USER MANUAL

Nasalance System
Model NAS-1 Multi User

May 2016

Glottal Enterprises, Inc.

1201 E. Fayette Street
Syracuse, NY 13210
315-422-1213
www.glottal.com

©2016 Glottal Enterprises, Inc. All rights reserved.

Nasalance System and NAS-1 are trademarks of Glottal Enterprises, Inc.

*This product is for research and teaching purposes only. It is not a medical device. It is not intended
to be used in the diagnosis, cure, treatment, mitigation or prevention of disease and it is not intended to
affect the structure or function of the body.
Nasalance System Manual May 2016 2

Page Table of Contents

3 A. General Information
A1 Introduction
A2 System Requirements
A3 Software Installation
A4 Nasalance System Hardware
A5 Assembling and Using the OroNasal Mask and Handle
A6 Assembling and Using the Partition-based Separator Handle
A7 Screen Rings
A8 Computation of Nasalance and Compensation
A9 Computation of Oralance
A10 Running the Nasalance System Software
A11 Input Level Adjustment
A12 Setting the Audio Playback Path

9 B. Nasalance
B1 Nasalance Flow Screen
B2 Options – both screens
B3 Axis Setting – both screens
B4 Target Range – flow screen only
B5 User Selection – both screens
B6 Threshold Level – both screens
B7 Playback Button and Playback Audio – both screens
B8 Average Nasalance Computation and Settings – both screens
B9 Compare-to-Model Screen
B10 Recording on the Compare-to-Model Screen
B11 Text Box
B12 Saving and Moving Segments Recorded on the Compare-to-Model Screen
B13 Loading Previously Recorded Segments on the Compare-to-Model Screen
B14 Highlighted Region
B15 Oralance option
B16 USB Switch Box

14 C. Checking Out Your System

C1 Calibration Verification
C2a Checking the Input Connection and Recording Level – Mask Input
C2b Checking the Input Connection and Recording Level – Separator Input
C2 Checking System Operation with Continuous Speech
C3 Setting the Threshold Level

20 Technical Support

21 Limited Warranty
 Nasalance System Manual May 2016 3

A. General Information
A1. Introduction
Welcome to the Nasalance System for the measurement of vowel nasalization. The Nasalance System
was designed to make it a meaningful tool for indicating graphically and numerically the degree of
velopharyngeal closure during voiced vowels and vowel-like consonants.

The Nasalance System uses two screens to accomplish all the needed functions. The software allows the
user to record sounds or phrases, play them back, hear them, graphically represent them and, with both
sight and sound, compare them to model productions or previous subject efforts.

After explaining how to install the Nasalance System on your computer, this manual describes the
features of each display screen.

After running the installation program, the Nasalance System program will appear on your Start-
Programs menu just like other Windows® compatible software. Left-click (click the left button of the
mouse) on Nasalance System to start the program.

A2. System Requirements

To use the Nasalance System software, you will need the following user-provided hardware:
• Desktop or laptop computer with Microsoft Windows XP operating system or newer, including
those Apple Macintosh computers that use an Intel microprocessor and can run Windows.
• Desktop or laptop computer with at least one free USB input jack.
• Minimum 50MB of free hard drive space, 100 MB recommended to accommodate saved speech
samples.
• Minimum 256MB of RAM, with 512MB or larger recommended, and 1GB for Windows7.
• External speakers or headphones connected to the computer’s audio system.
• CD-ROM drive for installing the program. However, a USB flash memory with 500MB available
can also be used (as with Netbook computers having no CD-ROM drive) by copying the program
onto the flash drive.

A3. Software Installation

1. It is recommended that you to close all applications before installing new software,
2. Insert the Nasality System CD into the CD-ROM drive. The Installation Wizard should start
automatically.
3. If the installation wizard does not start automatically, click on Start in the taskbar and select My
Computer. Double-click on the CD drive that contains the Nasality System CD
4. Click on Next in the window that appears. Follow the on-screen instructions.
5. If you already have a version of the Nasality System installed and insert a Nasality System CD, you
will be asked if you want to repair, modify, or remove the previous version. Click on the appropriate
button and click OK.

A4. Nasalance System Hardware

A Nasalance System contains the following parts:
o a dual chamber OroNasal mask with screen rings and a mask handle and electronics pack
w/USB cable,
and/or
o a partition-based separator handle with electronics and USB cable - either a single-user
handle or a multi-user handle allowing replaceable partitions.
o a set of three partitions, narrow, medium and wide, for a multi-user system.
 Nasalance System Manual May 2016 4

The OroNasal mask

The mask (normal Adult size) has eight large holes, containing eight removable stainless steel screen-
rings, one for each hole. These screen-rings use a specific grade of stainless steel cloth chosen to
maximize the sound transparency of the mask and minimize the resistance to outside noise.

