these references.
    • Discuss cycle to cycle variation.
  • CONCLUSION

    This brief statement summarises the aims and findings. You must also acknowledge the limitations of the research and what you would do to further explore the area in future work.
  • REFERENCES

    Your list of cited resources.

  • APPENDICES

    Any pertinent raw materials are included here.

Your report is expected to be no more than four single pages in length (A4, typed, 12 point, single spaced). You should include relevant graphs and tables in your results section. Penalties will be applied to assignments that are over the page limit.

References:

  1. Abberton, E. R., Howard, D. M. & Fourcin, A. J. (1989) Laryngographic assessment of normal voice: a tutorial. Clinical Linguistics and Phonetics, 3, 281-296.
  2. Baken, R. J. and Orlikoff, R.F. (2000) Clinical Measurement of Speech and Voice, 2nd ed., San Diego : Singular Publishing Group (RC423.B28/2000)
  3. Laver, J. (1980) The Phonetic Description of Voice Quality. Cambridge University Press, Cambridge.
  4. Mooshammer, C. (2010) Acoustic and laryngographic measures of the laryngeal reflexes of linguistic prominence and vocal effort in German, JASA, 127,1047-1058
  5. Winstanley, S. and Wright, H. (1991) Vocal fold contact area patterns in normal speakers: An investigation using the electrolaryngograph interface system. British Journal of Disorders of Communication, 26, 25-39.

You will have already been advised by email on how to obtain these papers.

The marking criteria for this assignment can be found here.