FIELD OF THE INVENTION

This invention relates to a device for recording glottal vibration using a piezoelectric sensor and a microphone, and further, the method of analysing and visualizing the recorded data, ft is a purely non-invasive, low cost device that can be handled by a person with basic training and can be used in research, diagnosis, education etc.
BACKGROUND OF THE INVENTION
Various devices have been developed in the past that can measure glottai vibration.
However, many of them have got limitations to which this invention answers to.
Typically, glottal information is recorded with different invasive and semi-invasive systems like endoscopy, Electro Giotto Graph etc. which requires professional expertise for handling the equipment and carrying out tests. Also cost of these systems are very high.
Endoscopic systems use high speed video camera inside oral cavity to record the movement of glottal folds visually. Major disadvantage of this system is that it requires medical experts to carry out the tests. Also, this interrupts with the normal voice and the cost of equipment and tests are also very high.
In Electro Giotto Graph, (he glottal activity is measured by sending a small amplitude high frequency electrical pulse from one side of the larynges and received on other side. The change in shape of glottal tissues causes a change in impedance. The resultant current change is measured by the device. This signal is filtered and produced as Electro Giotto Graph representing the movement of glottal fold tissues. It is a semi-invasive process but the cost of equipment is very high.
In EP1898786B1, Piezoelectric sensor is used to collect acoustic signals from the chest. Here the point of interest is the vibration of the chest wall. Speech components can
observed in this setup due to resonance inside audio track. This device calibrates the acoustics and separates the Speech sound and cough. However, using this device on the neck region as such will not be able to sense glottal vibrations properly and is therefore not obvious.
In the published paper- "Measuring glottal activity during voiced speech using a tuned electromagnetic resonating collar sensor" (Meas. Sci. Technol. 16 2015 2381-2390), the difference in capacitance in body tissue is used to calculate glottal activity. This has two electrical contacts with subject. Here the target is similar but the approach is totally

different. This approach is similar to Electro Glottograph, where the difference in impedance is used to calculate glottal activity and is less accurate.
In the published paper- "Laryngoaltimeter: A New Ambulatory Device for Laryngeal Height Control, Preliminary Results" (Journal of Voice, Vol-23, No-5, 2009,529-538), the vertical position of Larynx is estimated with a device Laryngoaltimeter which utilizes two microphones and comparing sound intensity level at two heights. Also in the discussion section, it is mentioned that "Laryngeal movement can also be detected by a piezoelectric sensor placed on the cricothyroid region of a subject" by referring to another paper "An Electronic Device Measuring the Frequency of Spontaneous Swallowing: Digital Phagometer" (Dysphagia. 11. 1996). But in both the papers, the area of interest is whole Larynx and the large movements (vertical high and swallowing) are considered for analysis. The vibration pickup from the Neck region is a mix of glottal movement, Resonance from the oral and nasai cavity and other muscle movements in larynx region. Hence, proper recording of the glottal information cannot be achieved.
In US 2010/0256503A1, visualisation of glottal behaviour in different phonation conditions including voiced or unvoiced using diffused illumination of glottis using an external light source and collection of light signal by means of external photo detector. However, it requires some energy(light) to travel through the person and its changes are recorded by the detector which is therefore not truly non-invasive.
Thus there is a need for a device and method to record glottal vibration information in a true non-invasive manner, with a simple external attachment and extract the glottal information by a low cost method.
OBJECT OF THE INVENTION
The principal object of the invention is to provide a low cost device for recording, analysis
and visualization of glottal vibrations.
Another object of the invention is to provide a method of glottal vibration analysis using a
plug & play hardware and computer based software interface.
DRAWINGS
Fig, 3 shows the device attached to She neck of a subject/person(31).
Fig. 2 shows the schematic illustration of the architecture of the device and the method.
Fig. 3 shows the cross-sectional view of the sensor placement in neck of the user.
Fig. 4 shows the effect of touch (noise introduced due to touch) to the sensor.