NOTE: To better assure proper mask calibration, it is recommended that you use gloves when
handling the screen rings during insertion or removal. This will help prevent the need for frequent
cleaning.

The mask has a molded partition on the inside to separate oral and nasal airflows for nasalance recording.
The mask also has a number of potential access ports on the outer surface, in the form of cylindrical
indentations. These access ports should be left unchanged. However, the access ports can be punctured to
allow other transducers or probes to penetrate the mask. Such penetration should be airtight. When a
punctured port is not being used, it can be sealed with a convenient removable plastic plugs that can be
provided upon request.

The mask handle and electronics pack

The handle allows the user to hold the mask comfortably in position. The electronics pack, a rectangular
box, is mounted on the handle, and includes:

• Two matched pressure-sensing elements mounted in a mask coupler

• Sensor activation circuitry
• Amplification and A/D conversion circuitry, with a USB output cable

The partition-based separator handle

The partition-style handle allows the user to obtain nasalance readings without the inconvenience of a
mask. In use, the partition plate sits against the upper lip. For the multi-user-model, the partition can be
chosen as one of the three easily interchangeable partitions that fits most closely to the maxillary anatomy
of the user.

NOTE: When using the partition-style handle, for consistent nasalance readings, it should be kept
at approximately the same angle to the face. For these reasons, we recommend the partition
handle primarily for certain linguistic applications, when the precise value of nasalance is not
important, or for situations in which a mask is inconvenient. The mask-style handle may be
preferable for obtaining data for publication.

A5. Assembling and Using the OroNasal Mask and Mask Handle
1. Put the USB plug on the lead cord from the mask handle into a USB input jack of your computer.
2. If the mask is not already assembled, snap one screen ring into each of the large holes in the mask,
making sure it is well-seated in the groove around the hole, preferably using gloves.
3. To insert the electronics pack in the mask, first hold the mask in one hand, grasping it by the nasal
separator on the inside surface. Hold the handle with your other hand. Align the two coupler plugs on
the electronics pack with the two small holes in the nose of the mask, oriented so that the computer
lead cord exits the underside of the electronics pack. Press the electronics pack into the mask until the
case touches the mask.
 Nasalance System Manual May 2016 5

4. Have the user hold the handle with his/her hand touching the back of the electronics pack, and place
the mask in position on the face, with the nasal separator resting on his/her upper lip. The mask should
touch the face firmly, especially along the upper lip, but without any great pressure.
5. Have the user talk into the mask. His/her jaw and mouth should be able to move freely - it is not
necessary to push the mask tightly against the face.

A6. Assembling and Using the Partition-based Separator Handle

The partition-based separator handle
The partition-style handle allows the user to obtain nasalance readings without the inconvenience of a
mask. In use, the partition plate is held against the upper lip. As in the photo below, a nasalance partition
handle should be held at an angle to the face that is about 10 degrees from perpendicular. This angle is
not critical but it is the angle at which the units are calibrated.

For the multi-user-model partition handle, the partition can be chosen as one of the three easily
interchangeable partitions in different sizes.

To use the multi-user version of the separator-style handle, determine which of the three partitions
supplied best fits the maxillary arch (curvature of the upper lip). Slide this partition fully back in the
handle and finger-tighten the knurled screw on top of the handle enough to hold the partition in place.
(Do not over-tighten.)

The more economical single–user partition handle has a fixed partition plate.

A7. Screen Rings

OroNasal masks and partitions for the separator handle are shipped clean, but not sterile. The masks and
partitions are reusable, and are easy to clean.

It is best to reserve a different mask for each user, and store it under sanitary conditions between uses. A
single handle and electronics pack can be used interchangeably with all masks. Extra masks can be each
conveniently enclosed in a plastic refrigerator bag. The masks and rings should be completely dry before
storing in a closed plastic bag. Partitions can be cleaned with alcohol, or if the application is not critical,
use any standard dish-washing method.
 Nasalance System Manual May 2016 6

Thorough cleaning of the mask

1. Remove the screen rings by gently bending the mask and popping out each ring.
2. Wearing gloves, wash the mask and screen rings separately in warm water with a mild liquid dish
detergent. Handle the parts gently. (Never wash parts in a dishwasher.)
3. Rinse the mask and screen rings with tap water and allow them to dry completely before storing.

