# US 2004/0083093 A1 · Method of measuring nasality by means of a frequency ratio

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2004/0083093 A1
- Document ID: US20040083093A1
- Status: Published application
- Kind: A1
- Application number: 10/280,877
- Filed: 2002-10-25
- Published: 2004-04-29
- Earliest claimed: 2002-10-25
- Inventors: Guo-She Lee; Terry B. J. Kuo
- Web page: https://patentgrep.com/patent/US20040083093A1

## CPC

- G10L 25/00: Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 (muting semiconductor-based amplifiers when some special characteristics of a signal are sensed by a speech detector, e.g. sensing when no signal is present, H03G3/34)
- G10L 17/26: Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

A method for measuring nasality senses a sound, amplifies the sensed sound signal, converts the sound signal to a digital sound signal, transforms the digital time domain sound signal to a frequency domain with a Fast Fourier Transformation (FFT), determines a cut frequency (fcut) calculates a low frequency power and a high frequency power of each window, calculates an acoustic low/high ratio (ALHR) and calculates an average acoustic low/high ratio (ALHRave). With such a method, the ALHRave reflects a person's nasality and nasal airway status, and the method is conveniently performed.

## Claims

### Claim 1

A method for measuring nasality comprising steps of:
sensing a sound enunciated for several seconds;
amplifying the sensed sound signal;
converting the amplified sound signal to a digital sound signal;
transforming the digital time domain sound signal to a frequency domain with a Fast Fourier Transformation (FFT), wherein, the time domain sound signal is divided into several calculation windows for FFT, and each window has multiple calculation points;
determining a cut frequency (fcut) to divide the sound frequency spectrum into a low frequency band and a high frequency band;
calculating a low frequency power (LFP) and a high frequency power (HFP) of each window, wherein the LFP is a sum of the power at frequencies below the cut frequency and LFP is a sum of the power at frequencies above the cut frequency;
calculating an acoustic low/high ratio (ALHR) of each window by dividing LFP by HFP; and
calculating an average acoustic low/high ratio (ALHRave) by averaging the ALHR of all of the windows.

### Claim 2 (depends on 1)

The method as claimed in claim 1, wherein the sampling rate of the sound signal should be greater than 16 kHz, and the-frequency resolution for Fast Fourire Transformation should be less than 10 Hz.

### Claim 3 (depends on 1)

The method as claimed in claim 1, wherein the sound is generated by a test subject enunciating a nasal consonant-vowel tone for several seconds.

### Claim 4 (depends on 1)

The method as claimed in claim 1 further comprising generating and sensing a background noise prior to sensing the sound to determine a start time of the sound signal.

### Claim 5 (depends on 4)

The method as claimed in claim 4, wherein the start of the sound signal is when the root mean square of the sound signal is two times larger than the root mean square of the background noise signal.

### Claim 6 (depends on 5)

The method as claimed in claim 5, wherein the sound signal is analyzed from 0.5 second from the start of the sound signal to the 3.5 seconds from the start of the sound signal.

### Claim 7 (depends on 1)

The method as claimed in claim 1, wherein the low frequency band is from 65 Hz to fcut; and
the high frequency band is from fcut to 8000 Hz.

### Claim 8 (depends on 7)

The method as claimed in claim 7, wherein the cut frequency is determined by square root times of a fundamental frequency.

### Claim 9 (depends on 1)

The method as claimed in claim 1, wherein the step of sensing the sound signal uses a high sensitivity microphone.