Fig. 5 shows the image of 3 dimensional view of the sensor enelosure(5). Fig. 6 shows the waveform graph of the signals recorded from a subject.
SUMMARY OF THE INVENTION
The invention relates to a device for recording, analysis and visualization of glottal vibration by using a Neck Band(3), an Electronic Hardware module(14) and a Software Module(16). Further, the invention also relates to the method employed in using the above components to record, analyse and visualize the glottal vibration. The said device uses Piezoelectric Seusor(6) to record glottal vibration from the subject/person(31) and Microphone(i) to record the voice. Here, the Scnsor(6) records vibrations from the Glottis(20) of the subject/person(31) through the Skin(4) through an Enclosure(5) in which it is located. The Enclosure(5) is non-conducting and non-magnetic. The Sensor(6) and the Microphone(1) are located in the Neck Band(3) and which is made of sewing elastic band and being comfortable to wear by the subjcct/pcrson(31). The signals from the Scnsor(6) and the Microphone(l) are processed by the Electronic Hardware module(14) converting them to digital signals that are further processed by the Software Module(16). The said method, after receiving the glottal vibration signals(7) and speech signals(8) simultaneously from a speaking subject/person(31), processes them separately for biasing and signal coupling(9)(10), supplying them to the respective analog filter and amplifier networks(ll)(12) and then converting them to separate digital signals. The digital signals are then transferred to the Software Module(16) located in a computer, through an Interface(15). The Software Module(16) performs storing and displaying of the real-time digital signals in views of- Signal Energy and/or Power spectrum and/or Spectrogram and/or Glottal instances and/or Glottal flow graph and/or Pitch and/or Zero crossing over formats and analysing the digital signals for speech analysis, speech recognition, speech comparison or glottal vibration related process and providing feedback of supplied or generated data to the speaking subject/person(31). Using this invention, tracking of glottal vibration for both speech research and clinical use within a comparatively low cost and less operational complexity conditions can be achieved. Pure non-invasive approach including full electrical isolation between the subject and the instrument is followed here to reduce complexity and operational hazard.

DETAILED DESCRIPTION OF THE INVENTION
The Invention relates to a device that comprises of two major components- a Hardware unit and a Software module(16). Further, the invention relates to the method employed in the said device for glottal vibration analysis using a plug & play hardware and computer based GUI software.
The Hardware Unit comprises of a Neck Band(3) and an Electronic hardware module(14). The Neck Band(3) comprises of at least one Microphone(l) attached to it by a semi-flexible attachment(2) and also an Enclosure(5) within or attached to the said band. The Neck Band(3) is made of sewing type elastic band with a Hook-and-loop fastener. The elastic nature of the band and use of the Hook-and-loop fastener for adjustment of length, provides the said band to comfortably fit around the different neck shapes of the users of different ages.
The Enciosure(5) is made of a non-conducting and non-magnetic material and gets positioned at the level of the Glottis(20) of a subject/person(31), touching the skin(4) above the Glottis(20), when the Neck Band(3) is worn. The EncIosure(5) comprises of at least one Sensor(6) inside it which is a Piezoelectric sensor, including the low cost metallic disc type having a maximum diameter of 30mm. The said enclosure may be made of ABS or similar polymer materials. This configuration eliminates noisy signals(22) that may be induced from the sensor directiy touching the skin whereas sensing through the Enclosure(5), clean signal(2l) can be achieved leading to high accuracy.
On the other hand, the Electronic hardware moduie(l4) is either located inside the enclosure(5), -in the Neck Band(3) or externally and comprising of biasing and signal coupling(9)(10), analog filter and amplifier(11)(12) and two channel analog-to-digital(13) circuits. The output of both the Microphone(l) and the Sensor(6) is received and processed by the said module and thereafter sent to the software module(16) as respective digital signals through Interfacc(l5) that can be wired or wireless chip to chip communication protocols.
The Software module(16) comprises of real time data analysis (18), data visualization(I7) and data storage(19) components and is located in and executed by a computer. The real time data analysis(18) component functions to calculate features of energy of the received signals, frequency domain transforms, zero crossing, spectrograms etc. Next, the data visualization 17) component outputs the data in the form of real time plots to display the signals and the corresponding plots with the calculated features. Also,

