vText

A. Introduction

vText is a simple voice recognition component designed for .Net and Java developers. You can configure it to either function as voice password or simple voice command but it is primarily designed to function as voice password or speaker recognition. It has recording, comparison and filter functions. The recognition period is either 5 or 10 seconds to compare the source signal and the input signal. The comparison returns high % score when its' frequency spectra match closely on the frequency domain, it is designed to allow you to input your voice signal at anytime within the 5s or 10s.

To function as simple voice command, you need to pre-record each user's voice first for each voice command. vText has parameters such as lower limit and upper limit which flexibly allow you to configure it to function either as voice password or simple voice command.  

When you should use vText ?

1. When you need further protection for your data.

i. Document signature e.q. cheque etc, you can use the viewWave method to plot the wave and print, the size and location can be changed at preview.

ii. Approve payment/transaction >= e.g. $100,000USD

2. When keyboard entry is not convenient. 

3. When system needs additional option for command entry for strategic operation.

4. When you want to customize and integrate voice password or simple voice command into your system. vText provides interface to basic elements needed for voice recognition through various computer languages such as Java and .Net.  

 

vText is suitable for use in any situation or system which requires additional protection for data. Normal keyboard password or smart card can be copied or stolen to make illegal access, with vText, these possiblilities are reduced because the voice biometric of each individual is different.. The voice commands used should have more than 2 words and have different words to allow easier differentiation.

 

The usage guide and download for vTer vText software component is readily available at:

1. http://www.basic-signalprocessing.com/downloadVText.htm (Java) ;or

2. http://www.basic-signalprocessing.com/downloadVTextNet.htm (.Net).

 

For installation on hardware device, notice should be given a few months earlier because the deployment process for each hardware is different.

B. The theory behind vText

vText was developed based on signal processing theory using Nyquist sampling theorem and frequency domain analysis, and basic mathematical design using Fourier transformation, vector and ratio analysis.

C. vText process flow

 


Note: You can use the threshold at around 77 as the minimum score to recognize the same Note
: To achieve a successful comparison within 3 attempts, you may need to design your database to store at least 2 voice samples for each users. These 2 samples must match >= 77% then when in use for comparison, the benchmark should set at 70. This will avoid user frustration and at the same time achieve reliability because at 70% it is easier to pass through and so far tests have shown no 2 voice samples from 2 different users have reached >= 70%. If there is, you can adjust to increase to an optimum value. 

 


Please email enquiry@basic-signalprocessing.com for enquiries or implementation.