Checking the condition of the screen-ring mesh

To perform properly, the stainless steel mesh of the screen rings must be free of skin oils, and largely
free of dust and other particles. For this reason, it is advisable to occasionally check the mesh with a
magnifying glass or jeweler's loupe (X5 or X10 recommended) while holding the ring up to a light
source. If you see any contaminants other than widely distributed dust particles (it is not possible to
totally clear the mesh, and this is not necessary), wash the rings as instructed above. You can also brush
the mesh gently with a toothbrush while the rings are submersed in water. The following figure shows
what you may see for screen rings in various states.

We have also found convenient for cleaning and drying screen rings the cans of compressed gas meant
for removing dust from electronic or photographic equipment, such as the product sold at Radio Shack,
Velocity dust remover spray, Model 64-4351, Catalog #64-4351.

Note: When measuring Nasalance, it is the ratio of oral to nasal flow resistance that is important, and
not the precise values of resistance. For this reason, some similar small amounts of dirt on all screen
rings will not affect the readings significantly and can be tolerated before cleaning is needed.

A8. Computation of Nasalance and Compensation for Acoustic Crossover

The Nasalance System features the calculation of nasalance from voiced speech, as recorded with a dual
chamber wire-screen pneumotachograph mask [1], or microphones above and below a partition held
against the upper lip. It features the calculation of nasalance from the amplitudes of the nasal and oral
voice components at or near the voice fundamental frequency, Fo. The resulting ‘Fo Nasalance’ is less
dependent on the particular vowel sound being spoken than is a nasalance ratio derived from sound
pressure energy in the first formant (the F1 Nasalance).

The percent nasalance is defined as %N = 100Anasal/( Anasal + Aoral), where Anasal and Aoral
are the amplitudes of the appropriately filtered nasal and oral airflows or pressures.
 Nasalance System Manual May 2016 7

Compensation for acoustic crossover - In any system for measuring nasalance, whether mask-
based or partition-based, there is some crossover of acoustic energy between the oral and nasal
channels. If the means of separation is well-designed, this crossover is generally assumed to be less
than about 10% of the signal, with the precise value depending on the geometry of the means of
separation and the directivity characteristics of the oral and nasal pressure sensing elements. The result
is that the measured nasalance for a vowel can never be less than the crossover percentage, even if there
is a complete velopharyngeal closure, and the measured nasalance for nasal consonants always falls far
short of 100%, the theoretical value. In addition, the differentiation between non-nasal and nasal values
is reduced.

The Glottal Enterprises system for measuring nasalance now contains a unique feature for reducing the
effects of acoustic crossover (patent pending). Before nasalance is computed, a small percentage of the
measured nasal energy is subtracted from the measured value, and the same percentage of the measured
oral energy is subtracted from the nasal value. The result is an improved differentiation between
nasalized and non-nasalized vowels, and a nasalance reading that reflects more closely the area of the
velopharyngeal opening.

The compensation feature can be turned off by an onscreen button, either before making a recording or
after recording is stopped. Turning off the compensation may be desirable if nasalance values are to be
compared to those from older systems for measuring nasalance that do not have this feature.

References:
[1] Rothenberg, M. “Measurement of Airflow in Speech,” J. of Speech and Hearing Res. 20, 155-176
(March 1977).

A9. Computation of Oralance

%Oralance is defined as [100 - %Nasalance]. The default setting is set to be Nasalance and Oralance
can be selected by means of the radio button at lower left on the screen, if this option is available.

A10. Running the Nasalance System Software

Click on Start in the taskbar and select All Programs. From the submenu that appears, click on
Nasalance System in the Glottal folder. When the screen for the NVS system appears, click on the box
labeled Nasalance at the top of the screen. The Nasalance system will open in the Flow mode, but either
the Flow or the Compare-to-Model mode can be selected at any time using the selection boxes at top-
left.

A11. Input Level Adjustment

Each of the computer screens in the Nasalance System has bar meters (lower right) for monitoring the
signal level coming from the OroNasal mask or partition separator. On each screen there are two such
meters, one for the oral signal and one for the nasal. Though the Record Level is not critical, the level
should be such that that the signal level indication is usually in the center third of the bar meter scale
when speech is being recorded. Both bar meters should be checked, the oral meter by vocalizing a non-
nasal vowel, as by saying /ababa …/, and the nasal meter by saying a nasal consonant, as a prolonged
/m/. (You will find that the nasal signal will generally be less than the oral signal and so the nasal
energy may not quite reach the middle third of the bar when the oral energy registers in the middle third.
This is normal.)
 Nasalance System Manual May 2016 8

The setting of the Record Level is performed automatically by the program. However, an on-screen
option is available for changing the gain by plus or minus 6 dB or 10dB, depending on the model. For
example, the plus 6 dB or 10 dB option may be needed for people with a low voice level, while the
minus 6 dB or 10 dB setting may be needed with loud singing.