the data storage(l9) component stores the recorded data and calculated features in standard formats like wav files, csv files etc. and the plots as image files.
To use the device, first, the Neck Band(3) is worn by the subject/person(31) such that the Enclosure(5) is positioned directly in front of his/her Glottis(20) and is firmly but comfortably secured. On powering up the device and the subject/person(31) speaking, the Hardware module(14) receives glottal vibration signals(7) and speech signals(8) simultaneously from the Piezoelectric Sensor(6) and the Microphone(l) respectively. These signals are then processed separately by the Electronic Hardware Module(14) by biasing and signal coupling. Here, the vibration signals(7) is filtered to attenuate the higher frequency artefacts from the resonance of speech in the oral and nasal cavity and from external sources in the environment allowing only the glottal vibration range to pass through and then is normalised to full scale. Further, the signals are then amplified and then supplied to the respective analog filter and amplifier networks(ll),(l2) and then is converted to separate digital signals. The software module(16) is located in a computer where the said digital signals are worked for speech analysis, speech recognition, speech comparison or glottal vibration related process by the real time data analysis(18) component. They are also displayed in real-time in views of- Signal Energy and/or Power spectrum and/or Spectrogram and/or Glottal instances and/or Glottal flow graph and/or Pitch and/or Zero crossing over formats by the data visualization(7) component. They are also stored by the data storage(19) component. During the above activities, feedback is also provided to the subject/person(31) on the history or real-time data processing. The system may be powered using external battery or may get power from USB port. Special precautions like shielding the connecting wires, use of decoupling capacitors and notch filters etc. can be taken to avoid electromagnetic interference. Along with the data recording and analysis features, the software moduie(16) also has a user organization and management system for managing different users recordings properly.
EXAMPLES
Test recordings of Glottal Vibration (23) and Corresponding Speech (24) from a male
subject is shown in Fig.6 where the signal was recorded for 14.4 seconds and the following
sounds were pronounced /a/ (25),/u/ (26), /m/ (27), /i/ (28), /s/ (29), /ʃ/ (30).
It is visible that for the non-voiced sounds like /s/ (29) or /ʃ/ (30) there is no glottal signal.

5. CLAIMS
1. A Device For Recording, Analysis and Visualization of Glottal Vibration and the
Method Thereof, the device comprising:
a Neck Band(3) having enciosure(5) within or attached to it;
at least one Sensor(6) placed inside the enclosure(5);
at least one Microphone(l) attached to the Neck Band(3) by semi-flexible
attachment(2);
Electronic hardware moduie(14) either located inside the enclosure(5), in the
Neck Band(3) or externally and comprising of biasing and signal coupling(9)(10),
analog filter and amplifier(l 1)(12) and analog-to-digital(13) circuits and
Software module(16) in a computer comprising-
real time data analysis(18) component calculating energy of the signal,
frequency domain transforms, zero crossing, spectrograms;
data visualization(17) component in the form of real time plots displaying the
signals and the corresponding plots with the calculated features and
data storage mechanism(19) storing the recorded data and calculated features
in standard formats like wav files, csv files etc. and the plots as image files,
wherein
the Sensor(6) is a Piezoelectric sensor;
the Enclosure(5) is made of a non-conducting and non-magnetic material
positioned at the level of the Glottis(20) and touching the skin(4) above the
Glottis(20) and
the Band(3) is a sewing type elastic band.
2. A Device For Recording, Analysis and Visualization of Glottal Vibration and the Method Thereof as claimed in Claim1 wherein the maximum diameter of the disc of the Sensor(6) is 30mm.
3. A Device For Recording, Analysis and Visualization of Glottal Vibration and the Method Thereof, the method comprising the steps of:

(a) Receiving glottal vibration signals(7) and speech signals(8) simultaneously from a speaking person(31);
(b) Performing biasing of the sensors (1)(6) and signal coupling to the received glottal vibration signals(7) and speech signals(8) separately and then supplying

them to the analog filter and amplifier networks(1l),(12) and then converting them to separate digital signals;
(c) Storing and displaying the real-time digital signals in views of- Signal Energy and/or Power spectrum and/or Spectrogram and/or Glottal instances and/or Glottal flow graph and/or Pitch and/or Zero crossing over formats and
(d) Analysing the digital signals for speech analysis, speech recognition, speech comparison or glottal vibration related process and providing feedback of supplied or generated data to the speaking person(31);
wherein
in step(a), the glottal vibration is recorded by at least one piezoelectric sensor(6) placed on the neck of the said person positioned above the glottis and touching the skin through a non-conducting material and
in step(b), the vibration signais(7) is filtered to attenuate the higher frequency artefacts from the resonance of speech in the oral and nasal cavity and from external sources in the environment allowing only the glottal vibration range to pass through and then is normalised to full scale.