If the level is too low, the result will be signal amplitudes that will too often be less than the threshold
setting, and, therefore, the nasalance will not display for some voice segments. A readjustment of the
Threshold setting to a lower number (see below) may rectify this situation. Note that this can be done
after the recording is made, or even when replaying a previous recording.

If the signal level is too high, one or both of the level meters will register in the red region, and the
audio system will overload. This will result in an error in the nasalance reading and, in addition, the
recorded sound may sound distorted. (The nasalance error will generally be such as to move the reading
closer to 50%.) However, the metering criteria are such that brief flashes of the lowest red bar on the
meter are usually acceptable.

A12 Setting the Audio Playback Path

With the Nasality System hardware plugged in to a USB jack (mask handle or partition handle), the
playback may not be channeled properly through a computer’s speakers or headphones. If this is the
case, it is possible to switch playback to the computer’s speakers using the Adjust Audio Properties
window for your version of Windows. In that window, if the Playback Device is set to iMic (the
identifier for all Nasalance Visualization System Hardware), you should change the Playback Device to
the system in the computer (in this case C-Media Wave Device). But do not change the Record
functionality from its iMic setting. The window shown here is for one version of Windows XP, but
other versions of Windows will have a similar window..
 Nasalance System Manual May 2016 9

B. Nasalance

B1 Nasalance Flow-mode Screen

The Nasalance System has two screens, Nasalance Flow-mode and Compare to Model that can be
chosen by clicking on one of the two tabs at the upper left part of each screen, on either screen, at the
lower right part of the screen.

When starting the Nasalance System, the Nasalance Flow-mode screen will open. The Flow-mode
Screen offers continuous real-time analysis and display. The chart on this screen displays the last 5 or 10
seconds (user selectable, with a default setting of 5 seconds) of nasalance values recorded after the Start
button is clicked. After starting the recording, the Start button changes to a Stop button and can be
clicked to stop recording and capture the last 5 or 10 seconds of the recording. Five seconds is the default
interval, but ten seconds can be selected by clicking on Options at the lower left.

The computer Space Bar offers a convenient alternative for starting and stopping recording. Just tap the
bar to toggle between the two.
 Nasalance System Manual May 2016 10

For brevity, the title of each of the following sections will indicate which screen,
Flow-mode or Compare-to-Model, the section applies to.
B2 Options – both screens
Clicking on Options on any screen in the Nasalance System brings up a sub-screen that allows the user to
change the time scale on the current chart display to 10 seconds from its default value of 5 seconds (as
mentioned under B-1 above) or change the audio sample rate from its default value of 11,025 samples per
second. The default value of sample rate results in an adequate audio quality for speech (approximately
AM radio quality) with reasonable computation times and memory requirements, and is recommended.
(A rate of 8,000 would shorten computation times and reduce memory requirements somewhat, while a
rate of 22,050 would improve audio slightly but double computation time and memory requirements.)

B3 Axis Setting – both screens

Either Percent Nasalance or Oralance can be selected, with Percent Nasalance (%N) the default setting.

B4 Target Range – flow-mode screen only

A Target Range can be marked on the chart by setting the slider at left to the center of the desired range
and the up/down arrows below the slider for the size of the range. Whenever recording is stopped, the
time that the user stayed within the target range is shown in the box above the chart, computed over the
entire 5 or 10 second duration of the chart, or the selected time interval, if one has been selected.

B5 User Selection – both screens

The User selection optimizes the computation for the expected pitch range of the user, and should be
chosen to match the gender and age category of the user, but is not critical for age selection. You can test
this for yourself by recording a sentence then switching between user categories.

B6 Threshold Level – both screens

The Threshold Level sets the minimum value of sound amplitude for which computation and display of
Nasalance will occur. A Nasalance value is computed only for those time intervals during which the
sound level exceeds this threshold. The threshold level can be varied by clicking on the up/down arrows.

The threshold comparison is made just prior to computing Nasalance. If both the filtered nasal signal and
the filtered oral signal are below the threshold, the value of Nasalance is set to zero and a yellow bar is
placed near the zero line on the display for that time instant.

The Threshold Level is set in terms of a percentage of a specified reference level. While recording is in
progress, whether on the continuous record function of the Flow-mode screen or the fixed length record
function of the Compare-to-Model screen, the reference level is taken as one-half the maximum amplitude
that can be recorded. After a recording is complete, the reference level is changed to the maximum
amplitude attained in the recorded segment.

The optimum value for the threshold percentage will vary slightly from person to person, but can be
expected to vary between about 20 to 30 percent. It is best to analyze a few signals using different
threshold values to determine which is the best value for your subject. For each subject, you can record a
sentence on the flow-mode screen, and then change the threshold to see the result. When the threshold is
selected properly, there should be no %N indication during silent intervals or unvoiced speech sounds,
and no gaps in the display during continuously voiced speech (with the possible exception of the closed
interval of a voiced stop, such as /b/ in “aba”, during which the voice signal is very weak.).
 Nasalance System Manual May 2016 11

If there are appreciable indications of %N during clearly silent segments, increase the threshold by a few
steps. On the other hand, if there are gaps in the display during a continually voiced vowel or a nasal
consonant, try decreasing the threshold by a few steps.
A recording of a recommended test sentence, as spoken carefully by an adult male speaker with no
known speech defects, is shown in Figures 1 and 2 on page 17 and 18. The charts in Figure 1 show this
recording analyzed at six threshold levels, bracketing the one or two that would be good choices for this
speaker, as indicated on the right of each chart.

In Figure 2, three of the threshold levels of Figure 1 are repeated with annotation. The annotation shows
in more detail the features of the Nasalance chart to observe when setting the Threshold level. The use of
a sentence such as that in Figures 1 and 2 in setting the Threshold Level is described in more detail under
Calibration.

B7 Playback Button and Playback Audio – both screens

When a 5 or 10 second segment has been captured (and recording stopped), it can be played back
auditorily by clicking on the Playback button. The Playback Audio selection at bottom right will
determine which channel is heard.

B8 Average Nasalance Computation and Settings – both screens

Average Nasalance (shown above the chart in the Flow mode or at left in the Compare-to-Model mode),
is computed over the entire 5 or 10 second window in the chart if no time window has been set on the
screen, or over a time window that has been highlighted on the chart using the mouse. (See B-14,
Highlighted Region.) Average Nasalance is displayed in the units on the screen.

The Average Nasalance computation does not include any interval during which the signal is below
threshold. In addition, a check box in the Nasalance Settings selection area at lower left allows the user
to also exclude from the computation those intervals during which there is likely to have been a nasal
consonant (%N greater than 60%). By eliminating silent intervals and nasal consonants from the
computation, the operator can use test sentences with a natural variety of consonants.
 Nasalance System Manual May 2016 12

B9 Compare-to-Model Screen

The screen labeled Compare to Model can be selected by clicking on that that tab at the upper left. This
screen differs from the Flow-mode Screen in that:

1. Only a fixed time interval is recorded, either 5 or 10 seconds. Recording stops automatically after that
interval.
2. Recorded segments can be saved.
3. There are two nasalance charts on the screen, enabling comparison of current and previous
productions. The top or Model chart can load a previously recorded segment.

The settings on the Compare-to-Model screen are largely the same as those for the Flow-mode screen
described above. Those settings and functions that differ are described in the following sections.

B10 Recording on the Compare-to-Model Screen

Recording on the Compare-to-Model screen is possible only on the lower chart. The Record interval is
either 5 or 10 seconds, as chosen under Options, with the default value being 5 seconds.

B11 Text Box

A horizontal box between the two chart boxes allows the user to add text, such as the printed version of
the word, phrase or sentence spoken in the Model chart. To type in text, place the cursor in the box, as in
a word-processing application, and begin to type. The text will be saved with the model when the model
is saved and will reappear whenever that model is loaded (see B-13).
 Nasalance System Manual May 2016 13

B12 Saving and Moving Segments Recorded on the Compare-to-Model Screen

A segment recorded on the lower chart of the Compare-to Model screen can be either moved to the upper
or Model chart or saved to a file in the computer, by clicking on the appropriate button to the right of the
lower chart. A Save button on the upper chart allows a segment moved from the lower chart to also be
saved.

B13 Loading Previously Recorded Segments on the Compare-to-Model Screen

To simplify saving and retrieving recorded segments, when the program is installed, a section of the
computer’s memory is relegated to saving nasalance files. Files from this section can be accessed and
shown on the upper, or Model, chart by clicking on the Load Model button. The User Settings (User and
Threshold) and the time scale for charts on the upper screen will be automatically set to the value used
when the segment was saved. However, Settings for Average Nasalance and Playback choice can be all
set as desired.

B14 Highlighted Region

To select any desired interval of a recorded time-nasalance display, place the mouse cursor at one
endpoint of the desired segment on the chart and press and hold down the left mouse button. While
holding down the left mouse button, move the mouse along the chart to form a highlighting rectangle
enclosing the chosen segment. To revert back to full display, click anywhere on the chart.

B15 Oralance Option

Clicking on the Oralance button at the lower left will cause the screen to display Oralance instead of
Nasalance.

Percent Oralance is defined as (100 – %Nasalance).

A sample screen comparing nasalance and oralance for the same recording is show here.
 Nasalance System Manual May 2016 14

B16 USB Switch Box

Glottal Enterprises now offers a USB switch box that can be used to switch between two users or
between a nasalance handle and a nasal emission handle, without unplugging and plugging USB
connectors. (Without the switch box, it is not possible to plug in two handles and choose between them
in software.)

It is suggested that, for convenience, if both a nasalance mask handle and nasal emission mask handle
are being used, all the screen rings be left in when switching a mask from nasalance to nasal emission,
even though the oral screen rings are not needed in the nasal emission application.

If a change is made in a recording parameter, such as the recording gain, while recording is ongoing,
recording must be stopped and restarted to register the change. This can be most easily accomplished
by tapping twice on the space bar of the computer’s keyboard.

C. Checking out Your System

C1 Calibration Verification
Mask calibration
Calibration (balancing the two channels) is performed at the factory and is generally not done in the
field. However, the following simple calibration check will show up most hardware, software and
connector problems, using the mask calibration verification coupling tube supplied with the unit.

1. With the system connected and operating, remove the mask from the mask handle.
2. Place the system in the Nasalance Flow mode, with the chart showing Percent Nasalance.
3. Holding the mask handle vertical, place one end of a short tube, as described below, over the two
nipples coming from the mask handle. This coupling tube is a simple device for keeping the sound
pressure at the two nipples the same. It comes with your system or a replacement is available from
Glottal Enterprises. If the tube you have comes with notches on one end, it is the end without
notches that is to go against the case. (The notches are only needed for checking a separator handle
calibration.)
4. Start recording by clicking on the Record button.
5. With your face anywhere from 1” to 3” from the tube opening (distance not critical), vocalize a loud
prolonged “oo” sound (as in “moon”) or alternatively an “m” sound. Do not move the coupling tube
during the recording.
6. Make sure that the volume level bar on the screen is in the proper operating range (2 to 5 segments
of the level indicator). If not, speak at a different level or readjust the record level, as by using the
+10 dB setting (see below).

The system should indicate a %Nasalance of between 47 % and 53%, indicating that the sound levels
recorded in the two channels match to within 5%. A value outside this range could indicate a defective
microphone or other equipment problem. Return the system to Glottal Enterprises for service.

If you need a replacement tube you do not need to get it from Glottal Enterprises, the tube dimensions
should be:
• Inside diameter between 1 3/8” and 1½”, or large enough to clear the nipples but small enough to
not extend beyond the case when in place.
• Length at least 2 ½”, with a length of 3” to 8” preferable.
PVC plastic plumbing couplings with these dimensions, including a 6” plastic sink drain “tailpipe”, are
available at most hardware or plumbing supply stores at a nominal cost.
 Nasalance System Manual May 2016 15

Separator calibration
The separator-style handle is factory calibrated and usually requires no recalibration in the field.
However, a calibration coupling tube that is supplied notched at one end can be used to make a rough
check The following figure shows the tube in place over the partition. The notches keep the tube
roughly centered over the partition. If held in place by hand, keep the hand far from the partition, so
that the sound reaching each microphone is the same.

Place the system in the Nasalance Flow mode, with the chart showing Percent Nasalance. With the face
anywhere from 1” to 2” from the tube opening, vocalize a loud prolonged “oo” sound (as in “moon”) or
alternatively an “m” sound. Do not move the coupling tube during the recording. Speak loud enough
so that the volume level bar on the screen is in the proper operating range (2 to 5 segments of the level
indicator).

The recorded nasalance should be approximately 50%. If readings are consistently outside the range
45% to 55%, the unit may require service.

C2a Checking the Input Connection and Recording Level – Mask handle input
Set the Nasalance System to the Flow-mode screen. Place the on the face of a person with no known
velopharyngeal deficit, and click on Record. When a held nasal consonant such as [m] is produced, the
Nasal bar meter (the left-most bar) should show 2 to 5 segments lit, with little or no indication on the
Oral bar meter.

When a held, non-nasalized open vowel such as [a] is produced, the Oral bar meter should show 4 to 6
segments lit, with little or no indication on the Nasal bar meter. (Note that the oral meter indication for
the vowel will usually be stronger than the nasal meter indication for the nasal consonant.) Even for a
speaker with what can be considered normal articulation, a held vowel may be inadvertently pronounced
 Nasalance System Manual May 2016 16

with some degree of nasality. To ensure that the vowel being pronounced is not nasalized, use
repetitions of the slowly spoken syllables /ba/ or /pa/ (bababa … or papapa….) instead of a constant
held vowel. In a repeated syllable sequence, the velum rises automatically for the proper production of
the stop consonants and remains closed for all or most of the intermediate vowels, if they are not
prolonged excessively. For a quantitative check, see also Mask calibration, in Section C1.

C2b Checking the Input Connection and Recording Level – Separator handle input
Set the Nasalance System to the Flow-mode page. Place the handle’s separator in place on the upper lip
of a person with no known velopharyngeal deficit. The separator plate should be chosen so as to fit well
to the upper lip of the subject. The separator plate should be angled slightly up as it meets the lip,
perhaps 10 or 15 degrees. Now click on Record (or tap the Space bar).

When a held nasal consonant such as [m] is produced, the Nasal bar meter (the bar on the left) should
show 2 to 5 segments lit, with little or no indication on the Oral bar meter.

When a held, non-nasalized open vowel such as [a] is produced, the Oral bar meter should show 4 to 6
segments lit, with little or no indication on the Nasal bar meter. (Note that the oral meter indication for
the vowel will usually be stronger than the nasal indication for the nasal consonant.) To ensure that the
vowel being pronounced is not nasalized, you can use repetitions of the syllable /ba/ or /pa/ (bababa …
or papapa….) instead of a constant held vowel. In a repeated syllable sequence, the velum rises
automatically for the proper production of the stop consonants and remains closed for all or most of the
intermediate vowels, if they are not prolonged excessively

C3 Checking System Operation with Continuous Speech

To check the system operation with continuous speech, change to the Flow-mode screen. With the
mask in place, record a test sentence. Use a sentence with a variety of consonants. For English
speakers, we recommend the sentence “Yes, maybe I will and maybe I won’t.” as used for Figures 1
and 2 below, but you may think of others equally convenient. We like it because it is also convenient
for setting or checking the Threshold Level, as explained below.

C4 Setting the Threshold Level

Figures 1 and 2 show the effect of varying the threshold for the recorded sentence “Yes, maybe I will
and maybe I won’t,” as recorded on the Compare-to-Model screen of the Nasalance system.

For the record level setting used for this recording, the Threshold setting of chart D appears to us to be
optimal, with settings for both charts C and D being good and acceptable. This judgment is explained
as follows.

The letters of the printed version of the sentence at the top of each chart in Figure 1 show approximately
at which points the various sounds of the sentence are being produced. At the setting of chart D, the
nasal consonants and are shown clearly as vertical blue bars. In this speaker’s dialect, the presence of
the N in “won’t” is conveyed primarily by the nasalized vowel, as in the nasal consonants in French, and
this is shown clearly in the display. The first vowel in ‘maybe’ is shown properly as somewhat
nasalized (nasalance values of larger than about 0.15), due to coarticulation effects (vowel adjoining a
nasal consonant). Other vowels are shown to be produced without nasality
(percent nasalance values below about 10%).

The solid yellow horizontal bar at the bottom if the display D shows correctly the periods during which
the signal level is below the criterion level, namely:
• before the sentence starts,
• during the /s/,
Nasalance System Manual May 2016 17

• during the /b/ closures,

• during the closure for the final /t/, including the noise burst produced at the release of the /t/
closure.

Note that the nasalance displays in Figures 1 and 2 show nasalance levels in green
instead of the red color used in current versions of the System.

FIGURE 1 – A recording of the test sentence “Yes, maybe I will and maybe I won’t”, analyzed using six
different settings of the Threshold Level, to illustrate proper and improper settings. Level differences
between steps were approximately 5 dB. For this speaker, the Threshold settings for C and D are acceptable.

To see the errors produced by an incorrect threshold setting, compare the correct pattern of chart D with
the patterns produced at the threshold settings for chart A (clearly too low) and F (clearly too high). To
expedite this comparison, these three charts are reproduced and annotated in Figure 2.

In chart A, for example, the low level voice murmurs coming from the closed oral cavity during the
periods of articulatory closure for the /b/ consonants are incorrectly registered by the system as showing
 Nasalance System Manual May 2016 18

sharp peaks in nasalance (marked by numbers 3 and 4). These peaks are artifacts of these very weak
sounds reaching the mask transducers, and have no significance. There are also some slight traces of
artifactual spikes during each /b/ at the setting in chart B, in which the threshold was raised about 5 dB.
Increasing the threshold level another 5dB, for chart C essentially eliminates these artifactual spikes.

With too high a threshold setting the sound from some of the nasal consonants and some vowel
segments falls below the threshold, and so the nasalance for these consonant and vowel segments is not
registered by the display.

For a given computer and audio level setting, the Threshold Level need not be changed between
subjects from day-to-day. However, an occasional check by recording the above sentence is quick and
simple. The correct setting is the same for both the Flow-mode and Compare-to-Model screens.
Changing masks should also have a negligible effect on the correct setting.

Since the correct threshold level is affected somewhat by the audio record level setting, you should
recheck the threshold setting if the audio level setting is changed.

FIGURE 2 – Three of the threshold settings used in Figure 1, from a recording of

the test sentence “Yes, maybe I will and maybe I won’t” by an adult male
speaker.

Some explanatory comments for Figure 2:

Referring to the CENTER chart, for which the threshold has been set properly, we see:
• When the threshold is set properly, the unvoiced continuant consonant (/s/) and the stop consonants
(/b/ and /t/) are indicated by a yellow bar, since the voice energy is below threshold.
 Nasalance System Manual May 2016 19

• The fully articulated nasal consonants in which the velum is fully raised (the /m/ and /mn/) are
indicated by a blue vertical bar, since the Nasalance Ratio increases to values over 1.5, the criterion
used here for a complete velar closure.
• The /n/ in “won’t is shown to be pronounced a nasalized vowel in this case (as is often used in
French), with only a brief velar closure (the narrow blue vertical line) before the labiodental closure
for the /t/ occurs/. The nasalized vowel is identified by the double headed arrow.

An important factor in accepting the above judgments is that they are consistent with the phonetic
structure of the sentence for normal English.

Referring to the TOP chart, for which the threshold is set too low, we see:
There are numerous intervals for which there are indications of Nasalance that are not consistent with
the sentence being spoken. Some of the more obvious of such aberrant intervals are identified by the
numbers 1 through 5. When the signal is very low, but accepted by a threshold that is too low, the
resulting value of Nasalance will be erratic and inconsistent, and this is shown in the Nasalance
computations at the locations identified by the numbers.

Referring to the BOTTOM chart, for which the threshold is set too high, we see:
There are a number of inconsistent gaps in the Nasalance values, such as those identified by the vertical
arrows. The Nasalance values that are plotted, however, are consistent with those of the (correct) center
chart. This is to be expected, since they are derived from oral and nasal signals that were strong enough
to exceed the high threshold.

*****
Nasalance System Manual May 2016 20

Technical Support
If you are unable to install the software, or have problems running this software, please contact Glottal
Enterprises Technical Support (e-mail is recommended).
Technical Support
Glottal Enterprises, Inc.
1201 E. Fayette Street
Syracuse, NY 13210 USA
E-mail: techsupport@glottal.com
Phone: 315-422-1213 Fax: 315-422-1216

Please provide the following information when you contact Technical Support:
1. Title of application you are running (NEM System software)
2. Other programs that may be running on your computer while the NEM System is running. (It is
recommended that other programs be closed while running the NEM System.)
3. Type of computer.
4. Version of Windows
5. Amount of installed RAM.
6. Description of the problem and when it occurs.
 Nasalance System Manual May 2016 21

Limited Warranty

Glottal Enterprises, Inc. warrants, subject to the conditions and restrictions contained below, that each product
manufactured and sold by Glottal Enterprises, Inc. shall be free from defects in materials and workmanship
for a period of one (1) year from the date of purchase from Glottal Enterprises, Inc.
Should a product manufactured by Glottal Enterprises, Inc. fail to conform to this written warranty, Glottal
Enterprises, Inc. will, at its option, either repair or replace the product at no cost to the purchaser or, at Glottal
Enterprises, Inc. option, refund the purchase price of the product.

The following shall not be considered defects in materials or workmanship:

Damage caused to the product by others or by any cause beyond the control of Glottal Enterprises, Inc.,
including but not limited to damage caused by misuse, abuse, mishandling, or by fire, flood, earthquake,
storm, or other acts of nature.

Claims under this Limited Warranty must be made in writing to Glottal Enterprise, Inc. within 1 year from the
time of purchase to the following address:

Glottal Enteprises Inc.

1201 East Fayette Street, Suite #15
Syracuse, NY 13210
USA

Claims should detail the nature of the problem and the name, address and telephone number of the person
making the warranty claim. Glottal Enterprises, Inc. reserves the right and must be given at least thirty (30)
days to examine the claimed defective product and to act on the warranty claim. This limited warranty
extends only to the original purchaser of the Glottal Enterprises, Inc. product and in any event no
compensation under this warranty shall exceed the original invoice price at the time of purchase from
Glottal Enterprises, Inc.

Glottal Enterprises, Inc.’s liability is limited to repairing or replacing defective products, or at its sole option,
refunding the purchase price. There is no implied warranty of merchantability. There are no
Warranties expressed or implied which extend beyond the conditions and remedies described herein.